text
stringlengths 0
1.19M
|
|---|
= |
hf/2) is ceramic-rich and its bottom surface (z |
= −hf/2) is metal-rich. The effective material properties of FGM layer are assumed temperature-dependent and are graded in the thickness direction according to a simple power law distribution in terms of the volume fractions of the constituents, as followswhere P denotes a generic material property, Pc and Pm are the corresponding values of the ceramic and metal constituents and n is the volume fraction exponent that is the positive real value.The constituent material properties are considered temperature-dependent as follows where P−1, |
P0, |
P1, |
P2 and P3 are the coefficients of temperature T (K) and are unique to each constituent.The temperature variation is assumed to occur in the thickness direction of laminated plate and the temperature field is considered constant in the xy-plane. Also, it is assumed that the changes in thermal conditions happen very fast and so there is no thermal dissipation in the processes (adiabatic processes). In such a case, the temperature distribution along the thickness can be obtained by solving a steady-state heat transfer equationK(z)=kphf/2<z<hp+hf/2kf(z)-hf/2<z<hf/2kp-hp-hf/2<z<-hf/2T(z)=Tp(z)hf/2<z<hp+hf/2Tf(z)-hf/2<z<hf/2Tp′(z)-hp-hf/2<z<-hf/2In the above equations, K indicates thermal conductivity and subscripts p and f are related to the piezoelectric and FGM layers, respectively.It is assumed that the upper and lower surfaces of laminated plate are subjected to the constant temperatures TU and TL, respectively. Therefore, Eq. can be solved by imposing boundary conditionsTp(hf/2)=Tf(hf/2),Tf(-hf/2)=Tp′(-hf/2)kpdTp(z)dzz=hf/2=kfdTf(z)dzz=hf/2,kpdTp′(z)dzz=-hf/2=kfdTf(z)dzz=-hf/2It should be noted that, in the present study, the steady-state heat conduction is considered when the mechanical load is transient. Qian and Batra The displacement field is assumed according to the HSDT u=u0+zϕx-c1z3ϕx+∂w0∂x;v=v0+zϕy-c1z3ϕy+∂w0∂y;w=w0where u0, v0 and w0 denote the displacements of mid plane and ϕx, ϕy are the rotations of a transverse normal about the y and x axis, respectively and c1 |
= 4/(3h2).The nonlinear strain–displacement relations, based on von Karman’s large deformation assumption using HSDT, are εxxεyyγxy=εxx(0)εyy(0)γxy(0)+zεxx(1)εyy(1)γxy(1)+z3εxx(3)εyy(3)γxy(3),γyzγxz=γyz(0)γxz(0)+z2γyz(2)γxz(2)εxx(1)εyy(1)γxy(1)=∂ϕx∂x∂ϕy∂y∂ϕx∂y+∂ϕy∂x,εxx(0)εyy(0)γxy(0)=∂u0∂x+12∂w0∂x2∂v0∂y+12∂w0∂y2∂u0∂y+∂v0∂x+∂w0∂x∂w0∂yεxx(3)εyy(3)γxy(3)=-c1∂ϕx∂x+∂2w0∂x2∂ϕy∂y+∂2w0∂y2∂ϕx∂y+∂ϕy∂x+2∂2w0∂x∂yγxz(2)γyz(2)=-c2∂w0∂x+ϕx∂w0∂y+ϕy,In the above relations, ɛxx, |
ɛyy and γxy denote in-plane strains. Also, γyz and γxz are transverse shear strains and c2 |
= 3c1., the strain vector can be divided into linear and nonlinear parts as belowwhere ε→={εxεyγxyγyzγxz}T and subscripts L and NL are related to the linear and nonlinear parts, respectively. In this manner, the nonlinear part of strain vector can be presented byIn this study, constant transverse normal displacement (zero transverse normal strain) in displacement field is assumed (Eq. ). In order to avoid Poisson locking, zero transverse normal stress (σz |
= 0) must be imposed in constitutive equations σ→=[Q]ε→-[e]TE→-[Q]α→θD→=[e]ε→+[K]E→+P→θS=α→T[Q]Tε→+P→TE→+cθIn the above relations, the scalar quantities θ, |
c and S are temperature, specific heat and entropy, respectively. Also, σ→,ε→,E→,D→,P→ and α→ denote the stress, strain, electric field, electric displacement, pyroelectrical constant and thermal expansion coefficient vectors, respectively. In addition, the mentioned vectors are presented as belowσ→=σxσyσxyσyzσxzTε→=εxεyγxyγyzγxzTE→=E1E2E3TP⇀=P1P2P3TD→=D1D2D3Tα→=αxαy000T[Q], [e] and [K] represent reduced elastic constant, reduced piezoelectric stress constant and dielectrical coefficient matrices, respectively. The piezoelectric stress constant matrix can be obtained bywhere [d] is reduced the piezoelectric strain constant matrix. Aforementioned matrices have the following forms[Q]=Q11Q12000Q21Q2200000Q4400000Q5500000Q66,[K]=K11000K22000K33[e]=0000000000e31e32000,[d]=0000000000d31d32000Functionally graded materials are isotropic materials. Whereas, in general, the piezoelectric materials have orthotropic properties. In this study, it is assumed that the piezoelectric material is transversely isotropic with the axis of transverse isotropy coincident with the poling direction (z-axis). Elements of reduced elastic constant matrix for orthotropic material (assuming σz |
= 0) are presented as below Q11=E11-ν12ν21,Q12=Q21=ν12E21-ν12ν21,Q22=E21-ν12ν21Q44=G12,Q55=G23,Q66=G13where E, ν denote the Young’s modulus and Poisson’s ratio, respectively. Also, the subscripts 1, 2 and 3 represent x, y and z axes (shown in , can be presented in another form as belowwhere σ→1 is the elastic stress vector and σ→0 denotes initial stress vector due to thermal and electrical loads. These vectors are presented byInitial stress has only in-plane components and therefore can be expressed byThe potential energy of an FGM plate with surface-bonded piezoelectric layers under thermal and electrical loads is given by U=∫V12ε→T[Q]ε→-ε→T[e]TE→-ε→Tβ→θ-12E→T[k]E→-E→TP→θ-12cθ2dv, the potential energy is obtained as belowU=∫V12ε→LT[Q]ε→L+12ε→NLT[Q]ε→L+12ε→LT[Q]ε→NL+12ε→NLT[Q]ε→NL-ε→LT[e]TE→-ε→NLT[e]TE→-ε→LT[Q]α→θ-ε→NLT[Q]α→θ-12E→T[K]E→-E→TP→θ-12cθ2dvIn the above relation, the term -ε→NLT[e]TE→-ε→NLT[Q]α→θ can be replaced by ε→NLTσ→0 using Eq. . This term is related to the stored potential energy from the initial stresses due to thermal and electrical loads. Therefore, the potential energy can be written as belowU=∫V12ε→LT[Q]ε→L+12ε→NLT[Q]ε→L+12ε→LT[Q]ε→NL+12ε→NLT[Q]ε→NL-ε→LT[e]TE→-ε→LT[Q]α→θ+ε→NLTσ→0-12E→T[K]E→-E→TP→θ-12cθ2dvThe kinetic energy of an FGM plate with surface-bonded piezoelectric layers is presented bywhere ρ is mass density and u→ is displacement vector as belowThe work done by external loads is given byWext=∫Vu→Tf→bdv+∫Au→Tf→sdA+∑i=1Nu⇀Tf→pi+∫AqeϕdAwhere f→b,f→p and f→s are body force, concentrated load and traction vectors, respectively. Also, qe, |
ϕ and N denote charge density, electric potential and number of concentrated loads, respectively. in Hamiltonian’s principle, the variational form of motion equation is obtained as below∫V-δu→Tρu→¨-δε→LT[Q]ε→L-δε→LT[Q]ε→NL+δε→LT[e]TE→+δε→LT[Q]α→θ-δε→NLT[Q]ε→L-δε→NLT[Q]ε→NL-δε→NLTσ→0+δE→T[e]ε→L+δE→T[K]E→+δE→TP→θ+δu→Tf→bdv+∫Aδu→Tf→sdA+∑i=1Nδu→Tf→pi+∫AqeδϕdA=0In the present study, the four-noded rectangular conforming element based on HSDT is employed. The considered element is C1-continuous with eight nodal degrees of freedom The generalized displacements u0, |
v0, |
ϕx and ϕy over the element are approximated in terms of corresponding nodal values by Lagrange interpolation functions. In addition, Hermite interpolation functions are used to approximate w0. These interpolation functions have been presented in Therefore, the nodal degrees of freedom over the element can be presented in terms of corresponding nodal values (u→(e)) bywhere [Nu] is the shape function matrix., the displacement vector u→ can be written as[H]=100-c1z300z-c1z300100-c1z300z-c1z300100000Linear and nonlinear parts of strain vector are related to u→(e) by matrices [Bu]L and [Bu]NL1 respectively as belowFor thin plates, in which the length-to-thickness ratio is greater than 6, the electric potential varies linearly across the thickness where ϕL and ϕU are the electric potential corresponding to lower and upper piezoelectric layers, respectively. The electric potential corresponding to each piezoelectric layer (ϕ) can be presented in terms of ϕ→(e) as followsThe electric field vector is defined as negative gradient of electric potential as below, the electric field vector corresponding to each piezoelectric layer is related to ϕ→(e) by matrix [Bϕ] as followsIt is noted that in this study, the electric field considered only has non-zero-valued component in z direction (Ez). Also, it is assumed that electric potential variation through the thickness of piezoelectric layers is linear. can be obtained, easily. By substituting Eqs. and assembling elements, the nonlinear finite element motion equations of FGM plate with surface-bonded piezoelectric layers considering electrical, thermal and mechanical loads are obtained as below[Muu]U→¨+([Kuu]L+[Kuu]NL+[Kuu]G)U→+[Kuϕ]Φ→=f→m+f→uθThe matrices and vectors in above equations are given by[Kuu]L=∫V[Bu]LT[Q][Bu]Ldv[Kuu]NL=∫V([Bu]LT[Q][Bu]NL1+[Bu]NL2T[Q][Bu]L+[Bu]NL2T[Q][Bu]NL1)dv[Kuu]G=∫V[Nu]T[S0][Nu]dv[Kuϕ]=∫V[Bu]LT[e]T[Bϕ]dv[Kϕu]=∫V[Bϕ]T[e][Bu]Ldv[Kϕϕ]=∫V[Bϕ]T[K][Bϕ]dvf→m=∫V[Nu]T[H]Tf→bdv+∫A[Nu]T[H]Tf→sdA+∑i=1N[Nu]T[H]Tf→pif→uθ=∫V[Bu]LT[Q]α→θdvf→ϕθ=∫V[Bϕ]TP→θdvf→q=∫Aqe[Nϕ]TdA, [Kuu]L, [Kuu]NL and [Kuu]G are linear, nonlinear and geometric stiffness matrices, respectively.It is noted that geometrically nonlinear effects has been appeared in the nonlinear stiffness matrix. Nonlinear stiffness is a function of plate deflection and has considerable effects on motion equations in large deformations. Therefore, finite element equations are nonlinear.Geometric stiffness matrix is related to stiffness from initial stresses due to thermal and electrical loads.[Meq]U→¨+[Keq]U→=f→m+[Kuϕ][Kϕϕ]-1f→q-[Kuϕ][Kϕϕ]-1fϕθ+f→uθwhere [Keq] and [Meq] are equivalent stiffness and mass matrices defined by[Meq]=[Muu][Keq]=[Kuu]L+[Kuu]NL+[Kuu]G+[Kuϕ][Kϕϕ]-1[Kϕu]In the present study, the Gaussian integration scheme has been implemented to evaluate integrals involved in different matrices. To prevent the shear locking phenomena, the reduced integration technique is used to integrate terms related to the transverse shear stress. In addition, nonlinear natural frequencies of plate corresponding to different amplitude vibrations are obtained by analysis of the nonlinear eigenvalue problem associated to Eq. . The solution procedure of nonlinear eigenvalue problem has been presented by Sundararajan et al. In the present study, a computer program has been developed to solve nonlinear finite element equations using FORTRAN 90. Numerical results to analyze nonlinear natural frequencies and dynamic response of plate are presented in the next section.In order to verify the presented finite element model, nonlinear static deflection of an FGM plate under mechanical loads is investigated and compared with the obtained results by Praveen and Reddy Load parameter is defined as (q0a4)/(Emh4), where q0, |
Em, |
a and h are distributed load, Young’s modulus corresponding to the metal constituent of FGM, side and thickness of plate, respectively. Also, dimensionless deflection is considered as W/h, where W is the nonlinear static deflection. To indicate variation of nonlinear deflection in terms of distributed load, the dimensionless deflection versus load parameter is shown in . Volume fraction exponent of FGM is considered as n |
= 2. According to , it can be seen that for low static loads, the obtained results in this study and the results presented by Praveen and Reddy As another example to verify the presented finite element model, normalized nonlinear fundamental frequencies of an FGM plate with integrated piezoelectric layers under thermal and electrical loads are investigated and compared with the obtained results by Yang et al. Zirconia (ZrO2) and aluminum (Al) are the constituents of the substrate FGM layer and PZTG-1195N is selected as the piezoelectric material. Yang et al. The plate is subjected to uniform temperature rise ΔT and electric voltage VU and VL in upper and lower piezoelectric layers, respectively. A stress-free temperature T0 |
= 0 C and FGM layer with n |
= 2 is considered.The normalized nonlinear fundamental frequency is defined as ωNL/ωLo, where ωNL is the nonlinear fundamental frequency of plate considering thermal and electrical loads. Also, ωL0 is the linear fundamental frequency of plate in the absence of thermal and electrical loads. Normalized amplitude is considered as Wmax/hf, where Wmax is the maximum vibration amplitude.Normalized nonlinear fundamental frequencies of plate corresponding to different normalized vibration amplitudes under temperature rise ΔT |
= 0, 300 C and applied electric voltage VU |
= |
VL |
= −200 V are presented in . In a coupled thermo-piezoelectric problem, frequencies are generally complex due to energy dissipation caused by heat conduction indicates that the differences between the results obtained in this study and the results obtained by Yang et al. Having validated the model to a certain extent, nonlinear natural frequencies and dynamic response of an FGM plate with integrated piezoelectric layers are presented in the tabular and graphical forms.The FGM constituents are silicon nitride (Si3N4) and stainless steel (SUS304). Young’s modulus and thermal expansion coefficient corresponding to these constituents are assumed temperature-dependent and presented in . Other material properties are considered temperature-independent and shown in . The upper surface of FGM layer is ceramic-rich whereas the lower surface is metal-rich. The piezoelectric material is PZT-4 with material properties presented in . The hybrid FGM plate is square with side of a |
= 0.4 m. The thickness of the substrate FGM layer is hf |
= 0.04 m and the thickness of each piezoelectric layer is hp |
= 1 mm. Constant temperatures TU and TL are applied on upper and lower surfaces of the laminated plate, respectively. A stress-free temperature T0 |
= 300 K is used. Electrical load is considered as applied voltages VU and VL in upper and lower piezoelectric layers, respectively. Clamped boundary condition is considered for all edges of plate as belowx=0,au0=v0=w0=∂w0∂x=∂2w0∂x∂y=ϕx=ϕy=0y=0,bu0=v0=w0=∂w0∂y=∂2w0∂x∂y=ϕx=ϕy=0Also, simply supported boundary condition is considered as:Dimensionless (linear or nonlinear) natural frequency (ω¯) is defined aswhere ω is linear or nonlinear natural frequency (rad/s). Also, ν, |
ρ0 and E0 denote Poisson’s ratio, mass density and Young’s modulus corresponding to metal constituent of FGM at the room temperature (T0 |
= 300 K). Dimensionless maximum amplitude vibration of plate is defined as Wmax/hf, where Wmax is maximum amplitude vibration of the plate.In a finite element analysis, it is desirable to have the convergence studies to estimate the order of mesh size to be necessary for the numerical solution. indicates the convergence studies for the first two dimensionless nonlinear natural frequencies of a fully clamped FGM plate with surface-bonded piezoelectric layers. It is observed that a mesh size of 10 |
× |
10 is sufficient to get a reasonable order of accuracy. The analysis in the subsequent problems is carried out with this mesh size. shows dimensionless linear and nonlinear fundamental frequencies of a fully clamped FGM plate with surface-bonded piezoelectric layers under three thermal conditions with no electrical load (VU |
= |
VL |
= 0). It is seen that linear and nonlinear natural frequencies are the highest for ceramic plate and lowest for the metallic plate. It is due to this fact that Young’s modulus of ceramic plate is maximum. Therefore, the ceramic plate has the highest stiffness and so the highest natural frequencies. Also, by increasing the volume fraction exponent, Young’s modulus decreases and so linear and nonlinear natural frequencies of FGM plate decrease.Nonlinear analysis compared considers additional stiffness for the plate. So, natural frequencies obtained based on nonlinear analysis are higher than linear frequencies. Results show that nonlinear natural frequency increases with the increase of vibration amplitude. Therefore, the effect of geometric nonlinearity, which is to raise stiffness and so natural frequencies of the plate, is more important at large vibration amplitudes. show that the linear and nonlinear natural frequencies increase with the decrease of temperature. Also, temperature rise results in decreasing the linear and nonlinear natural frequencies. Indeed, temperature rise makes compressive in-plane forces in the plate and consequently, geometric stiffness of plate decreases. This results in decreasing the natural frequencies. Tensile in-plane forces are made in the plate by decreasing the temperature. In this case, geometric stiffness of the plate increases and therefore natural frequencies increase. shows dimensionless linear and nonlinear fundamental frequencies of a fully simply supported FGM plate with surface-bonded piezoelectric layers under three thermal conditions with no electrical load (VU |
= |
VL |
= 0). Comparing the results presented in this table with the corresponding results in indicates that linear and nonlinear natural frequencies of fully clamped plate are higher than the natural frequencies of the same plate with fully simply supported boundary conditions. It is because of the fact that clamped boundary condition makes higher stiffness in the plate compared to simply supported boundary condition.In order to investigate effect of thermal load on nonlinear natural frequencies of the plate in graphical form, a fully clamped FGM plate with surface-bonded piezoelectric layers is considered. The lower surface of the laminated plate is subjected to constant temperature TL |
= 300 K. Temperature on upper surface of plate is varied from TU |
= 300 K to TU |
= 1000K. shows variation of dimensionless nonlinear fundamental frequency of the plate corresponding to vibration amplitude Wmax/hf |
= 0.4 versus variation of temperature on the upper surface of plate for different volume fraction exponents. It is seen that temperature rise results in decreasing the nonlinear natural frequencies of the plate. shows effect of applied voltage in piezoelectric layers on dimensionless linear and nonlinear fundamental frequencies of the fully clamped plate. Results have been presented for thermal condition TU |
= |
TL |
= 300 K and electrical loads VU |
= |
VL |
= 0 and VU |
= |
VL |
= 200 V. It is seen that in the considered condition, applied voltages in piezoelectric layers have no considerable effect on linear and nonlinear natural frequencies. It is because of the fact that the thickness of both piezoelectric layers is small compared to the thickness of FGM layer. In this particular condition, electrical load makes small in-plane forces in the plate and so geometric stiffness do not change considerably. This results in small change in natural frequencies of the plate. In Section 5.2.3.2, the thickness of FGM layer will be reduced to magnify the effect of applied voltages in piezoelectric layers on vibration behavior of the laminated plate. shows dimensionless fundamental frequencies of FGM plate with surface-bonded piezoelectric layers for two boundary conditions, fully clamped (CCCC) and fully simply supported (SSSS). It is seen that linear and nonlinear natural frequencies of fully clamped plate are higher than the same plate with fully simply supported boundary conditions. It is clear that clamped boundary condition makes higher stiffness in the plate compared to simply supported boundary condition.It is well known that significant geometrical nonlinearity is induced when a plate has deflection of more than approximately one-half of its thickness, especially when there are immovable edge constraints shows free vibration of the plate based on linear and nonlinear analysis. In this study, dynamic responses are presented for the central point of the plate. It is seen that results obtained from linear and nonlinear analysis have no considerable difference. It is because of this fact that vibration amplitude (about 6 mm) is small compared to the laminated plate thickness (42 mm). shows free vibration of the plate from initial condition caused by applied uniform distributed force q |
= 2000 MPa based on linear and nonlinear analysis. It is observed that linear and nonlinear dynamic responses are quite different. In this case, vibration amplitude is large (about thickness of laminated the plate) and so geometric nonlinearity is considerable. Therefore, nonlinear analysis compared to linear analysis considers larger stiffness for the plate. As a result, dynamic response from nonlinear analysis has lower amplitude and higher frequency compared to linear analysis. indicates that nonlinear vibration amplitude is about 30 mm that corresponds to Wmax/hf |
= 0.75. Referring to , dimensionless nonlinear natural frequency associated to the mentioned vibration amplitude can be obtained by interpolation about 20.14 that corresponds to period of 0.24 ms. This value coincides to period of nonlinear vibration shown in . It is clear that period of linear vibration has not been changed relative to the case of small amplitude vibration (shown in shows free vibration of the plate from initial condition caused by different applied uniform distributed loads (q |
= 100, 200, 400 MPa) based on nonlinear analysis. It is observed that frequency of vibration is independent of the vibration amplitude. Also, by doubling the applied load, vibration amplitude has become double. In this case, vibration amplitudes are small relative to the thickness of the plate and therefore nonlinear analysis coincides to linear analysis.Free vibration of the plate from initial condition caused by uniform distributed loads q |
= 1000, 2000, 4000 MPa are shown in based on nonlinear analysis. It is seen that frequency of vibration increases with the increase of vibration amplitude. Also, by doubling the applied load, vibration amplitude has not become double. It can be seen from that vibration amplitudes are large and so geometric nonlinearity has considerable effect on vibration behavior of the plate. shows effect of volume fraction exponent on large amplitude free vibration of the fully clamped plate from initial condition caused by uniform distributed load q |
= 2000 MPa. With the increase of volume fraction exponent, plate stiffness decreases and therefore vibration amplitude increases whereas frequency of vibration decreases. indicates large amplitude free vibration of the plate for different upper surface temperatures TU |
= 300 K, 600 K, 1200 K. Temperature on lower surface of the plate is maintained constant at TL |
= 300 K. Temperature rise decreases the plate stiffness and so increases the vibration amplitude and decrease the vibration of frequency. In addition to the dynamic oscillation, there is a static offset induced by the temperature gradient (thermally induced static deflection). In this study, the thermally induced static deflection for TU |
= 600 K and TU |
= 1200 K are 1.07 mm and 1.5 mm, respectively. Therefore, the thermally induced static deflection is small compared to the total thickness of the laminated plate (42 mm) and has small effect on nonlinear vibration behavior of the plate.In order to investigate nonlinear forced vibration behavior of the laminated plate, a harmonic distributed load with amplitude of 2000 MPa and period of 0.1 s is applied on the surface of plate. Upper and lower surfaces of the laminated plate are subjected to the temperatures TU |
= 600 K and TL |
= 300 K, respectively. Applied voltages in upper and lower piezoelectric layers are VU |
= |
VL |
= 200 V. Volume fraction exponent for FGM layer is considered as n |
= 2. shows nonlinear time response of FGM plate with surface-bonded piezoelectric layers under thermal, electrical and mechanical loads. indicates effect of FGM volume fraction exponent on nonlinear forced vibration of the laminated plate. It is seen that with the increase of volume fraction exponent, plate stiffness reduced and so vibration amplitude increases. Also, it is observed that vibration frequency is the same as frequency of the applied force.Effect of the upper surface temperature on nonlinear forced vibration of the plate is presented in . Temperature on lower surface is maintained constant at TL |
= 300 K. It is observed that temperature rise results in the increase of vibration amplitude. As mentioned, with the increase of temperature, plate stiffness reduces.In the considered condition, the thickness of piezoelectric layers is small compared to the thickness of FGM layer. Therefore, applied voltages in piezoelectric layers have no important effect on plate stiffness and do not affect time response of the plate, considerably.Nonlinear frequency response of fully clamped hybrid FGM plate for different values of FGM volume fraction exponent is shown in . In this figure, excitation is a distributed load with amplitude of 250 MPa. It can be observed that the nonlinear effect bends the resonance peaks to the right. In addition, with the decrease of FGM volume fraction exponent, nonlinear frequency response shifts rightwards, as expected. indicates effect of thermal load on nonlinear frequency response of fully clamped hybrid FGM plate for a distributed load with amplitude of 500 MPa. It is observed that with the decrease of temperature of upper surface, nonlinear frequency response shifts rightwards. Indeed, increase in temperature of the upper surface decreases the stiffness and natural frequency of the plate.In the considered condition, the thickness of both piezoelectric layers is small compared to the thickness of FGM layer. Therefore, the applied voltages in piezoelectric layers have no considerable effect on vibration behavior of the plate. In order to magnify the effect of piezoelectric voltages on nonlinear frequency response of the plate, the thickness of FGM layer is reduced to 4 mm. Nonlinear frequency response of fully clamped hybrid FGM plate is shown in for two cases: no applied voltage in piezoelectric layers and applied voltage in piezoelectric layers making VU |
= |
VL |
= 0. It is seen that nonlinear frequency response shifts to the left with the applied voltages in piezoelectric layers. indicates that the effect of applied voltages is considerable when the thickness of piezoelectric layers is not small compared to the thickness of FGM layer. In fact, applied voltages in piezoelectric layers make compressive in-plane forces in the plate and consequently, geometric stiffness of plate decreases. This results in decreasing the natural frequencies and increasing the vibration amplitude of the plate.In this study, nonlinear natural frequencies and time response of FGM plate with surface-bonded piezoelectric layers have been investigated under thermal, electrical and mechanical loads. In addition, nonlinear frequency response diagrams were presented and effect of some parameters, .i.e. FGM volume fraction exponent, temperature gradient, and piezoelectric voltage are investigated. In this regard, the finite element formulation based on HSDT has been developed that includes geometric nonlinearity using von Karman’s assumptions.The material properties of FGM are assumed to be temperature-dependent and are graded in the thickness direction according to a simple power law distribution in terms of the volume fractions of the constituents. The temperature field considered is assumed to be a uniform distribution over the plate surface and varied in the thickness direction and the electric field considered only has non-zero-valued component Ez.Numerical results show that volume fraction exponent, thermal environment and vibration amplitude have considerable effects on free and forced vibration of the hybrid FGM plate. In contrast, applied voltage in piezoelectric layers has not important effect on nonlinear natural frequencies and dynamic response of the plate.Tensile properties and strengthening effects of Al 3003 alloy weldment reinforced with TiO2 nanoparticlesThis paper examines the significance of strengthening mechanisms in the fabricated Al 3003 nanocomposite weldment using accumulative roll bonded (ARB) Al 1100 – TiO2 filler rod. The wt. % of titanium oxide (TiO2) content is varied as 0, 0.75, 1.5, 2.25 and 3 to the total weight of the weld zone. The weld possesses an ultimate tensile strength of 249 ± 04 MPa in 3% reinforced whereas 137 ± 04 MPa in unreinforced condition. The results obtained from the tensile test were cross checked with the characterization studies such as X ray line profile analysis (XRD), Field emission scanning electron microscopic analysis (FESEM), Transmission electron microscopic analysis (TEM), Electron backscattered diffraction analysis (EBSD) and Energy dispersive x-ray spectroscopic analysis (EDS). The tensile test results reveal that Al – 3 wt% TiO2 nanocomposite weld exhibits higher strength which is not only due to drop in grain size but also due to mechanisms such as grain refinement strengthening, Orowan strengthening and thermal expansion coefficient mismatch strengthening and occurred incommensurate with the progressive addition of TiO2 nanoparticles in the weldment. However, it has been found that the mechanism due to Orowan strengthening has played a significant and greater role for the overall strength of the weldment.Incorporation of nanosized hard ceramic particles in the matrix material boosts the mechanical properties of composites. Improvement in properties leads to variations in hardness, yield strength, ultimate tensile strength and toughness. At present, Al – Mn alloy, known as Al 3003 alloy, is a material which possesses moderate strength and properties. Additionally, it can be applicable to radiators, oil tanks and heat exchangers in refrigerators, etc. []. In order to attain the required properties and enrich its applications extensively, it is essential to modify its grains into fine/ultrafine level through the addition of second phase particles and/or severe plastic deformation technique. Various reinforcements such as Al2O3, AlN, B4C, TiC, ZrC, ZrO2, etc. were tried by previous researchers by incorporating into Al alloy to take it to superior properties. There is a lack of detailed investigation on TiO2 reinforced metal matrix composites. TiO2 holds excellent thermal and mechanical properties such as hardness, strength, coefficient of thermal expansion (CTE), thermal shock resistance [Among several arc welding processes, gas tungsten arc welding process has the capability of joining both ferrous and nonferrous alloys by applying heat to take the base metal to molten stage with the help of filler rod and shielding gas in order to add the second phase particles and protect the weldment from the atmosphere, respectively []. Fabrication of nanocomposite fillers is regarded as the emerging trend in scientific community to uplift the load-bearing capacity of the welded joint. Generally, the possible approaches which are used to nanocomposite filler metal are powder metallurgy and stir casting route. Though powder metallurgy route aids to obtain the homogenous distribution of reinforcement in the matrix, porosities due to incomplete bonding between powder particles and less productivity hinder this technique to fabricate filler metal. In stir casting route, even though this method is simple and effective for high volume production, poor wettability, clustering and weak interfacial bonding between matrix and reinforcement are undesired eventualities [Accumulative roll bonding (ARB) technique is a suitable process to overcome the aforesaid shortcomings. ARB requires several cycles of cleaning, brushing, sandwiching, rolling at 50% reduction and parting off the sheets in which the nano-sized ceramic particles are well spread in-between the sheets []. Several researchers have suggested this technique to obtain the required properties. The reason for selecting ARB is to create a homogenous distribution of ceramic nano/microparticles in the filler rod and increase the interfacial bonding within the composites.Fattahi et al. (2015) blended TiC/Al powders through mechanical alloying (MA) process and sheath rolling route. During MA, TiC nanoparticles are homogenously mixed with Al matrix boosting, the dispersion strengthening of the material which results in enhancement of yield strength of the composite. Sivasankaran et al. (2017) synthesised the gliding Cu – Al2O3 for the duration of 20 h to transform the micron-sized particles to nano-sized. Continuous MA decreases the grain size due to high energy collision, cold welding and fracturing. At last, grain refinement strengthening has involved more and increased the tensile properties of the composite. Schmidt et al. (2011) sprayed the nanosized reinforcement particles between two sheets and roll bonded to produce nanocomposite filler metal in order to modify the structural and mechanical properties of weld. While staking, the powder particles were accommodated within the gap made by scratching. During rolling, the nanosized particles produced extraordinary dislocations leading to increased yield strength of the filler metal which is a major contribution of dislocation strengthening [The objective of this investigation is to: (a) Assess the feasibility of roll bonded Al – TiO2 filler metal by joining AA 3003 to create nanocomposite weldment. (b) Examine the structural morphologies involving XRD, SEM, TEM, EDS and EBSD analyses and mechanical property evaluation in terms of ultimate tensile strength, yield strength. (c) Compute the contribution of each strengthening effect on the overall strength of the weldment.Al 1100, Al 3003 alloy are chosen as the filler metal and base metal to weld with the purity of 99.5% each their respective chemical compositions are given in . Al 1100 and TiO2 nanoparticles were chosen as to prepare nanocomposite filler metal through ARB method. The as-received TiO2 is in anatase phase with 98.5% purity. The filler metal fabrication process and GTA welding using the fabricated filler metal are illustrated schematically in . As received TiO2 nanoparticles were ultrasonicated using acetone for 10 min in order to avoid the clustering between each particle. (a) shows the TEM micrograph of TiO2 particles in as-received condition which showed that TiO2 particles are cluster free and average size is 27 ± 04 nm in dimension which has been measured using Image J software. (b) depicts the frequency versus particle size chart which ensured that most of the particles are in the range of 25–30 nm and the maximum size of the particles are not more than 40 nm. (c) depicts the XRD pattern of TiO2 nanoparticles which assists to observe the TiO2 peaks in the composite weldment. The Al 1100 and Al 3003 alloy were procured from M/s. Coimbatore metal mart, TN, India and TiO2 is purchased from M/s. Loba Chemie, India. The as-received Al 1100 sheets (300 × 200 × 1 mm of length x width x thickness) were scratched using stainless steel brush and cleaned (degreased) using acetone to remove the oxide layer, dirt and other dust particles over the sheets. Afterward, ultrasonicated TiO2 nanoparticles were sprayed in between the degreased sheets and allowed to dry for a few minutes. Then the sheets were stacked like a sandwich and rolled up to 50% reduction in cross-section. After rolling, the sheets were parted off into two sheets and the same cycle is repeated. Then the sheets were heated up to 300 °C for 5 min to relieve the stress. Further, the sandwiched sheet was rolled up to eight passes to distribute the reinforcement particles uniformly without clustering. Finally, the fabricated filler metal is used to weld the base metal. The base metal is taken with the single V groove of 45° in the center to perform the welding operation which is schematically illustrated in . The cross section of the weld is polished and etched using Keller's reagent to examine the microstructural behaviour (reinforcement distribution) in the matrix. Tensile strength is measured by machining the specimen as per ASTM-E8 standard followed by fracture surface morphology, identified using SEM in order to find out the undergone fracture is either brittle or ductile. explains the machines/equipment used for rolling, welding, characterization and ultimate tensile strength (UTS). represents the XRD pattern of Al – x wt.% TiO2 nanocomposite weldment (x = 0, 1.5, 2.25 and 3). The presence of TiO2 particles have been ensured in 2.25 and 3 wt% reinforced nanocomposite weld. Both the intensity and width of the major peak (1 1 1) is increased incommensurate with the reinforcement content. The matrix peak positions are inconsistent in the fabricated nanocomposite weldment which appeared feasibly because of lattice distortion of the Al by reinforcement particles []. In 1.5 and 3 wt% reinforcement specimens, Al peaks moved from their real position when compared with the unreinforced specimen, which was due to the difference between the lattice constant of Al and TiO2 particles. This indicates the occurrence of crystallite refinement during the ARB process []. The results not only indicated the existence of reinforcement content but also the absence of intermetallic compounds which revealed that there was no chemical reaction has taken place between Al and TiO2 nanoparticles.FESEM micrographs of Al/TiO2 nanocomposite weld is illustrated in (a), (b) and (c) depict 0, 1.5 and 3 wt% reinforced weldment, respectively. The decrease in grain size was due to the addition of reinforcement particles. Inset in (c) depicts the magnified view of 3 wt% reinforced nanocomposite weld, in which TiO2 particles can be clearly seen that the cluster-free ceramic particles were attached to the grains and seated into nearby grain boundaries. This type of distribution was necessary to attain outstanding properties. Due to the addition of TiO2 content, the α-Al phase was highly refined and inhibited the grain growth dimensionally and stimulated the homogeneity in the Al grains []. Zener pinning effect and heterogeneous nucleation mechanism (HNM) on reinforcement lead to grain refinement as a result rate of grain growth decreases. Zener pinning effect implies that force acts to hinder the migration of boundary and the relevant equation is expressed as [], FZ = f/Kr, where f, k, r and FZ are defined as the volume fraction of TiO2, Zener coefficient, the radius of TiO2 and particle pinning force. According to HNM, ceramic particles work as the heterogeneous nucleation sites (HNS) leading to matrix grains growth initiation when solidifying. The mismatch between the lattice of TiO2 particles and Al matrix, elevate the mechanical properties of HNS due to a decrease in the activation energy for nucleation [ (d – f) depict the EDAX elemental mapping of Al/3 wt% TiO2 nanocomposite weld. The uniform distribution of elements namely titanium and oxide was confirmed in Al weldment due to the addition of ceramic particles in the filler. Hence, ARB is an effective technique to disperse the reinforcement uniformly in the entire filler metal and promotes wettability and enhances the bonding strength [ (g – i) displays the TEM micrographs of Al/3 wt% TiO2 nanocomposite weld. Fine/ultrafine grains are revealed in (g) which are mainly due to grain refinement led by Zener pinning effect and HNM on TiO2 particles, with the rate of grain growth decreasing eventually. The presence of clear interface was due to the absence of defects such as voids, cracks and clustering. (h) shows the homogenous distribution of TiO2 particles observed in the matrix grain, attributed by the fast cooling rate when welding []. Yield strength of the weld is relying on bond strength (interfacial) that transmits the load from the matrix to reinforcement [ (i) depicts the dislocation, dislocation pinning and strain fields. The witness of detected dislocations in the weld is due to (a) severe deformation of Al matrix to produce the composite weld and the deformed materials are prone to cause dislocation. The grains in the weldment were recrystallized and deformed multiple times before the plasticized nanocomposite weld got cooled. (b) There is a mismatch between the thermal properties of Al and TiO2. The heat involved in the expansion and contraction of both is different. Hence, a large amount of dislocations are created due to thermal mismatch. Finally, it contributes to hardness and strength of composites []. Further, more amount of dislocation pinning are seen in the weldment. In this, pinning of dislocation was due to the addition of second phase particles [ (a - c) illustrates the EBSD images of the 0, 1.5 and 3 wt% reinforced nanocomposite weldment. The effect of TiO2 addition on Al grain can easily be observed. An extensive grain refinement occurred in the composite layer. The refinement of grains are due to: (i) plastic deformation during welding and (ii) incorporation of hard ceramic particles which paves the way to pinning effect. The existence of TiO2 particles makes the matrix stronger for the free growth of α-Al grains. The corresponding grain growth is controlled and the grain size are compacted in the nanocomposite weld. This effect is multiplied with the addition in wt.% [In order to investigate the relationship between microstructural features and strengthening properties of fabricated nanocomposites weldment, at times, both the analytical and numerical modeling are required. This will enable prediction of yield strength, deformation and damage characteristics [The yield strength of any metal/alloy/composite can be calculated through various strengthening mechanisms []. The aim of the present work is to analyze the strengthening mechanism involved in the Al 3003 weldment with the dispersion of TiO2 nanoparticles. Through characterization studies, it is clearly noticed that TiO2 dispersion, generation of dislocation, grain size reduction have occurred within the weldment. The UTS and YS of Al/TiO2 nanocomposite weld are shown in (a and b). The UTS and YS of reinforced nanocomposite weldment are directly proportional to reinforcement, in addition. This is mainly attributed to the superior interfacial bonding strength, wettability of TiO2 surrounded by the matrix and the proper spreading of reinforcements in the weldment. Hence, the accomplishment of the desired properties is made by the route of ARB []. Discussion on strengthening mechanisms that are likely to have occurred in the present investigation is detailed below.There are several ways to elevate the strength and toughness of metals and non-metals. The contribution on the total yield strength of Al – x wt.% TiO2 nanocomposites is given in . In this investigation, TiO2 nanoparticles in the Al, obstruct the motion of dislocation in the boundary. Grain boundary strengthening is the result of grain boundary strengthening and grain boundary pinning. The grain refinement strengthening contribution can be expressed (Hall–Petch relation) [where ΔσGB, ΔσO, K (0.04 MPa m1/2), d are the grain boundary strengthening, friction stress, average grain size and constant respectively.Thermal mismatch, in theory, explains that change in temperature will produce stress in the composite and lead to the variation in material strength. Therefore, the influence of thermal mismatch on the strength of the material is generally considered in the composite. The incorporation of TiO2 ceramic particles in the nanocomposite weld results in improvement of dislocation density and increases the yield strength. Dislocations can be formed during solidification of the weld pool due to CTE variance between the Al and TiO2 particles. The following equation has enabled to evaluate the increase in the yield strength which is due to CTE mismatch [where α = 1.25 (constant), G = 26.9 GPa (shear modulus), b = 0.28 nm (Burgers vector), ΔT = 633 K (the melting and ambient temperatures variation), ΔC = 14.3 × 10−6 K−1 (CAl = 23.6 × 10−6 K−1 and CTiO2 = 9.3 × 10−6 K−1), Vp (volume fraction of reinforcement) and dp (average particle size of reinforcement).In Orowan/dispersion strengthening mechanism, the strength of the composite would be improved without loss of ductility. The reinforcement particles hindered the dislocations motion which leads to generating a dislocation loop around the particles. Further migration of dislocation is prevented by generated loops. This results in an Al matrix to strain hardened and makes the weld stronger. The improvement in yield strength by ΔσOrowan is stated as [Δσorowan=0.81MGb2Π(1−υ)1/2ln(dp/b)dp(123Π2VP−1)where M = 6.5 (Taylor factor) and n = 0.3 (Poisson ratio). The yield strength of TiO2 reinforced nanocomposite welds were computed and then compared with the experimental results. The experimental values were much lower than the calculated values. Nevertheless, the obtained results indicated good agreement with the experimentally investigated values. The parameters and variables used in the equations can be varied as per the processing methods. The predicted Δσ can be calculated by Ref. [ illustrates that the Orowan strengthening contribution among the overall strength is higher in the fabricated nanocomposite welds, followed by CTE strengthening and grain boundary strengthening in sequence. clearly shows the strengthening mechanism contribution takes place in the overall strength in Al – 3 wt% TiO2 nanocomposite weld. It can be concluded that the presence of TiO2 nanoparticles contribute more on Orowan mismatch strengthening and it has played a leading role in enhancing the yield strength of Al/TiO2 nanocomposite welds [The contribution of various strengthening mechanisms on total strength for Al – x wt.% TiO2 nanocomposite weldment (x = 0, 3 and 6) is portrayed in . It is further understood based on calculations through relevant equations intended for various strengthening mechanisms, the order of percent contribution towards strengthening is found to be 53.35% (dispersion strengthening), 23.7% (grain refinement strengthening) and 22.87% (dislocation strengthening). Thus it can be seen that the presence of nanoscale reinforcement particles contributes more on dispersion strengthening. Therefore, the incorporation of nano-titania particles on Al 3003 matrix will enhance the mechanical properties of the weldment much better [ (a - c) show the fractography analysis of unreinforced and 0, 1.5 and 3 wt% TiO2 reinforced nanocomposite weld along with the tested specimen after tension. In 0 and 1.5 wt% TiO2 reinforced specimens, fractures has occurred in the weld zone whereas in 3 wt% fracture has occurred in the base metal. This reveals that the strength of the weld is higher than the base metal. The fracture morphology of the unreinforced nanocomposite weld exhibits dimples of uneven sizes i. e larger dimples whereas fine dimples are seen in 1.5 and 3 wt% TiO2 reinforced specimens because of grain refinement. The dimensions of the dimples in Al-3 wt.% TiO2 specimen are visibly smaller than that with 1.5 wt% TiO2 which depicts that the refinement should have occurred in crystalline level. Absence of voids around the witnessed TiO2 nanoparticles indicates superior interfacial bonding strength between Al and TiO2. The incorporation of reinforcement to the weldment has caused hindering the motion of dislocations and upliftment of the dislocations density ( (g – i)) and the strain-hardening rate.In this research, novel Al 1100 - TiO2 nanocomposite filler metal is used to join Al 3003 weldment using ARB and GTAW techniques. The effect of TiO2 nanoparticles on the tensile properties and microstructure of the composite weld was studied in detail and the strengthening mechanism of both unreinforced and TiO2 reinforced composite welds was further investigated. From the obtained results, the following conclusions were summarized.XRD results not only revealed the presence of TiO2 phases exists in the Al nanocomposite weldment but also the absence of intermetallic compounds which showed that there was no reaction took place between the matrix and reinforcement during welding.Characterizations such as (i) TEM analysis results showed the proper distribution of nanolevel ceramic particles in the weldment without any cluster formation. (ii) FESEM micrographs displayed the refinement of grains occurred incommensurate with reinforcement addition. (iii) EBSD images proved that the size of the grains tends towards decreasing due to particle pinning growth.The strength of the composite welds was improved considerably when compared with Al - 0 wt% TiO2 weld. The following are the reasons assisted to increase the strength of the composite weldment. (i) Homogeneous distribution of reinforcement particles, (ii) grain refinement, (iii) dislocation density, and (iv) decrease in inter-particle spacing between the particles.In nutshell, among the involved strengthening mechanisms, Orowan (dispersion) strengthening has played a significant and greater role in the overall strength of the weldment followed by grain refinement strengthening and CTE mismatch (dislocation) strengthening.The following is the Supplementary data to this article:Supplementary data to this article can be found online at Improving the tensile strength of carbon nanotube spun yarns using a modified spinning processA modified process for the dry spinning of carbon nanotube (CNT) yarn is reported. The approach gives an improved structure of CNT bundles in the web drawn from the CNT forest and in the yarn produced from the twisted web leading to improved mechanical properties of the yarn. The process enables many different mechanical and physical treatments to be applied to the individual stages of the pure CNT spinning system, and may allow potential for the development of complex spinning processes such as polymer–CNT-based composite yarns. The tensile strength and yarn/web structure of yarn spun using this approach have been investigated and evaluated using standard tensile testing methods along with scanning electron microscopy. The experimental results show that the tensile properties were significantly improved. The effect of heat treatments and other yarn constructions on the tensile properties are also reported.Research studies on CNTs have revealed a unique atomic structure with interesting related properties such as high tensile strength and Young’s modulus, high aspect ratio, and good electrical and thermal conductivities Significant efforts have been carried out to fabricate macroscopic structures to make use of the CNT’s unique properties. Many promising techniques and results have been published The first publication on spinning continuous CNT yarn from Professor Fan’s group at Tsinghua University To date, two distinct methods have been utilized to improve the properties of CNT yarns, firstly the development of longer and better aligned CNT forests on wafers At the micro-structural level, within the yarn, there is a hierarchical structure containing two levels: (i) individual CNTs at the fundamental level and (ii) bundles of aggregated CNTs (so called secondary fibres). These bundles form a continuous network, called the web, with a preferred orientation along the yarn axis.a and b) show that the CNTs are not individualised in the web, instead they tend to be stripped from the forest wall in a range of bundle sizes. According to recent publications in which r is the CNT diameter and d is the gap between CNTs. The CNT interaction forces which cause bundles to be formed are based on van der Waals forces. The cohesive energy for the formation of a bundle over retaining a free-standing tube depends on tube diameter This interactive energy of CNT bundles have been approximated by summing up the inter-CNT atom–atom pair energies Φi, using a Lennard Jones potential as follows where ri is the distance between the ith atom pair, σ is the C–C bond length and A and y0 are constants. The parameters for the Lennard Jones potential (A, y0) have been further described by Girifalco and Lad where b is a constant value. While F1 depends on the interactive energy of CNT bundles described above, Qian et al. The mechanical properties of yarn depend not only on the interaction of CNTs in bundles but also on the degree of condensation (or packing) of CNT bundles in the yarn structure. Since the inter-tube interactions in CNT bundles is very weak This paper focuses on the dry spinning process as it avoids the use of volatile liquids which introduce additional hazards and operating costs. Zhang et al. c) from substrate-based ‘forests’ of MWCNTs (The success of the dry spinning process was attributed in part to the degree of twist applied during spinning. The role of twist was to reduce the inter-CNT distance whilst fixing the relaxed cross-section shape for individual CNTs in the bundle and also increasing the van der Waals forces (Eq. that twist (Ω) applied to the CNT web (2) produces the required inter-bundle lateral cohesion after the web was pulled out from forest wall on the wafer (7). The current dry spinning process schematically described in The CNT web pulled out of the wafer is poorly orientated (see b). Since the CNT web is twisted immediately after being pulled out from the forest this arrangement is locked in place. Thus it is difficult for the CNT bundles to be further aligned in order to improve the structure of yarn.The geometry (length) of the web is changeable and not controllable during spinning. (The length of the web is defined as the distance from the forest wall where CNTs were pulled out to the convergence point on web/yarn.) This causes heterogeneity in the tension of the web. Furthermore, the uniformity of the web tension depends on the ratio of the length to the width of the web strip. When the ratio is smaller (shorter triangle) the heterogeneity of web tension is higher. The experimental work showed that for the current dry spinning process, the length of triangle web is very short and the convergence point (6) (see ) tends to move close to the wall of CNT forest (7).Since there is no separation between the different stages (e.g. web formation and twist insertion) of the current spinning process, processing web and yarn is either very difficult or not efficient.In order to overcome these disadvantages, a modified dry spinning method is proposed and presented in the next section.The initial web of CNT bundles influences the packing structure of the yarn. For example, a web with a poorly oriented structure (b) results in yarns with less than optimal packing of the bundles because a large proportion of the CNT bundles are not aligned in the direction of the yarn. This produces yarns with inferior properties to those predicted from the properties of individual CNTs, that is, generally the tensile strength of CNT yarn spun by either the wet or dry-state spinning method is much lower than the strength of individual CNTs We postulate that the dry spinning of CNTs brings together the characteristics of both traditional staple spun yarn and synthetic filament spinning, in which the web and yarn structures are the key factors affecting the yarn’s physical and mechanical properties. In addition, the bundles of CNTs in web/yarn have similarities to the staple fibres of a spun yarn. Based on this idea and the interactions of CNTs in bundles described in Section , this work proposes a modified spinning system which seeks to enhance the CNT and bundle interaction by increasing the drawing and alignment of the CNTs in the yarn direction.For the modified spinning process, the system was partitioned into three distinct zones (see ): web formation zone (A), tensioning/drafting zone (B), and twisting zone (C). The following description is general in nature and detailed calculations of our new process will be represented in the future publication.In the web formed from the forest, the bundles are oriented along the yarn axis in a variety of ways, some of them are bundles of parallel CNTs and others are poorly aligned or coiled (b). It is desirable to have an aligned network of bundles of CNTs in the web structure as this is expected to improve the properties of CNT yarns and derived products. In order to improve the web structure, a web needs to be kept stable during the spinning process. With a CNT forest of given parameters (forest density, CNT length and diameter), web stability is improved by constant length of the web and uniform tension. The geometrical and kinetic model of the web can be obtained via the governing equation of the web formation with the following parameters: the tension Tw, the torque Ω caused by twisting and the combination force fs of the CNT interaction, CNT–wafer interaction and the geometry of web (see a). This will be represented in the future publication.After being pulled out from a wafer, the web was stabilized via the regulation of the tension and torque of a tensioning rod system described in the next section (a). Since the web structure is only weakly bonded, this stabilization and isolation from the later spinning stage was important for allowing further treatments to be performed on the web (for example heat treatment) during the spinning process.In spite of the uniform tension, there are still random bundles as well as parallel ones in the web. As yarn passes through a capstan effect rod system (CERS) (), the increased tension extends and aligns the bundles. Provided that the breaking extension of the yarn was not reached, some strain can occur. The change in tension as yarn passes around a rod (described in ) and the corresponding radial pressure on the yarn are given respectively as follows where μ, Tw, Tl, θ1 and N1 are the friction coefficient of yarn on the rods, the web (incoming) tension, outgoing tension, contact angle and radial pressure of yarn on the rod 1, respectively. The change in tension at the last rod of the CERS is given bywhere To is the tension of CNT yarn rolled up on the bobbin; n is the number of rods and θi is the contact angle between yarn and ith rod.Although the yarn was subjected to gradually increasing tension (Tw |
< |
T1 |
< |
T2 |
< ⋯ < |
To) through the CERS (Eq. ), the alignment and slip effects are observed to occur principally at the first rods, where CNT sliver (defined as bundles compacted of low density and without twist) is running through the CERS. The radial pressure imposed on the yarn at the rods, predominantly the later ones causes an increased CNT interaction in bundles and yarn. It was assumed that this process consists of two effects: (i) alignment of nanotubes in the fibres during the initial tensioning and (ii) the condensing or closer packing of the nanotubes together. The former effect increases the contact length between bundles and the latter one decreases the distance between tubes. Both effects are important to increase the tensile properties of the yarn, as the more aligned the fibre bundles, the higher will be the yarn compaction. The net effect being an increase in the tensile strength of yarns produced using this process compared to the process described in Section As indicated earlier, twist is imparted to reduce the inter-CNT distance and then increase the van der Waals interaction force to produce the required inter-bundle lateral cohesion of the yarn. In the modified spinning process, the total twist reaches from the final rod of the CERS to the bobbin (). A lower twist can be imposed on the yarn when running through the CERS by the adjustment of radial pressure of the yarn on rods via the contact angles (Eq. ). This technique allows the influence of twist to occur after the CNT sliver has been stretched, as well as protecting the yarn from exceeding its breaking force.Results show that the tenacity of a CNT yarn depends on stress transfer in shear mode between neighbouring CNT bundles. To achieve a higher tensile strength, the CNTs and bundles need not only to have a minimum length, but also they need to share maximum contact area.It should be noted that the tenacity of a yarn depends not only on the spinning process but also on the quality and characteristics of the drawable forest, for example the CNT diameter and length and the density of the CNT forest. In fact, whilst the longer carbon nanotubes will increase inter-tube contact areas and therefore yield higher tensile strength, decreasing the CNT diameter may also increase the yarn tensile strength Although the optimization of structural and engineering parameters of the modified process has still to be fully resolved, a significant improvement in the quality of CNT yarn has been obtained, especially enhancements in load transfer (the tensile strength) and the electrical conductivity which will be discussed in a later publication.In this work, the experimental plan involved using MWCNT drawable forests grown on silicon wafers using Chemical Vapour Deposition (CVD). The CNT’s length and outer diameter are approximately 300–400 μm and 7.5–8.5 nm, respectively. In order to remove the variability of the specification of the CNT forests from this investigation, the comparative results have been developed using MWCNT forest from the same wafer.The modified system for spinning continuous twisted CNT yarn allows significant improvement in the alignment of bundles in the yarn. b and c shows a typical segment of web and sliver. The micrographs indicate that the bundles in web and sliver structure obtained by the modified process are parallel and highly aligned in the direction of the long axis when compared with the web in a produced using the previous process as shown in The scanning electron microscopy (SEM) images also depict the structural difference of sliver located between rod 1 and rod 2 of the CERS using the modified process without heat treatment (Although the significance of the effect of CERS on the surface of CNT yarns has not yet been ascertained, the micrographs () show the surface layer of MWCNT yarns spun using the modified process and twisted 15,000 turns per meter (TPM) with a heat treatment of 200 °C and 400 °C on the web.Twisting can enhance stress transfer between CNT bundles under tensile forces is a SEM image of a micro-snarling on a CNT yarn with 17,000 TPM. The snarling occurs due to the imbalance between torque (caused by twisting) and the yarn tension.In order to investigate the tensile strength of CNT spun yarn, four drawable wafers of the same batch have been used to spin CNT yarns using the former process described in Section and the modified process. All are spun using the same spinning parameters. Twist factor of 15,000 TPM and the take-up speed of 6 m/h were used. The diameters of yarn samples were determined before testing their breaking force and extension on the tensile tester. The gauge length was 10 mm. The results of 12 typical samples shown in depict that the tensile strength of the yarn from the new approach has increased significantly when compared with tensile results using the former process. The tensile strength and strain data covers a range from 970 MPa to 1.4 GPa and 5–8%, respectively. The median tensile strength of 12 typical samples approaches 1.2 GPa. The stress–strain curves also depicted that the new process produces yarn with reduced breaking elongation and increased Young Modulus. From the work reported earlier The improvement in the yarn structure was obtained by stretching CNT bundles and enabling more effective yarn compaction. This densification treatment was not only from the twist but also by the radial pressure through the CERS.The separation of the different stages of distinctive mechanical and structural features of web and yarn allows for different treatments to be performed at individual stages of the spinning process. Several treatments have been carried out such as, heat treatment on the web, multi-web spinning from multi-wafers (), and yarn annealing. Some of these have been represented in the next sections. CNT yarn has been spun from two-webs split on one single wafer (Tran and Smith, 2008). The convergent point is located at the first rod of the CERS. Although it depends on many factors (the distance between the strips and position of the convergent point), the results of many different test samples on different wafers showed that the tensile strength was improved dramatically in comparison with one strip of web. It has been shown that the van der Waals force causes the inter-tubular CNT interactions related to the bundle and web formation and yarn compaction and that the magnitude of the force depends on thermal load and temperature After extraction from the forest, the CNT web goes through a mini furnace whose temperature is adjustable from 200 °C to 600 °C. shown that the temperature affects the mechanical properties of yarn via its effect on the van der Waals interaction between CNTs. Although the stretch of bundles and alignment level of CNTs depends on other factors such as tension, take-up speed and torque on the web and yarn, a reduction of inter-CNT interaction allows bundles to align more easily under the tension (from the CERS) and then the yarn is more compact under radial pressures. In , the vertical bars describe the variation of the tensile strength of yarn samples at the different temperatures of heat treatment on the web. The dotted line represents the average values of tensile strength of six groups of yarn samples. Although further investigations are required with an optimisation of the CERS, the preliminary data has shown that the tensile strength reaches the highest value, at a web heat treatment temperature of 200 °C and then it tends to decrease as the temperature is further increased.It is important to appreciate that the process of web heat treatment is carried out to improve the drafting properties of the CNT web. Web heat treatment can, but is not directly related to, strength improvements gained through the yarn annealing process. Yarn annealing has the objective of increasing the compaction of CNT yarn mentioned in Section , but not detailed by the authors as an objective of this work.A modified process for dry spinning a web of CNTs into a yarn has been evaluated. The spinning system introduced a zone that allows some controlled tensioning of the CNT web and this significantly improved the alignment of the CNT bundles in the web/sliver and was transformed into a highly compacted yarn by twisting. This method led to a significant improvement of mechanical properties of yarn with the tensile strengths about double that of earlier studies. The present modified process also allows different treatments in the individual zones, during the spinning, and the effect of heating of the web prior to twisting was evaluated and showed an optimum effect on the tensile strength at 200 °C. These preliminary results demonstrate that further refinement of this approach may contribute to the clarification of how the specific properties of drawable CNTs translate to the bulk properties of manufactured macroscopic structures. The optimization of the process parameters requires further investigation. The partitioning of the spinning system into separate zones has given new insights into the mechanisms involved and may allow further development of other types of CNT yarns, such as CNT composite yarns.Degradable, injectable poly(N-isopropylacrylamide)-based hydrogels with low gelation concentrations for protein delivery application► A biodegradable and injectable hydrogel was synthesized. ► The gelation characteristics relevant to physiological applications were studied. ► The hydrogel had a comparatively lower polymer concentration. ► The activity of the proteins was maintained after release.CMC-g-PNIPAAm copolymers were developed by decorating the backbone of carboxymethylcellulose (CMC) with linear chains of poly(N-isopropylacrylamide) (PNIPAAm), with the ultimate aim of synthesizing a biodegradable and injectable hydrogel that also possesses a low gelation concentration. Their aqueous solutions were found to undergo a reversible subphysiological phase transition at the concentration of 2 wt%. The value is much lower than that reported for many PNIPAAm-based copolymers. The phase transition behavior, gelation time, injectability, viscosity, swelling, degradation and cytocompatibility were explored. The model drug lysozyme was easily incorporated into the hydrogel by mixing with the gel precursors prior to heating. In vitro release of lysozyme from the in situ forming hydrogel was studied. Secondary and tertiary structure analysis and biological assays of the released protein showed that encapsulation and release did not affect the protein conformation and functionality. These results indicate that this biocompatible and injectable hydrogel system may be useful as a potential vehicle for therapeutic proteins for sustained release applications.Today therapeutic protein and peptide drugs are an important source of medicine. These drugs are marketed almost exclusively for treatment of illnesses, including some diseases which were previously incurable Hydrogels are physically or chemically cross-linked natural or synthetic polymers. Hydrogel formulations provide many advantages for the encapsulation and controlled delivery of protein and peptide drugs Injectable hydrogels can be formed by ionic cross-linking Another disadvantage limiting the application of PNIPAAm as injectable hydrogel is its high critical gelation concentration at which gelation occurs (15–20 wt% or above) Our objective in this work was to develop a biodegradable PNIPAAm-based hydrogel that also possesses a low gelation concentration. To accomplish this objective, a graft copolymer based on a carboxymethylcellulose (CMC) backbone bearing thermo-sensitive PNIPAAm side chains was prepared. Graft copolymers of CMC with PNIPAAm have already been synthesized before The sodium salt of carboxymethylcellulose (CMC, Mw 250,000, carboxymethylation degree of 0.91 per disaccharide unit) was supplied by Wealthy Chemical Industry Co., Ltd. (Suzhou, China). N-isopropylacrylamide (NIPAAm) provided by ACROS (New Jersey, USA) was recrystallized from a toluene/hexane mixture just before use. Lysozyme used for release studies was provided from Bio Basic Inc. (Toronto, Canada). N,N′-Azobisisobutyronitrile (AIBN) was recrystallized from methanol before use. 2-Aminoethanethiol hydrochloride (AET·HCl), 1-(3-(dimethylamino)propyl)-3-ethyl-carbodiimide hydrochloride (EDC) and other chemicals were used as received. All solutions were prepared with de-ionized water.Amino-terminated PNIPAAm was prepared by radical polymerization. Briefly, NIPAAm (17.70 mmol) and AET·HCl (4.40 mmol) were dissolved in 20 ml of methanol. The solution was deaerated with nitrogen bubbling for 30 min prior to the addition of AIBN (0.24 mmol). After inducing polymerization under stirring in a water bath at 65 °C for 6 h, the crude solid was precipitated in an excess of diethyl ether, dialyzed for 3 days, and lyophilized.The CMC-g-PNIPAAm copolymers were prepared according to a previous reported procedure slightly modified .) The pH of the mixture was adjusted to 5.5. After 24 h, the reaction mixture was dialyzed against water and lyophilized. Then the dried products were extracted in a Soxhlet extractor with methanol for 72 h. The grafting ratio and grafting efficiency were estimated by Eqs. Grafting efficiency(%)=WCP−WCMCWPNIPAAm-NH2×100where WCP, WCMC and WPNIPAAm-NH2 are the weights of CMC-g-PNIPAAm copolymer, CMC and amino-terminated PNIPAAm, respectively.The graft copolymers has been determined by Fourier transform infrared (FTIR) spectroscopy, which was carried out with KBr disk using a Nicolet NEXUS 670 FTIR spectrometer. The polymer after degradation in phosphate buffered saline (PBS, pH 7.4, 0.1 M) for 15 days at 37 °C was also characterized.Polymer molecular weights were characterized by size exclusion chromatography (SEC) using a Waters Alliance GPCV2000 (USA) equipped with three detectors on line: a differential refractometer, a viscometric detector and a multiangle laser light scattering (MALLS) detector from Wyatt (USA). Aqueous sodium nitrate (0.1 M) at ambient temperature with nominal flow rate of 0.5 ml/min was used as the mobile phase. Weight-average molar mass (Mw) was obtained as characteristic of the polymers. Molecular weight of the copolymer after degradation in PBS for 15 days at 37 °C was also measured.Polymer gelation concentration was measured by varying CMC-g-PNIPAAm copolymer concentrations from 0.5 to 3 wt%. Appropriate amounts of polymer and solvent were taken in glass tubes. The tubes were initially kept at 25 °C for 30 min, and then quickly transferred into a water bath at 37 °C. The gelation was visually measured by observing cessation of the liquid flow inside the tube when it was tilted.Hydrogel injectability was evaluated by its ability to be injected through a 26-gauge needle at the concentration of 2 wt%. This was based on the consideration for future animal injection.Hydrogel gelation time was measured using a UV/vis spectrophotometer (Lambda 35, Perkin-Elmer, America), equipped with a circulating water bath. The system was pre-warmed to 37 °C before testing. Then the CMC-g-PNIPAAm copolymer solution (2 wt%) at room temperature was quickly put in a 5 ml of cuvette. The time needed for the solution to become completely opaque (transmittance% ≤0.1) was considered as a gelation time.The viscosities of the CMC-g-PNIPAAm copolymer solutions (2 wt%) were determined using a rotational viscometer (NDJ-1, Shanghai Shangtian Precision Instrument Co., Ltd., China) at 25 °C (rotor No.: 3, rotation: 60 rpm).Phase transition properties of the aqueous solutions of amino-terminated PNIPAAm and CMC-g-PNIPAAm copolymers (2 wt%) were monitored as a function of temperature using a UV/vis spectrophotometer, equipped with a temperature control system. The samples were heated at a rate of 0.2 °C/min from 25 to 35 °C. Transmittance measurements were recorded every 1 min. The phase transition temperature was determined as the point at which 50% of the overall transmittance change had taken place.Steady state fluorescence measurements of samples containing pyrene were performed with a fluorescence spectrophotometer (LS 55, Perkin-Elmer, America) using 1 cm path length quartz cuvettes. For measuring pyrene fluorescence, samples were excited at 342 nm. Excitation and emission slits with a nominal bandpass of 5 nm were used for all measurements. The ratio of the first (373 nm) and third (384 nm) vibronic peak intensities (I1/I3) was calculated from pyrene emission spectra Scanning electron microscopy (SEM) was employed to investigate the porous structure of the hydrogels. 1 ml of CMC-g-PNIPAAm (2 wt%) copolymer solution in PBS was placed in a water bath at 37 °C to allow gelation. The hydrogel was quickly frozen in liquid nitrogen and lyophilized using a LABCONCO Freeze Dryer. The dried samples were sputtered with gold and the microstructures and morphologies of the cross-sections of the hydrogels were then examined by a JSM-5600LV SEM (Japan).The equilibrium swelling ratio (ESR) of the hydrogels was measured in PBS at 37 °C. 1 ml of CMC-g-PNIPAAm (2 wt%) copolymer solution in PBS was placed in a 37 °C water bath to allow gelation. Then extra 20 ml of PBS was added and the hydrogels were kept until they swelled to equilibrium. After excess surface water was removed, the fully swollen samples were weighed at 37 °C. All experiments were done in triplicate. The ESR can be calculated by Eq. where We and Wd are the weights of equilibrium hydrated gel and dry gel, respectively.1 ml of CMC-g-PNIPAAm (2 wt%) copolymer solution in PBS was placed in a test tube on water bath at 37 °C and allow gelation. Then extra 20 ml of PBS was added. Different tubes were prepared for each hydrogel formulation (). At regular intervals, the hydrogels were lyophilized and weighed. Each sample was evaluated in triplicate. The weight loss ratio (WLR) was calculated by Eq. where W0 and Wt are the weights of hydrogels before and after degradation, respectively.Cytotoxicity evaluation of hydrogels and their degradation products were performed by 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) assay using HEK 293T cells, which were grown in DMEM with 10% of fetal bovine serum (FBS) and were cultured in a humidified atmosphere containing 5% CO2 at 37 °C. CMC-g-PNIPAAm copolymer solutions were sterilized and then injected into a 24-well plate and incubated at 37 °C to form hydrogels. Then 200 μl of culture medium containing 1 × 105 cells was added into each well. This was left at 37 °C for 24 h. Then the medium was replaced by fresh DMEM. After 48 h, the cell cultures were washed with PBS solution and MTT assay was conducted. Untreated cells were taken as a control with 100% viability. The cell cytotoxicity of hydrogels was defined as the relative viability (%) which correlates with the amounts of liable cells compared with the cell control. Reported values were mean of three replicates.To study the release behavior of the protein, CMC-g-PNIPAAm hydrogels loaded with lysozyme were prepared. Lysozyme (3 mg) was mixed with 1 ml of CMC-g-PNIPAAm (2 wt%) copolymer solution to form a homogenous solution, and then the mixture was placed in a 37 °C water bath to allow gelation. Extra 20 ml of PBS was added. At fixed time intervals, 5 ml of the release medium was withdrawn and replaced immediately with fresh buffer solution. The lysozyme content was determined using a UV/vis spectroscopy at 281 nm. All experiments were done in triplicate.Circular dichroism (CD) spectra were recorded between 200 and 260 nm at room temperature on an Olis DSM 1000 CD spectrometer (America) for lysozyme freshly prepared and released from the hydrogels after 10 days. All measurements were carried out in 1-mm quartz cuvettes in PBS at the concentration of 0.1 mg/ml. Spectra resulted from accumulation of 2 scans were averaged. Blank spectra of the buffer without lysozyme, obtained under identical conditions, were subtracted.Fluorescence emission of native and released lysozyme was measured using a fluorescence spectrophotometer at room temperature. Emission spectra were recorded between 320 and 400 nm on excitation at 300 nm. Both the excitation and emission slit widths were set at 10.0 nm.The biological activity of released lysozyme was determined using Micrococcus lysodeikticus as substrate The synthesis of functional PNIPAAm is usually achieved by introducing an efficient chain transfer agent in the polymerization medium. In the specific case of amino-terminated PNIPAAm, AET·HCl is undoubtedly the best candidate since thiols are widely used and well-known chain transfer agents . Through EDC, a presumably comb-type graft copolymer was synthesized by grafting linear chains of PNIPAAm along the length of a CMC backbone, which was supposed to exhibit a sharp phase transition below physiological temperatures at a low gelation concentration. Grafting ratio and grafting efficiency were determined, and the results are shown in . Take CP1 for example, the weight of extracted CMC-g-PNIPAAm copolymer, CMC and amino-terminated PNIPAAm was 127.7, 50.0, and 150.0 mg, respectively. The grafting ratio and grafting efficiency were calculated by Eqs. , and were 155.4% and 51.8%, respectively. Grafting ratio represents the number of amino-terminated PNIPAAm chains grafted onto a CMC backbone molecule, while grafting efficiency gives the percentage of grafted amino-terminated PNIPAAm based on the initial amount of amino-terminated PNIPAAm used in the reaction. With the increasing content of amino-terminated PNIPAAm, grafting ratio increases and grafting efficiency decreases.FTIR measurements were determined to confirm the preparation of amino-terminated PNIPAAm and CMC-g-PNIPAAm copolymer, as shown in a) has two strong characteristic peaks at 1620 and 1410 cm−1, attributed to Cb). In addition, two bands at 3436 and 3310 cm−1 for the primary amine (NH2) are observed, indicating that −NH2 was introduced to PNIPAAm polymers. Other characteristic peaks in the amino-terminated PNIPAAm spectrum (b) can be ascribed to 1650 cm−1 for Amide I, 1542 cm−1 for Amide II, 1387 cm−1 for −CH3 stretching vibration in CH(CH3)2, and 634 cm−1 for C–S stretching vibration, respectively. For CMC-g-PNIPAAm copolymers (c), the characteristic peaks at 1650, 1542, 1387 cm−1 and 634 cm−1 are observed for positive identification of the introduction of amino-terminated PNIPAAm. Meanwhile, there exists the characteristic peak of CMC at 1062 cm−1 for the C–O(H) stretching vibration, indicating that amino-terminated PNIPAAm has been successfully introduced onto CMC.The capacity to form a gel and provide a structural matrix under practical conditions is important for injectable hydrogels. The capacity is affected by several factors, one of which is gelation concentration. Usually gelation concentration and the stiffness of the resulting hydrogel run parallel. In the study, CMC-g-PNIPAAm copolymers were dissolved at different concentrations to determine polymer gelation concentration at 37 °C. Aqueous solutions of varying concentrations were transparent at 25 °C. At an elevated temperature (37 °C), the transparent solutions formed opaque solutions to rigid hydrogels depending on their concentrations. When the concentration is as low as 0.5 wt%, the copolymer concentration is below the critical gelation concentration, and a gel cannot form. However, soft weak gels were formulated at higher concentrations (1–1.5 wt% depending on the grafting ratio) which may be near the limit of gelation. Upon increasing the concentration up to above 2 wt%, the copolymers underwent a phase transition from a clear solution to a rigid hydrogel. Therefore, CMC-g-PNIPAAm copolymers were prepared at the concentration of 2 wt% in this study. The value is much lower than that reported for many PNIPAAm-based copolymers Hydrogel injectability was evaluated by its ability to be injected through a 26-gauge needle, as demonstrated in . Phase transition first occurred in the outer layer of the solution drops when CMC-g-PNIPAAm copolymers were injected directly into PBS solutions at 37 °C. After about 40 s, phase transition occurred completely for the whole drops and rigid hydrogels formed. This was based on the consideration for future animal injection.Gelation time is crucial because slow gelation would cause delocalized gel formation due to diffusion of the gel precursors away from the injection site and lead to drug overdosage, while fast gelation would clog the needle before injection . It is noted that gelation occurs in a reasonable time (61–86 s) for further clinical application. There is a hysteresis was observed between the determined gelation time and actual gelation time shown in , which is because of the effect of the detecting system. The gelation time decreases as the grafting ratio increasing from CP1 to CP4, which might be attributable to the presence of stronger hydrophobic interactions among PNIPAAm chains and the complicated entanglement of macromolecular chains.Ideal injectable hydrogels should behave as a low viscosity fluid at room temperature to be injected into body via syringe. The viscosity of CMC-g-PNIPAAm solutions (2 wt%) measured by a rotational viscometer at 25 °C is listed in , we can see that the viscosity of CMC-g-PNIPAAm solutions increases with the grafting ratio increasing. Viscosity of a polymer solution depends on concentration and size (i.e. molecular weight) of the dissolved polymer. For a fixed polymer concentration, polymer chain interactions such as Van der Waals attractions and entanglements increase with increasing molecular weights. These interactions tend to fix the individual chains more strongly in position and resist deformations at higher stress. Accordingly, among the hydrogels tested in the viscosity examination, CP4 is observed to exhibit the highest viscosity. However, all CMC-g-PNIPAAm solutions have low viscosities, which is an appropriate feature for the injectable applications.The phase transition behavior of the two polymers, amino-terminated PNIPAAm and CMC-g-PNIPAAm, in PBS is presented as a function of temperature in . Both amino-terminated PNIPAAm and CMC-g-PNIPAAm with a concentration of 2 wt% are soluble in PBS at room temperature. During the heating cycle, amino-terminated PNIPAAm transforms to a milky white solution, whereas CMC-g-PNIPAAm turns an opaque gel. All samples appear to undergo rapid phase transition at their LCST, as indicated by the sudden decrease in transmittance. The subsequent cooling cycle results in gel–sol transition and fully reversible gel melting. For all samples, the gel–sol transition temperatures are lower than the sol–gel transition temperatures, indicating some hysteresis between the gelation and gel flowing temperatures. The observed hysteresis loop, typical for polymeric network, is a kinetically controlled phenomenon reflecting the resistance to disintegration of entangled polymer chains forming the hydrogel network for CMC-g-PNIPAAm solutions. All CMC-g-PNIPAAm solutions show a sol–gel transition at 33 °C during the heating cycle, independent of their grafting ratio. Similarly, during the subsequent cooling cycle, gel–sol transition at 31 °C is observed for all CMC-g-PNIPAAm solutions.At the polymer concentration of 2 wt%, amino-terminated PNIPAAm solution does not form a rigid hydrogel at 37 °C but gives white precipitates. This is because that the concentration is not high enough to promote sufficient hydrophobic interaction between polymers to create physical cross-links to form a gel. Nonetheless, 2 wt% of CMC-g-PNIPAAm solution gels successfully at 37 °C, indicating that CMC plays an important role in stabilizing the integrity of three dimension structure of polymer hydrogels. This is because that the number of entanglements increases due to complications of the macromolecules as a result of motion of segments.Pyrene is a strongly fluorescent molecule. The I1/I3 ratio of the fluorescence spectra of pyrene is used to obtain information on the phase transition behavior of various PNIPAAm-based copolymers, and it reflects the polarity of the microenvironments where the probe is hosted The change in the ratio of I1/I3 in pyrene emission spectra of amino-terminated PNIPAAm and CMC-g-PNIPAAm solutions with temperature is shown in . The figure shows that I1/I3 ratio decreases with increasing temperature depending upon the increasing hydrophobicity of the medium by the phase transition of the thermally sensitive polymer. An inflection point in the spectra of amino-terminated PNIPAAm and CMC-g-PNIPAAm solutions is observed at 33 °C, which is close to LCST determined by UV/vis spectroscopy, as shown in . These behaviors exhibit an increase of the hydrophobic association of PNIPAAm segments and a decrease in polarity of microenvironment around pyrene when the temperature was higher than 33 °C.Samples were prepared at 37 °C, rapidly frozen by immersion in liquid nitrogen and lyophilized to generate a dry sample while preserving the hydrated pore structures, thereby allowing image collection via SEM, as shown in . All samples exhibit highly porous sponge-like structures and all the porous have extremely thin walls. In addition, all the pores in the hydrogels appear very regular and are oriented along a common direction such that they look like an array of high porous filaments. The pores are small, probably because that chain entanglement further enhances the formation of the hydrogel networks. From the figure, we can see that the grafting effect on the porous structure is not so significant.Previous studies have shown that swelling behavior of the hydrogel-based drug delivery systems has a great effect on their drug release behaviors . All the samples swelled rapidly and reached equilibrium within 2 h. Grafting ratio of CMC-g-PNIPAAm significantly influences the ESR of CMC-g-PNIPAAm hydrogels, i.e. the ESR of CP1 is 18.2 g/g, while which is 9.6 g/g for CP4. CP1 reveals less grafting PNIPAAm and loose structure, consequently increasing the exposure of hydrophilic polymer chains to water molecules at 37 °C, and significantly leading to a greater swelling ratio. In addition, because of its low grafting ratio, there are many unreacted hydrophilic groups in CMC such as carboxyl and hydroxyl groups, which also contribute to the greater hydration capacity in contrast to other CMC-g-PNIPAAm hydrogels. CP4 has clusters of physically associated PNIPAAm chains through interactions and entanglements, thus preventing water absorption into CMC.The development of biodegradable drug delivery hydrogels is nowadays of increasing interest and deserves the attention of both academic and medical research. In this study, in vitro degradation of CMC-g-PNIPAAm hydrogels was monitored by examining the weight loss of hydrogels with incubation time in PBS at 37 °C, as presented in . All hydrogels demonstrate progressive mass loss during the degradation. Under current degradation conditions (without enzyme), hydrolysis of amide bonds dominates degradation. This can be testified by the FTIR spectra of the degraded polymers (, in which the band at 1387 cm−1 corresponding to the −CH3 stretching vibration in CH(CH3)2 of PNIPAAm disappears. In addition, all the four polymers exhibit an obvious increase at about 1100 cm−1 and 920 cm−1 which are attributed to the C–O stretching vibration of CMC. The results clearly show that degradation had occurred. The degradation process can be described as follows: initially, the hydrogels absorb water, a small number of amide bonds are broken and the lattice size of the networks enlarges, resulting in large swelling stress in the hydrogels. Then, when the strength of the swelling stress becomes higher than that of hydrophobic interactions, which hold together the thermo-sensitive chains within hydrophobic domains, polymer dissociation occurs and the whole network will be disjointed, eventually leading to disintegration of the hydrogels , CP4 hydrogel is much more slowly degraded than other hydrogels, whereas CP1 hydrogel shows significant weight loss and it completely collapsed after 7 days of incubation in PBS at 37 °C. For CP4 hydrogel, the higher grafting ratio of polymers results in more robust and dense hydrogel network and lower hydrolysis rate.To further confirm degradability of the hydrogels, molecular weights of the hydrogels before and after degradation were measured by SEC, as shown in , we can see that all hydrogels display reduced molecular weight after incubation in PBS for 15 days, which demonstrates that the degradability of the hydrogels.Critical for a potential clinical use of the developed biomaterials is their biocompatibility. Therefore, the cytotoxicity of CMC-g-PNIPAAm and their degradation products were evaluated by MTT assay. The viability of HEK 293T cells at 48 h in the presence of CMC-g-PNIPAAm at different concentrations is reported in . The results demonstrate that, CMC-g-PNIPAAm hydrogels and their degradation products do not significantly affect the mean viability of HEK 293T cells at concentrations of 0.1, 0.5, 1.0, and 2.0 mg/ml in relation to the control. Moreover, the cell viability is not significantly different among all concentrations. It was reported that CMC is a biodegradable polymer and has no toxicity Compared with other hydrogel-based drug delivery systems, injectable hydrogel can incorporate drugs by simple pre-mixing. In this study, the model drug lysozyme was easily incorporated into the hydrogels by mixing with the gel precursors to form a homogenous solution prior to heating. The aim here was to investigate whether the synthesized CMC-g-PNIPAAm hydrogels could control the release rate of lysozyme.The release behavior of lysozyme in PBS at 37 °C is depicted in . It is apparent that there is a burst release of lysozyme within the first day. This is most likely because that lysozyme molecules which are at or near the solvent–hydrogel interface escape rapidly into the release medium. In addition, lysozyme occupied the water channels of the hydrogel from which it was released by diffusion along a concentration gradient inside and outside of the hydrogel. In the initial stage of release, the concentration gradient is great and the dynamic exchange of water and drug is fast, resulting in a rapid initial release of lysozyme. Subsequently, a steady release of lysozyme was observed, which may be dominated by the diffusion of the drug together with the degradation of the hydrogels. In the last stage of release, a decrease in the drug release rate is attributable to the electrostatic interaction between the drug and the polymer. Under physiological conditions, lysozyme is positively charged since its isoelectric point (pI) is about 11 From the drug release profiles, it can be concluded that the drug release kinetics can be tailored by changing the grafting ratio of the CMC-g-PNIPAAm hydrogel. Hydrogels with a lower grafting ratio display a lower cross-linking density, which leads to an enlarged mesh size, thus enabling the entrapped drug to diffuse out more readily. Nevertheless, with a higher grafting ratio, hydrogels are denser and may entrap lysozyme within the matrix for a longer period.Not all lysozyme is completely released. The remaining part may be entrapped in the hydrogels as the positive charge of lysozyme can enter into electrostatic interaction with the negative charge of CMC under physiological conditions. In addition, protein release rarely reaches 100%, which is most likely caused by physical entrapment of the molecules in highly entangled networks of the hydrogel, allowing less free motion of the diffusant It is important that a protein release system not only delivers the proteins at a controlled release rate but also maintains the activity of the proteins. Protein aggregation events and protein–hydrogel interactions resulting in protein denaturation may occur during formulation, storage, and release periods CD was used to examine the secondary structural characteristic of lysozyme. shows the CD spectra of native and released lysozyme. As seen in , the CD spectra of the released lysozyme closely resemble those of lysozyme in the native state, which suggest that the secondary structural of lysozyme has little change during the release process.It has been reported that fluorescence emission is much more sensitive to changes in the environment of the tryptophan, thus fluorescence spectroscopy is an excellent method to investigate tertiary structure changes of proteins shows the fluorescence emission spectra recorded with excitation at 300 nm. It can be seen from that the emission spectra of all released lysozyme are similar to that of the native lysozyme with respect to the emission maximum, indicating that the tertiary structure of lysozyme is not affected during the release process.Lysozyme functionality after release was conducted by a classical photometric assay, which detects changes in the turbidity of a bacterial suspension, M. lysodeikticus, caused by the enzymatic activity of lysozyme. Released lysozyme from CP1, CP2, CP3, and CP4 yields an activity of 173, 169, 162 and 158 substrate units/mg of lysozyme in the reaction solutions, respectively. These values are essentially identical to that measured with freshly prepared lysozyme, which is 179 substrate units/mg of lysozyme. Therefore, lysozyme functionality preserves upon incorporation and release through the hydrogels.In attempts to develop an injectable hydrogel with a low gelation concentration to deliver lysozyme, we decorated the backbone of carboxymethylcellulose (CMC) with thermo-responsive PNIPAAm. Grafting of amino-terminated PNIPAAm onto the CMC backbone was achieved by a coupling reaction. Polymer grafting was confirmed using FTIR and did not significantly alter the LCST of the copolymers. The gelation concentration was determined to be 2 wt%. The hydrogel displayed excellent compatibility with HEK 293T cells. The released lysozyme was tested, and the results showed that the encapsulation and release did not affect the secondary and tertiary structure of lysozyme, nor was their functionality diminished. Compared to many in situ forming hydrogels of PNIPAAm, the CMC-g-PNIPAAm hydrogel had a comparatively lower polymer concentration for in situ gelation. These attractive properties demonstrated that the injectable hydrogel system may be useful as a carrier for therapeutic protein for sustained release applications.Novel approach to additively manufacture high-strength Al alloys by laser powder bed fusion through addition of hybrid grain refinersGrain refinement is effective in restraining hot tearing, reducing anisotropy, eliminating defects, improving processability, and enhancing the mechanical properties of high-strength aluminum components additively manufactured by laser powder bed fusion (LPBF). However, achieving the desired strength and ductility in LPBF-fabricated high-strength aluminum alloys post grain refinement is a predominant challenge. We have therefore designed and developed a novel hybrid grain refiner (solute/ceramic nanoparticles) which can effectively refine grains and enhance the mechanical properties of LPBF-fabricated high-strength aluminum alloys. Adding a Ti/TiN hybrid grain refiner to the LPBF-fabricated 7050 alloy can produce ultrafine grains with an average size of 775 nm, resulting in an ultimate tensile strength and ductility of up to 408–618 MPa and 13.2–8.8%, respectively. These tensile properties are comparable to those of conventional wrought 7XXX alloys. During LPBF processing, the hybrid grain refiner exhibited interesting synergistic grain refinements and strengthening mechanisms between the solute and the ceramic nanoparticles. During solidification, not only in-situ particles formed by the chemical reaction of the solute in liquid Al and the externally added ceramic nanoparticles can act as the nuclei of α-Al respectively, but also solute can inhibit the agglomeration of ceramic nanoparticles to promote their nucleation efficiency. Moreover, the strength can be further improved by doping the solute atoms at the ceramic nanoparticle/Al interface. The improvement in elongation benefited from the uniform dispersion of the various particles.The laser powder bed fusion (LPBF) additive manufacturing of high-strength aluminum alloys has drawn considerable attention Generally, grain refinement of LPBF-fabricated high-strength Al alloys can be achieved through various approaches LPBF feedstock powder can be modified by adding other alloying element powders, such as Zr, Sc, Si, Fe, or Ti To overcome the above-mentioned limitations and achieve an ultrafine microstructure with exceptional mechanical properties, a novel hybrid grain refiner comprising solutes and ceramic nanoparticles for LPBF-fabricated high-strength Al alloys was developed. The solutes promote the formation of the constitutional supercooling zone and the nanoparticles act as nuclei. The 7050 alloy is a typical high-strength aluminum alloy widely used in the aerospace industry, and the Ti/TiN hybrid refiner has been used in the fabrication of cast aluminum alloys to successfully achieve grain refinement. This research investigated the role of a hybrid refiner in an LPBF-fabricated 7050 and optimization of the Ti/TiN combination was carried out. The tensile strength and ductility properties were measured, and the strengthening mechanisms are discussed. The outcomes of this study have potential applications in the design and fabrication of ultrafine microstructures during other rapid solidification processes, such as welding and laser surface modification.Ti/TiN hybrid refiner was prepared using ball milling (Fritsch Pulverisette-5). TiN nanoparticles (purity 99.9%) and Ti microparticles (purity 99.9%) were mixed for 30 min in a ball mill at a volume ratio of 1:1. The volume ratio of Ti and TiN was controlled using a measuring cylinder. A planetary mono mill at a speed of 300 rpm was used with a milling duration of 10 min followed by 10 min of rest. The mass ratio of the material to steel ball mass ratio is 1:10. The circulation of a short-time mill with resting can avoid the contamination of Fe from the steel balls . Subsequently, 7050 powders with various contents of Ti/TiN(2%, 4%, and 6%, mass fraction), 7050/TiN(0.18%, 0.36%, and 0.54% of TiN, mass fraction), and Ti (1.82%, 3.64%, and 5.46%, mass fraction), were mixed using a TURBULA®Shaker-Mixer machine for 2 h. These well-mixed powders were selected for fabrication using LPBF.According to ASTM B557M-10, dog-bone-shaped samples for the tensile testing (a and b) of 7050, 7050-TiN, 7050-Ti, and 7050-Ti/TiN alloys were designed and fabricated in a flowing argon environment by LPBF with an oxygen content of less than 0.3%. The building thickness of the sample was 2.5 mm with a margin of 1 mm for polishing. LPBF was conducted using a Concept Laser M2 instrument. The scanning strategy is illustrated in c. A “random island” strategy with a rotation of ~90° between successive layers was employed to decrease the accumulation of residual stress. The fabrication parameters were: laser power (P) of 210 W, powder layer thickness (t) of 30 µm, hatch spacing (h) of 50 µm, island size (w) of 1.5 mm, and laser scanning speed (V) of 115 mm/s.The size distribution of the powders was measured using a Malvern laser particle size analyzer. The relative densities of the LPBF-fabricated samples were measured through Archimedes method (ASTM B962). The microstructures and elements of the raw powders and LPBF-fabricated samples were examined using scanning electron microscopy (SEM, ZEISS Merlin) and transmission electron microscopy (TEM, FEI Talos F200X). The thin films from the LPBF-fabricated components for TEM analysis were prepared using a focused ion beam (FIB, FEI Helios Nanolab 600i). The phase constituents were analyzed using X-ray diffraction (XRD, Rigaku Smartlab) in the scanning range of 2 theta from 10° to 90°. The grain orientations, sizes, and boundary angles of the LPBF-fabricated samples were characterized using electron backscattered diffraction (EBSD) attached to the SEM. Tensile properties (stress–strain relations) were tested on an Instron 3382 machine with a 10 mm gauge extensometer. A constant strain rate of 0.2 mm/min was used for testing. Each tensile test was repeated three times to improve data reliability.To deeply understand the strengthening mechanisms of the Ti/TiN hybrid refiner in the LPBF-fabricated 7050 alloy, we calculated various parameters of Al/TiN interface binding before and after the doping of Ti atoms at the interface using first-principles calculations. The Vienna Ab-initio Simulation Package (VASP) was used in this calculation. The exchange–correlation functional with the generalized gradient approximation parameterized by local density approximation (LDA) was used. The truncated kinetic energy of the plane wave was 450 eV, and the convergence criterion of the total energy of the system was 10 × 10−6 eV. Monkhorst-Pack method was used to generate a special K-point grid with a center of 8 × 8 × 8 Γ. The conjugate gradient algorithm was used to generate the Hellmann–Feynman forces of all atoms in the supercell less than 0.01 eV/Å where ΔEif and ΔEfs denote the binding energies of the interface and free surface segregated by solute, respectively; EifTi and EfsTi denote the total energies of the ideal Al/TiN interface and free surface doped with Ti, respectively; and ETi represents the total energy of the Ti atom. The crystal information of experimentally determined Al and TiN is shown in , and the Ti-doped Al/TiN interface systems are shown in Nano-sized TiN particles were chosen as the gain refiners because of their high melting point (3223 K), excellent chemical stability (no reaction with liquid Al ) in Al, was selected as the solute to guarantee the grain refinement during solidification. The data used in the calculation of Q were obtained from Refs. b and c), and spherical gas-atomized 7050 powders (d). Particle size distribution parameters of 7050 powders were Dv(10), Dv(50), and Dv(90) at 10.9, 32.4, and 70.7 µm, respectively. A Ti/TiN hybrid refiner was prepared by ball milling Ti and TiN powders (see Methods). After milling, the TiN nanoparticles were assembled on the surface of the Ti microparticles (e). 7050/TiN hybrid particles were also prepared in the same manner (f) for the purpose of comparison with Ti/TiN. In e and f, sites with particle sizes of 30–60 nm (marked by arrows) were considered as TiN nanoparticles. The XRD results (g) show that there is no phase transition in these hybrid particles.The LPBF-fabricated 7050–0.18TiN, 7050–1.82Ti, 7050–2Ti/TiN (0.18, 1.82, and 2 represent the mass fractions), and 7050 samples were compared. The optical microscope images () show that the LPBF-fabricated 7050 and 7050–0.18TiN samples have cracks, columnar grains, and porosities along the building direction, whereas 7050–1.82Ti and 7050–2Ti/TiN are crack-free, with uniform and discreet porosities. The black spots in c and d are small pores. The relative densities of the 7050, 7050–0.18TiN, 7050–1.82Ti, 7050–2Ti/TiN were 98.5%, 98.9%, 99.6% and 99.7%, respectively. The EBSD characterizations of the four samples () indicate that cracks propagated through the intergranular regions in the LPBF-fabricated 7050 and 7050–0.18TiN samples. Significant columnar grains with sizes of 91.8 µm and 88 µm and high isotropy values of 3.473 and 2.245 for the (001)[001] orientation were observed for LPBF-fabricated 7050 and 7050–0.18TiN samples, respectively (a-d). Increasing the Ti content to 1.82% (LPBF-fabricated 7050–1.82Ti) generated fine equiaxed grains with a size of 2.3 µm and an isotropy value of 1.793 for the (101)[001] orientation. With a further increase in Ti/TiN to 2% (LPBF-fabricated 7050–2Ti/TiN), an ultrafine equiaxed Al grain structure with an average size of 775 nm and a low isotropy value of only 1.19 for the (001)[001] orientation were achieved () show the cracks and grain size details consistent with the EBSD results. Additionally, oxides were observed adjacent to the cracks in the LPBF-fabricated 7050 and 7050–0.18TiN samples, whereas in the LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN samples, Ti-containing particles were dispersed in the α-Al matrix.a and b) to determine the phase constituents in the LPBF-fabricated samples. Al, Al2CuMg and MgZn2 phases were identified in all samples (a), whereas Al3Ti phases appeared only in the LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN (b). Interestingly, no TiN peaks were observed in the LPBF-fabricated 7050–2Ti/TiN sample. This may be due to the low content of TiN and limited resolution ratio of the XRD. The TEM images show various particles within the α-Al matrix or on the grain boundaries (d illustrates that Particle 1, considered as a TiN nanoparticle due to the size of 50 nm and the elements consisting of Ti and N (), exists within an α-Al grain. The high-resolution TEM (HRTEM) images with the incident electron beam parallel to the [010] zone axes of both TiN and Al prove the existence of the TiN nanoparticle within the α-Al grain. The interfacial atomic coincidence demonstrates a coherent interface between TiN and Al (). This coherent binding at the TiN/Al interface implies strong interfacial binding ) has a diameter of hundreds of nanometers and a (110) lamellar spacing of 0.287 nm, as shown in the HRTEM image. These Al3Ti formed in situ should have an L12 structure under LPBF f shows the presence of the η phase (MgZn2) with a diameter of ~62 nm and a C15 structure segregates at the boundary Tensile tests for the LPBF-fabricated 7050, 7050-(0.18, 0.36, and 0.54)TiN, 7050-(1.82, 3.64, and 5.46)Ti, and 7050-(2, 4, and 6)Ti/TiN samples were performed at room temperature (25 ± 3 °C). Typical engineering stress–strain curves from the test results are shown in d). However, when the grain refiner content was further increased, the UTS decreased. This UTS reduction is related to the high content of particles in the matrix The SEM investigations of the fracture surfaces after tensile testing () revealed ductile and intragranular fractures for the LPBF-fabricated 7050–2Ti/TiN sample (a-c), while brittle fracture characteristics and intergranular cracks were observed in the LPBF-fabricated 7050 sample (d and e). The improvement in ductility of the LPBF-fabricated 7050–2Ti/TiN can be attributed to grain refinement and the existence of uniformly dispersed nanoparticles inside the α-Al matrix Considering that cracks in the LPBF-fabricated 7050 and 7050–0.82TiN samples determine the real tensile properties, quantitative comparisons were only conducted based on the UTS data from LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN. For these two samples, the strengthening mechanism can be described by the grain boundary strengthening (ΔσHall–Petch), Orowan loop strengthening (ΔσOrowan), and load-bearing (Δσload) mechanisms and is as follows.Numerous grain boundaries can act as obstacles for the movement of dislocations during deformation, hard nanoparticles can serve as nuclei, and pinning effects inhibit grain growth during the LPBF process. From the EBSD results, the average grain sizes of the LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN is 2.3 µm and 775 nm, respectively (). The ultimate tensile strength gained from the Hall–Petch strengthening (or grain–boundary strengthening) can be calculated as:where k = 0.06 MPa m1/2 is the strengthening effect coefficient for aluminum, and d is the average grain size in the LPBF-fabricated samples The contribution from the Orowan strengthening of nanoparticles can be estimated by where Gm is the shear modulus of the matrix (~25.5 GPa), b is the Burgers vector (~0.286 nm), dnano is the size of the nanoparticles (50 nm for TiN and 210 nm for Al3Ti), and V is the volume fraction (~0.9 vol% for TiN and ~3.5 vol% for Al3Ti). The densities of TiN, Al3Ti, and Al are 5.2, 3.3 and 2.7 g/cm3, respectively, and thus, the relative densities of TiN and Al3Ti compared to Al, respectively, are 1.9 and 1.2. In the LPBF-fabricated 7050–2Ti/TiN alloy, the mass fraction of TiN was 0.18%; therefore, the volume fraction was 0.18%/1.9 ≈ 0.9%. Assuming that the solid solution degree of Ti in the Al matrix was 0.2 wt% (according to the Al–Ti binary phase diagram), the mass fraction of Ti that participated in the formation of Al3Ti was 1.82–0.2% = 1.62%, and the mass fraction of Al3Ti was 4.36%. Sequentially, the volume fraction of Al3Ti was 4.36%/1.2 ≈ 3.5%. φ is a constant of order 2. In 7050–1.82Ti, the Orowan strengthening comes from Al3Ti, whereas in 7050–2Ti/TiN it comes from Al3Ti and TiN. The ΔσOrowan was thus determined to be 27 and 103 MPa for the 7050–1.82Ti and 7050–2Ti/TiN, respectively. Additionally, the load-bearing effect also contributes to the strengthening of the LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN specimens because of the interfacial bonding between Al3Ti or TiN nanoparticles and Al matrix, which can be estimated by where Vp is the volume fraction of the particles, and σi is the interfacial bonding of Al/TiN. The UTS data for crack-free LPBF-fabricated 7050 (σ7050) is not available in the literature. The calculated σload values for the LPBF-fabricated 7050–1.82Ti and 7050–2Ti/TiN were noted as 360-σ7050 and 374-σ7050, respectively. The values 360 and 374 were obtained by subtracting ΔσHall-Petch and ΔσOrowan from the ultimate strengths of the two alloys, respectively.The calculated strength increments are listed in . In the calculation, the dislocation density-strengthening mechanism was neglected. This assumption was based on the fact that no apparent dislocations surrounding the TiN nanoparticles or precipitates were observed in the LPBF-fabricated 7050–2Ti/TiN samples (). The contribution of Ti solutes to the strengthening effects is also assumed to be negligible because of the equal amounts of Ti in the 7050–1.82Ti and the 7050–2Ti/TiN samples. Ti/TiN hybrid refiner contributes 118 MPa to the total strength enhancement in the 7050–2Ti/TiN, compared to that of the 7050–1.82Ti. The enhancements of ΔσHall–Petch, ΔσOrowan, and Δσload are 28, 76, and 14 MPa, respectively.From the results given above, it is noticeable that Ti and TiN in the Ti/TiN hybrid refiner produce a significant synergism in grain refinement, thereby enhancing the mechanical properties of the LPBF-fabricated 7050 alloy. The enhancements of ΔσHall–Petch, ΔσOrowan, and elongation are closely related to the grain refinement effects and uniform particle dispersion dominated by special solidification behavior, whereas that of Δσload relies on the Al/TiN interface strengthening in a Ti-rich environment. The effects of the interfacial binding strength between α-Al matrix and nanoparticles are discussed in the following sections in terms of solidification behaviors of the LPBF-fabricated 7050-Ti/TiN alloys.The solidification behaviors of the LPBF-fabricated 7050, 7050-TiN, 7050-Ti, and 7050-Ti/TiN alloys are schematically illustrated in In LPBF-fabricated 7050, hot tearing cracks were caused by a large residual thermal stress (a) generated during solidification and propagated along the coarse columnar grain boundaries. The thermal stress tore the fragile intergranular regions containing the elements Mg, Zn, and Cu, and the segregation of oxides, which were brought from the surface into the metal liquid during processing To restrain the epitaxial growth of Al, TiN nanoparticles were introduced into the molten pool. The TiN nanoparticles, with their potential to act as nuclei due to the low misfit of the lattice constants between TiN and Al, were expected to refine grains and suppress cracking in the LPBF-fabricated 7050-(0.18–0.54)TiN alloy during solidification (b). However, cracks were observed in the 7050-TiN alloys. This may be because the thermal undercooling was insufficient to surpass the critical undercooling for the activation of heterogeneous nucleation on the TiN nanoparticles. According to the free growth theory (ΔTn= 4σSL/Svdp, where σSL is the solid–liquid interface energy, Sv is the entropy of fusion per unit volume, and dp is the diameter of the particle) e). Particle agglomeration within the liquid is thermodynamically spontaneous, driven by a negative change in the interfacial free energy from the particle pair system to the coalesced doublet-particle system shows the agglomeration of TiN nanoparticles at the grain boundary in the LPBF-fabricated 7050–0.54TiN sample. Therefore, a uniform distribution and good dispersion of the inoculant are critical and even essential for heterogeneous nucleation. Previous work shows that only when the volume fraction is 1–20%, or when the mass fraction is 1–30%, the ceramic nanoparticles can be considered as valid nuclei LPBF-fabricated 7050–1.82Ti has a crack-free and relatively fine microstructure (f). The mechanism of grain refinement can be described as follows. First, Al3Ti particles form and promote heterogeneous nucleation, that is, at a lower ΔTn. Tan et al. ), implying that a lower undercooling for Al3Ti is required for nucleation. As a result, the epitaxial growth of grains and hot tearing were restrained. However, although the grains can be refined in LPBF-fabricated 7050-Ti sample, the mechanical properties, especially elongation, are still not adequate when compared to those of wrought alloys. This may be related to the size, dispersion, and fraction of the in-situ Al3Ti particles Ultrafine grains were achieved in the LPBF-fabricated 7050–2Ti/TiN alloy, which mainly benefited from the synergism of Ti/TiN (d). The fine grains imply that, comparing with the 7050–1.82Ti system, in addition to the ΔTCS originated from Ti solute and lower ΔTn resulted from the in-situ formed Al3Ti, TiN nanoparticles contribute to further decrease in ΔTn in a Ti-rich liquid during solidification of the molten pools. During LPBF processing, if irradiated by laser beams, the 7050 powders melt at first owing to their low melting point (933 K), and then available Ti particles dissolve into the liquid Al homogeneously, owing to the high diffusion coefficient of Ti in high-temperature liquid Al (e.g., Cd = ~4.94 ×109 m2/s, at a temperature of 1173 K) g and h). All the variations influenced the properties of the solidified Al/TiN interface.In the LPBF-fabricated 7050–2Ti/TiN with a TiN nanoparticle content of 0.9 vol%, the enhancement of Δσload (14 MPa) implies a high binding strength σi of Al/TiN interface of up to 1073 MPa (see Methods for calculations). This finding is consistent with the calculation results from Lin et al. , the Al(001)/TiN(001), Al(011)/TiN(011), and Al(111)/TiN(111) interface structures are doped with Ti atoms. The crystallographic parameters of the experimentally determined Al and TiN are listed in . Density functional theory (DFT) calculations for the binding energy difference (ΔE) between the Al/TiN interfaces with and without Ti atoms were performed. ΔE has been widely used as a descriptor of the interface stability , indicate that after doping with Ti, the ΔE values of the Al(001)/TiN(001), Al(011)/TiN(011), and Al(111)/TiN(111) interfaces were − 0.5704 eV, − 0.5708 eV, and − 0.8099 eV, respectively. All the calculated ΔE values were negative, suggesting that Ti atoms enhance the interface stability between TiN and Al atoms, and further strengthen the Al/TiN interface binding.We designed and produced a novel hybrid grain refiner consisting of solute and ceramic nanoparticles for the LPBF additive manufacturing of high-strength aluminum alloys. Adding a Ti/TiN hybrid grain refiner to the LPBF-fabricated 7050 alloy can produce ultrafine grains with an average size of 775 nm, resulting in an ultimate tensile strength of up to 408–618 MPa and ductility of up to 13.2–8.8%. These tensile properties are comparable to those of conventional wrought 7XXX alloys. During LPBF processing, the hybrid grain refiner exhibited interesting synergistic grain refinements and strengthening mechanisms between the solute and the ceramic nanoparticles. During solidification, not only in-situ particles formed by the chemical reaction of solute in liquid Al and external added ceramic nanoparticles acted as the nuclei of α-Al respectively, but also solute inhibited the agglomeration of the ceramic nanoparticles to promote their nucleation efficiency. Furthermore, the strength of the LPBF-fabricated 7050 alloy was further improved by the doped solute atoms at the ceramic nanoparticle/Al interface. The strengthening mechanisms were attributed to the Hall–Petch, Orowan, and load-bearing transfer from the particles inside the matrix and further strengthening of the solute upon binding on the ceramic nanoparticle/Al interface. The improvement in elongation benefited from the uniform dispersion of the various particles.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Decoupling loading effect in simultaneous sensing and actuating for dynamic measurementWhen doing dynamic measurement using an electromechanical transducer for both sensing and actuating simultaneously, loading effect of the transducer is obviously large enough to jeopardize measurement accuracy. This paper examines the role of a newly introduced ‘transduction matrix’ in decoupling the loading effect from a piezoceramic inertial actuator (PIA) which is employed to detect pseudo mechanical impedance of a structure as both exciter and sensor simultaneously. The two by two transduction matrix characterizes the relationship between the input electrical variables and the output mechanical variables of the PIA. Once it is properly calibrated, the mechanical impedance of the structure can be indirectly detected via measuring the input electrical impedance of the actuator. In this presentation, the appropriateness of a numerical model of the PIA is first validated through comparison with experimental results. Then, the influences of design parameters of PIA on its transduction matrix as well as measurement accuracy are studied numerically in detail. The results show that, although the transduction matrix varies significantly due to changes of design parameters, the detected mechanical impedance remains the same for a given structure, as the loading effect of the PIA is satisfactorily decoupled.In literature, piezoceramic patches used as actuating or sensing elements for dynamic measurement have been widely reported. When supplied with electrical field, PZT patches create strain due to inverse piezoelectric effect and function as actuators. On the other hand, when loaded with external strain, PZT patches produce electrical charges due to direct piezoelectric effect, which makes them suitable for sensing In principle, employing PZT for dynamic sensing and actuating simultaneously is also feasible (for example, The homemade piezoceramic inertial actuator employed in the study consists of an inertial mass, an aluminum plate, a PZT patch and necessary frameworks (). The changeable inertial mass is affixed at the center of the top surface of the aluminum plate. A square PZT patch is bonded to the bottom surface of the aluminum plate and the upper and lower frames clamp the plate. When ac power is supplied to the PZT patch, it extends and contracts in x–y plane due to inverse piezoelectric effect; this causes the transactional bending of the aluminum plate, which in turn accelerates the inertial mass. If the PIA is properly linked with a point on the test structure, the reaction force of the motion of the inertial mass is exerted to the structure. In this process, PIA serves as an actuator. During actuating, the response of the structure affects the motion of the PIA and results in additional deformations of the PZT patch. As a result, the input electrical impedance is changed accordingly due to the direct piezoelectric effect. If the electromechanical interaction taken place in PIA is properly modeled, the mechanical impedance at its output port can be evaluated from the measured electrical impedance at the input port.Since the piezoceramic inertial actuator is a linear dynamic transducer without internal energy generation or consumption, it can be characterized by a four-pole model as shown in [T¯] is named the transduction matrix of PIA and its four elements are complex transduction functions with both amplitude and phase information. They are in fact four different frequency response functions describing the relationships between one input variable and one output variable of the PIA in frequency domain. From Eq. , T¯11=EFv=0 and T¯21=IFv=0 are the transduction functions of the input voltage and input current to the output force when the PIA is clamped onto the ground. Similarly, T¯12=EvF=0 and T¯22=IvF=0 are the transduction functions of the input voltage and current to the output velocity of PIA when the actuator is freely suspended., the mechanical impedance at the output port can be calculated from the measured input electrical impedance of the actuator:where F is the excitation force exerted to the host structure, v the corresponding response velocity, Z¯M the point mechanical impedance of the test structure (which is also the excitation point), and Z¯e=EI is the input electrical impedance of the PIA device.The transduction matrix of a PIA has to be identified before the method is applied to a structure. For either numerical or experimental identification, the electrical voltage and current at the input port of PIA is measured when it is powered under two boundary conditions: mechanically clamped and mechanically free. Under clamped condition, the transuction functions T¯11 and T¯21 are determined; under free condition, the other two transduction functions T¯12 and T¯22 are evaluated.To identify the transduction functions numerically, the finite element package ANSYS was employed to model and simulate the dynamic behavior of PIA. provides the physical properties and element types used in the finite element model. In the analysis, the actuating point () of the PIA is subjected to clamped and free boundary conditions, repectively and harmonic analyses are carried out to calculate the four transduction functions.In experiments, the clamped condition is achieved by clamping the PIA on a 70 kg base, which is hundreds of times heavier than the actuator, and the free condition is approximated by suspending the PIA using a rubber band; the natural frequency of the suspended system is far below the interested frequency range of the PIA. depict the experimental set-up for determining transduction functions under grounding and free boundary conditions respectively (see Ref. , it can be seen that for all four transduction functions the numerical and experimental results have good coincidence with small errors over a wide frequency range except the neighborhood of resonance frequencies. These discrepancies may be due to two reasons. Firstly, the numerical simulation is based on ideal conditions in which the PZT patch is perfectly integrated onto the aluminum plate; while, in the experiment, the PZT is glued on the aluminum plate by epoxy which to some degree does not provide perfect bonding as ideal as assumed in numerical calculations. Secondly, the use of force transducer affects experiment accuracy. In spite of these discrepancies, the comparison concludes that (1) the method for identifying transduction matrix can be conducted both experimentally and numerically with tolerable errors, and (2) the employed finite element model is accurate enough in predicting the dynamic behavior of PIA and thus can be exploited to investigate the ‘decoupling’ function of transduction matrix in the following sections.After determining the transduction functions, a simple case study is carried out to demonstrate the effectiveness of the proposed method and serves for comparisons to be done in the next section. A clamped–clamped aluminum beam, which is 786 mm in length, 99 mm in width, and 10 mm in thickness, is utilized as the test structure. The piezoceramic inertial actuator is affixed at 123 mm to the left end of the beam, as shown in (not to scale). When harmonic voltage is supplied to the PIA, its input electrical impedance is measured by an HP4192A LF impedance analyzer. The pseudo mechanical impedance at the excitation point was then calculated from the measured electrical impedance using the identified transduction functions according to Eq. The same testing was then conducted numerically. In the model for finite element analysis using ANSYS, the test aluminum beam was modeled by element solid 73 (243 in number). A sinusoidal voltage is supplied to the PIA and its input electrical impedance is numerically determined. With the available transduction matrix numerically obtained earlier, the pseudo mechanical impedance of the aluminum beam is calculated from the input electrical impedance of the actuator according to Eq. , the amplitude of the detected mechanical impedance of the beam, both experimentally and numerically, are illustrated. It can be observed that the results are comparable.In dynamic measurement, sensors are made small to minimize its disturbance or loading effect to the measurand in a measured system. On the other hand, actuators are made large so that the power it exerts can be kept as desired regardless the varing loading from the sytem under its actuation , it is seen that the transduction matrix characterizes and quantifies the forward actuating and backward sensing functions in a transducer. Through a simple mathematical operation, it can decouple the loading effect of the tranducer exerted onto the system being actuated and realizes quantitative measurement of the force, motion as well as pseudo impedance of the structure at the actuating point. In the following, taking advantage of the validated numerical model of the PIA in the above, numerical experiments are conducted to reveal the capability of the introduced transduction matrix in decoupling, or removing the loading effect.In our PIA design, the inertial mass is changeable to suit different required loading in dynamic testing. Changes of the inertial mass obviously alter the dynamic property of the PIA and thus its transduction matrix. Using the numerical model, the effects of inertial mass changes (90.5, 180.0 and 271.5 g) to the transduction matrix were simulated. As shown in , the resonance of all four transduction functions moves as expected to lower frequencies as the inertial mass increases. When these different PIAs are utilized to measure the point mechanical impedance at the same point of the same test beam structure used in the above section, the results should possess no discrepancies from each other if the above-claimed capability of decoupling properly exists. In , the electrical impedance numerically ‘measured’ from the three PIAs are plotted and the considerable differences among them are observable. After the decoupling done by Eq. , however, the resulted mechanical impedance at the actuating point from the three PIAs gives the same point mechanical impedance as presented in . This concludes that the transduction matrix decouples load effect from a transducer appropriately when doing actuating and sensing simultaneously.As a double check, the effect of the dimension changes of the PZT patch of a PIA on impedance measurement is also examined. PZT thickness changes of 0.3, 0.4, and 0.5 mm should greatly affect the dynamic behavior of the actuator and thus alter the corresponding transduction matrix. As illustrated in , as the thickness increases, the four transduction functions exhibit considerable differences in both magnitude and resonance frequency. When these PIAs are affixed to the same test beam, the corresponding electrical impedance also presents big differences (, the mechanical impedance of the beam obtained from the transduction matrices and electrical impedance almost coincides, which further verifies that the transduction matrix does decouple loading effect quite ideally.In this paper, the method for detecting point mechanical impedance of structures using a piezo-ceramic inertial actuator was further studied. A numerical model was created to model the dynamic behavior of the PIA and the simulation results coincide well with experimental results obtained from a simple structure test. Using the numerical model, two numerical experiments were done which successfully demonstrate the role of the introduced transduction matrix in decoupling the loading effect of the actuator when conducting simultaneous sensing and actuating. Through this decoupling capability of transduction matrix of an electromechanical transducer, quantitative dynamic measurement is thus made possible in simultaneous sensing and actuating.Shih-Fu Ling was trained in vibration and noise engineering in Purdue University, USA. He had 10 years industrial experience in automobile design and manufacturing in Taiwan. After joining Nanyang Technological University in Singapore in 1991, he has concentrated his research in dynamic measurement, simultaneous sensing and actuating, and electromechanical transducers.Xiaoyan Hou just submitted her PhD thesis to Nanyang Technological University, Singapore where she had devoted to development of a transducer for pure moment exertion and rotary speed measurement simultaneously. She holds a BEng degree from Department of Precision Machinery and Instrumentation, University of Science and Technology of China.Yi Xie is currently working as a research engineer in oil industry in Singapore. He was a research fellow in Nanyang Technological University working on simultaneous sensing and actuating. Before coming to Singapore, he did research in dynamic measurement as a graduate student in University of Science and Technology of China and a postdoctor in Tsing-Hua University, China.Fatigue behavior of laser consolidated IN-625 at room and elevated temperaturesLaser consolidation (LC) is an emerging manufacturing process that produces net shape functional and metallurgically sound components by material addition based on a CAD model. Due to the rapid solidification inherent to the process, LC materials generally have excellent material properties. However, there is limited information available on the fatigue properties of LC materials. Throughout this work, the fatigue behavior of LC IN-625 was evaluated at both room temperature and 650 °C and compared with wrought and investment cast IN-625 materials. The results show that the LC IN-625 has a fatigue resistance higher than the cast material but lower than the wrought material. Tensile and fatigue test results along with the examination of microstructures and crack initiation sites are presented.Laser consolidation (LC) is an emerging manufacturing process being developed to produce net shape functional and metallurgically sound components by material addition based on a computer aided design (CAD) model. The process is illustrated in . Starting from an existing component or substrate, three-dimensional parts are built using a laser beam and metal powder stream to deposit material along the cross-section of a component layer by layer. The motion of the substrate relative to the laser beam and metal powder stream is controlled using a multi-axis computer numerical controlled (CNC) motion system. The airborne metal powder is injected into a molten pool created by the focused laser beam and rapidly solidifies onto the previous layer. By adding successive layers, three-dimensional metallurgically sound components free of cracks and porosity can be formed with relatively high tolerances and surface finish Due to the rapid solidification that is inherent to the LC process, various materials processed by LC such as Ni-alloys, Co-alloys, Ti-alloys, stainless steels and tool steels have been reported to have excellent material properties The LC process produces unique microstructure and therefore unique material properties. In this study, the axial fatigue behavior of the Ni-based Inconel alloy 625 (IN-625) processed by LC was evaluated in air at room and elevated temperatures (650 °C) and compared with two other conventional methods of fabrication: wrought and investment casting. The microstructure of LC IN-625 is essentially a cast microstructure but it is refined due to the rapid solidification inherent to the process. It also shows a directionally solidified dendritic microstructure that is free of cracks and porosity. The tensile properties of LC IN-625 have been reported to be better than the cast IN-625 material and comparable to the wrought IN-625 material The chemical compositions of the wrought (sheet), investment cast and powder used to make the LC specimens are listed in . All IN-625 materials were purchased from commercial suppliers. The potential residual stresses induced by the different manufacturing methods were eliminated by subjecting all the specimens to a standard stress relieving treatment of 850 °C for 15 min in an air environment inside an electric furnace followed by air-cooling The wrought specimens were sheared and machined from IN-625 sheet of 0.9 mm thickness. Two orthogonal directions (A and B) were labeled on the sheet and used to identify the orientation of the tensile and fatigue specimens tested throughout this study. Optical micrographs of the wrought material are shown in along the side and across the thickness of the sheet. There is essentially no microstructural difference between the two sections. The grains are equiaxed and relatively fine which is the desired outcome of hot working processes The cast specimens were produced from an investment cast bar (15 cm × 15 cm × 2.5 cm thick) of IN-625 purchased from a commercial supplier. Specimens were sliced to proper thickness (0.9 mm) using wire electrical discharge machining (EDM) and milled to final shape. The specimens tested were well below the outer surfaces of the bar to avoid using material in the chill zone of the casting. The thin re-cast layer that is typically produced by wire EDM cutting was found to be completely removed by the standard manual polishing procedure used on all the specimens. Optical micrographs of the cast material shown in reveal a relatively coarse dendritic structure. An overall examination of the dendritic structure of the entire section did not indicate any preferred orientation of the dendritic structure. No cracks were found in the material.Another section of investment cast IN-625 was polished and observed with the Scanning Electron Microscope (SEM) using Secondary Electron (SE) and Backscatter Electron (BSE) detectors. The SE image shown in reveals a significant amount of porosity in the cast material, which appears to be shrinkage pores at interdendritic regions (region 1) as well as gas evolution during solidification (region 2) due to their round and elongated shapes with smooth contours. The shape of the void at region 3 suggests the location of an incoherent particle or inclusion that may have been plucked out during the grinding and polishing procedure. The BSE image in b is more sensitive to the average chemical makeup of the specimen and reveals regions having different compositions within the material. The chemical composition of the white particles (such as regions 4, 5 and 6) was analyzed using Energy Dispersive Spectroscopy (EDS) and found to be rich in Nb content. The EDS spectra for those three regions showed a composition of 0–0.78% Mg, 2.38–2.83% Cr, 6.51–8.32% Ni and 88.84–90.87% Nb. Two other EDS spectra were taken from the matrix (regions 7 and 8) for reference purpose and had a chemical composition close to the nominal values listed in The LC specimens were manufactured in-house from commercially available IN-625 powder with a particle size of −45 μm/+10 μm. The LC process was conducted inside a glove box where the oxygen content was maintained below 50 ppm throughout the process. The specimens were built on a base plate in the form of square tubes with approximately 0.9 mm wall thickness using an average laser power ranging from 30 to 300 W and powder feed rate ranging from 2 to 20 g/min. The sides of the square tubes were separated and four specimens were machined from each tube. Two sets of specimens produced using different processing conditions were tested in this study. The first set was produced using low-energy input parameters through low power and/or high traverse velocity and the second set was produced using high-energy input parameters. All the mechanical tests were conducted in the direction parallel to the build direction.The optical micrographs of LC IN-625 sample manufactured from low-energy input parameters shown in reveal a fine directionally solidified dendritic microstructure. The columnar dendrites in a are aligned almost parallel to the build direction. The size of the dendritic cells shown in b is in the 2–3 μm range. These observations are consistent with previous work done on this material has the same characteristics as the low-energy input parameters. The main difference between the two samples is the size of the dendritic structure. Close examination of Figs. b reveals that the cell size of the columnar dendrites is approximately 2–3 times larger for the high-energy parameters than the low-energy parameters. This is likely due to the lower cooling rates produced from the high-energy parameters. The darker regions in b are interdendritic regions that were revealed during metallurgical characterization of the specimen. There was no porosity or cracks observed inside either sample produced from different LC parameters.The room temperature tensile properties of all three forms of IN-625 material in the SR condition were evaluated and the results are listed in along with some reference values. The results show that, in general, the LC IN-625 material has tensile properties superior to the investment cast material and slightly lower but comparable to the wrought material. The LC material produced with high-energy parameters had a slightly lower strength but a higher ductility than the material processed with low-energy parameters. The tensile properties of the cast IN-625 material were notably lower than the reference values. However, it should be noted that there are several factors that can significantly affect the strength of castings and the material and tests conditions for the reference values were not indicated. On the other hand, the tensile properties of the wrought material used for comparison throughout this study exceeded the reference data.The heating and temperature control of the specimen were achieved using an induction heater integrated with a proportional, integral and derivative temperature controller and thermocouples connected to the specimen. Using a dummy fatigue specimen having 5 thermocouples connected along its reduced section, the induction coil geometry was calibrated and produced a very uniform and stable temperature distribution along the reduced section of the specimen. During the calibration, the steady-state temperature at the minimum section was 650 ± 4 °C with a temperature gradient of less than 10 °C within 10 mm from the minimum section. During the actual fatigue experiments, the temperature was controlled and monitored throughout the entire duration of each test using 2 thermocouples connected 10 mm away from the minimum section. The fatigue tests were only started once the specimen temperature had reached the set point and allowed to stabilize for a period of 20 min.All the mechanical tests were performed on an Instron 8516 servo-hydraulic testing system with a 100 kN load cell. The tensile tests were conducted according to ASTM Standard E8M-01 and the room temperature fatigue tests were conducted in air according to ASTM Standard E 466-96 (Re-approved in 2002). All the tensile and fatigue specimens were manually polished along the loading direction down to 600-grit Silicon carbide paper. The specimen geometry used for the tensile and fatigue tests are shown in respectively. The grip sections of the specimens used for elevated temperature fatigue were made longer to provide enough clearance between the grips for the induction coil. The run-out limit was set to 107 cycles for the room temperature tests, and 2 × 106 cycles for tests at 650 °C. The loading regime was a stress-controlled, sinusoidal waveform with a min/max stress ratio of R |
= 0.1 (tension–tension), and a frequency of 60 Hz.The results of the fatigue tests at room temperature are displayed in . The LC IN-625 material tested in the build direction had a fatigue resistance significantly higher than the investment cast but lower than the wrought material. There was no significant difference between the LC material produced from low-energy parameters and high-energy parameters. The endurance limit of the LC material was just under 450 MPa, which is about 200 MPa higher than the investment cast material and about 100 MPa lower than the wrought material. There was essentially no difference in fatigue resistance between the 2 orthogonal directions (A and B) on the wrought IN-625 sheet. This is consistent with the experimental results reported by Grubb In principle, the LC process is a form of casting process where the material is first melted and then re-solidified to form a desired shape. The microstructure of the LC IN-625 material is essentially a cast microstructure. However, due to the rapid solidification that is inherent to the process it is much more refined and uniform compared to the investment cast material. Furthermore, the LC material is free of cracks and porosity. This proved to be beneficial to the room temperature fatigue properties. Even though the cell size of the columnar dendrites for the high-energy parameters appears to be 2–3 times larger than that of the low-energy parameters, the difference in cell size does not seem to be sufficient to affect the room temperature fatigue resistance of LC IN-625. The microstructure of the high-energy parameters LC IN-625 is still relatively fine in relation to the cast material. It is recommended that additional experiments be conducted on LC IN-625 to evaluate a wider range of cell size and demonstrate the effects of cell size on the room temperature fatigue resistance of LC IN-625.Conventionally cast Ni-based superalloys generally have a coarse dendritic microstructure with residual porosity The wrought IN-625 sheet was evaluated within this study to establish another baseline for comparison purposes. Generally, wrought materials have superior mechanical properties than castings because deformation processing can eliminate casting porosity and produce a uniform grain structure with a more homogeneous chemical composition. Therefore, it is unsurprising that the wrought material had the highest room temperature fatigue resistance compared to the investment cast and LC IN-625. It had an equiaxed and relatively fine microstructure which offered more resistance to cyclic deformation and a higher fatigue endurance limit than the LC (refined cast structure) and investment cast IN-625 materials.During high cycle fatigue the crack initiation stage is very important and consists of the largest portion of the total fatigue life. The mode of fatigue crack initiation for each form of IN-625 was investigated by SEM examination of fracture surfaces. SEM images of the overall fracture surface and crack initiation site for a typical specimen of wrought, LC and investment cast IN-625 are included in respectively. For comparison purposes, the specimens selected for examination had comparable fatigue lives in the mid range of the fatigue curve and were all in the SR condition. The wrought specimen selected for examination was tested in direction A on the sheet and the LC specimen examined was produced from low-energy process parameters.b. Inside the investment cast specimen however, the crack initiation site was well below the surface. There were also several voids identified at the initiation site that are visible in b. The shape of these voids suggests that they may have been the locations of incoherent particles that were plucked out during fracture. The composition of these particles could not be obtained directly from the fracture surface but the polished section of cast material analyzed with EDS shown in identified numerous Nb-rich particles randomly distributed throughout the cast material. Further investigation would be required to positively identify the source of these voids. In any case, this cluster of voids would have certainly produced an area of high stress concentration within the specimen and caused early crack initiation which resulted in relatively poor fatigue resistance of the investment cast material.The high temperature fatigue tests were conducted on all three forms of IN-625 in air at 650 °C using the same loading regime as for the room temperature tests. The results of the elevated temperature fatigue tests are displayed in . The wrought material was tested in direction A on the sheet and LC specimens produced from low-energy parameters were tested in the build direction. The elevated temperature environment caused a significant decrease in fatigue resistance for all three forms of IN-625 compared to room temperature but the trend was the same as for the room temperature tests. The LC IN-625 material still had a fatigue resistance higher than the investment cast material but lower than the wrought material. The endurance limits were found to be approximately 450 MPa for the wrought material, 325 MPa for the LC material and around 200 MPa for the cast material.Generally, the strength of metals decreases with increased temperature. At elevated temperatures, additional thermally activated and time-dependent factors (such as creep/relaxation and oxidation) are acting in combination with mechanical fatigue. Consequently, the fatigue strength also decreases with the increased temperature. The service temperature for IN-625 is reported to range from cryogenic to 982 °C (1800 °F) As for the room temperature experiments, the mode of fatigue crack initiation for each form of IN-625 tested at 650 °C was investigated by SEM examination of fracture surfaces. SEM images of the overall fracture surface and crack initiation sites for a typical specimen of wrought, LC and investment cast IN-625 are included in respectively. For comparison purposes, the specimens selected for examination had comparable fatigue lives and were all in the SR condition.At elevated temperature, a change in failure mechanism from cycle-dependent to time-dependent was observed from the fracture mode. At elevated temperature, intergranular fracture usually replaces transgranular fracture, which is more common at room temperature. This is caused by the interaction of creep and fatigue mechanisms and also oxidation when the material is exposed to air respectively revealed intergranular crack initiation at 650 °C. This indicates that at this temperature, the initiation of a fatigue crack in these materials relied heavily on the exposure time. To further emphasize this point, there were two crack initiation sites on the LC fracture specimen shown in b and c. The two initiation sites appear to have acted simultaneously but independently before they eventually combined into one larger crack. They both have similar topographic features and their local crack propagation regions are relatively the same size. On the investment cast specimen there were no signs of intergranular cracking at the initiation site shown in b. The crack initiated at a relatively large interdendritic pore (8–10 μm) located below the surface. This pore produced a region of high stress concentration within the specimen and caused early crack initiation, which resulted in relatively poor fatigue resistance of the investment cast material.Throughout this study, the LC material was only tested in the build direction. However, as shown in the microstructure of the LC IN-625 material has a dendritic structure with a preferred orientation. Therefore, it is highly probable that the fatigue behavior of this material is dependent on the loading orientation relative to the build direction. In addition, the microstructure of the LC material is dependent on the process parameters. The coarser microstructure of LC material produced from the high-energy parameters did not influence the room temperature fatigue behavior but was not evaluated at 650 °C. Since the fatigue process in the LC material is sensitive to time-dependent factors, the coarser microstructure that is free of cracks and porosity may be beneficial in this case. Consequently, with a more comprehensive understanding it may be possible to tailor the mechanical properties of LC components to the requirements of specific applications by varying the processing parameters and build orientation. Additional experiments are planned to verify these observations.The LC IN-625 material produced from the low-energy input parameters has a finer microstructure and superior tensile properties than the LC specimens built with high-energy input parameters.At room temperature, LC IN-625 in the SR condition tested in the build direction has a fatigue resistance higher than the investment cast material but lower than the wrought material. There was no significant difference in fatigue resistance between the LC samples produced from low-energy parameters and high-energy parameters.The crack initiation in the wrought and LC material at room temperature occurred near the surface by stage I cracking. However, inside the cast material, the crack initiated at a cluster of voids, which created a high stress concentration and resulted in relatively poor fatigue resistance.At 650 °C, the fatigue resistance of all three forms of IN-625 material decreased but the trend was the same as the room temperature results. The LC IN-625 produced with low-energy parameters in the SR condition tested in the build direction has a fatigue resistance higher than the investment cast material but lower than the wrought material.At 650 °C, the mode of crack initiation was intergranular for both the LC and the wrought IN-625 material, while a relatively large pore caused crack initiation inside the cast material.Stipagrostis pungens is a traditional medicinal annual plant growing in Tunisia. It is quite a small plant suitable for Tunisia arid areas. This paper deals with the study of the chemical composition and the cooking of this raw material, in order to produce lignocellulosic fibres. The chemical composition of S. pungens has showed that it contains quite high amount of extractives, lignin (around 12%) and polysaccharides (71%). The α-cellulose amounts are acceptable (44%). Finally, the ash content was around 4.65%. With the insight of such data, we have undertaken the soda-anthraquinone cooking of S. pungens and produced lignocellulosic fibres. Different cooking temperatures (120 and 140 °C) and times were tested. A yield of about 43% w/w was obtained and the produced pulps had a kappa number of around 11. Finally, the isolated fibres were used to produce laboratory handmade standard sheets. The structural and mechanical properties of the prepared samples were close to those of other common annual plants-based fibre-mats.The worldwide pulp and paper industry has been growing rapidly. As a result there has been a huge demand for pulp and paper making raw materials. In some countries, pulp and paper industries have been facing environmental pressure and insufficient forest resources. They use non-wood, annual plants and agricultural residues for paper production in order to solve this dead end situation and find new applications for these raw materials. There are many studies that have been carried out over many years. They were devoted to investigate the use of annual plants or/and agricultural wastes as alternative sources of fibres. These include annual plants such as: Cajanus cajan (), Amaranthus caudatus L. and Jerusalem artichoke (), as well as agricultural residues including rapeseed straw (In this perspective, we try to exploit a lignocellulosic material widely available in Tunisia, as a source of cellulose fibres, namely: Stipagrostis pungens. Thus, S. pungens (Poaceae) is a perennial grass widely distributed on the North Africa arid regions and it is also distributed in tropical West Africa and West Asia. The leaves dimensions are as tall as 1 m and they are valuable sources of forage (). Whereas, it is extensive and deep root system diminish erosion. This spontaneous plant is used in popular medicine to treat indigestion, stomach ache, wound cicatrisation and constipation (). The antimicrobial activity of the S. pungens extracts was studied (). It was found that, the ethyl acetate extracts of the leaves of S. pungens L. show a significant antibacterial activity on Pseudomonas and on a large spectre of fungi. Furthermore, the chemical components (polysaccharides, lignin, and phenolic acids) from S. pungens was performed in comparison with two perennial sub desert grasses: Stipa tenacissima, Ligeum sparteum (). Hemicelluloses, the main matrix cell wall polymers of this species were isolated from leaves, characterised and converted as plastic material (Thus, this paper reports the characterisation of the Tunisian cellulosic product S. pungens and the evaluation of their properties, in the papermaking context. To the best of our knowledge, no report available was devoted to the chemical composition and pulping of this vegetal species.The first part of this work was devoted to the determination of the chemical composition (lignin, holocellulose, cellulose, extractives and ash). Then, soda-anthraquinone cooking, which is known to be tailored for pulping annual plants (), was tested. The ensuing pulps were characterised in terms of yield, kappa number and degree of polymerisation (DP). Finally, the physical and mechanical properties of hand sheets were tested and they are presented and discussed. The obtained results (chemical composition of pulps and physical properties of papers) are compared to other wood, non-wood and annual plants.The raw material used for various tests is the aerial part of S. pungens (). Depending on the area in which they grow, it may have different morphological and chemical characteristics. So it is necessary to give a description of the materials that formed the basis for this study in chemical and morphological terms.This plant was taken from the region of Kasserine (latitude 35° 10′ 3″ N and longitude 8° 50′ 11″ E) in December 2012. It was air dried under natural conditions during January 2013 (average relative humidity = 63%; average temperature = 16 °C). They were then washed in order to eliminate sand and dried again under the same conditions. For chemical composition determination, S. pungens samples were ground and 40–60 mesh fractions were selected. The procedures were performed according to standard T264 cm-07. Before pulping, S. pungens was cut into small pieces with lengths of about 2–3 cm.The morphology of S. pungens stems was studied by scanning electron microscopy (SEM). The images were collected using ABT unit 55 which is characterised by an accelerating voltage of 0.5–30 kW and a resolution of 4.5 nm.Chemical analysis of the aerial part of S. pungens was determined by classical methods. The evaluation of extractive substances was carried out in various liquids according to common standards. The solubility in hot and cold water, in NaOH 1% and in ethanol–toluene was determined according to TAPPI standards methods: T207 cm-08, T212 om-07 and T204 cm-07 respectively, the rate of ash was determined according to standard T211 om-07. The amounts of lignin, holocellulose, α-cellulose, and the kappa number were also evaluated using the following respective TAPPI standard methods: T222 om-06, the method of , T203 cm-99 and T236 om-06. All the experiments were duplicated and the difference between the values was within and experimental error of 5%.Elemental analysis was carried out by the “Service Centrale d’Analyse (Vernaison, France)” of the “Centre National de la Recherche Scientifique” (CNRS). Carbon, Hydrogen, and Oxygen contents were measured.The thermal stability of the fibre samples were characterised using a TA instrument Q-50 thermogravimetric analyzer (TA Instruments, USA). The samples were heated from 30 °C to 800 °C at a heating rate of 10 °C/min, under a stream of nitrogen.Soda-anthraquinone pulping was carried out and it was performed in a rotating system consisting of 6 vessels each with a capacity of 1 L (multi-shell), heated electrically, under controlled temperature. Briefly, 60 g of chips S. pungens were cooked in 600 mL of a freshly prepared caustic soda. The amount of anthraquinone was kept at 0.1% (w/w), with respect to the oven dried (o.d.) initial solid material.After cooking, a suspension of brownish cellulose pack of fibres in the black liquor was obtained. Cellulose fibres become fairly well individualised and flexible after the disintegration of the cooked chips. Then they are separated from black liquor, washed with water until obtaining a clear filtrate. The resulting pulps were characterised in terms of yield, kappa number and degree of polymerisation. The cooking yield was calculated as the ratio of weight of o.d. material after washing to that of the initial raw material. The viscosity of pulp dissolved in a cupriethylenediamine (CED) solution was determined according the standard method TAPPI T230 om-99. This value was then converted into degree of polymerisation (DPv) using the following equation (where η (mPa s) is the intrinsic viscosity. Briefly, this method implies the soaking of the pulp samples in water with further centrifugation at 3000 × |
g during 15 min, and the WRV were calculated using the following equation:Ww is the mass of the wet sample after centrifugation and Wd is that after drying of the wet sample at 105 °C to a constant weight. The pulp drainability was evaluated by measuring the Shopper Riegler degree (SR – ISO 5267-1).The unbeaten screened pulp suspensions were diluted to 2 g/L. Then, conventional hand sheets with a basis weight of 60 g/m2 were prepared on a Rapid Khöten sheet former following the standard method IS0 5269-2. Prior to testing, the hand sheets were conditioned (23 °C, 50% relative humidity – ISO 187) and structural and mechanical properties were determined by measuring basis weight, thickness, bulk and permeability, as well as the tensile, burst and tear strength according to their respective standards ISO 536, ISO 534, ISO 5636-3, ISO 1924-3, ISO 2758 and ISO 1974. The elemental composition of the paper was determined by Scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS). presents SEM micrographs of the surface morphology of S. pungens's stems. The stems have a homogeneous aspect with a rough surface. The observation of a cross-section of the sample clearly shows the different layers that form the main components of the studied material, namely the bundles of fibres (fibres long cells or ultimate), timber vessels and vascular short cells called sclereids. Moreover, this material has several porous fibres ( shows the results of the chemical analysis of S. pungens and some lignocellulosic plants. A comparison of the data obtained for S. pungens with those reported in the literature for the other raw materials reveals the following remarks:The amount of extractives in hot and cold water is very high to compare with that found in hardwood (2–8%) and also in the non-wood sources like Date palm rachis (). Nevertheless, it is comparable to some annual and perennial plants such as Sorghum stalks (). The solubility in 1% NaOH (43%) is very high to compare with some sources of wood like Eucalyptus globulus () and comparable to that of Sorghum stalks () and other annual plants such as Wheat straw (). The amount of Klason lignin is comparable to typical values for annual and perennial plants, non-wood and hardwood sources, which are typically less than 20%. The amounts of holocellulose (71%) and α-cellulose (44%) were similar to those found in wood, non-wood sources like Rapeseed straw () and comparable to some known for annual plants (Sorghum stalks). These quite large contents allow envisaging the development of such crops, as a source of cellulose and lignocellulosic fibres for paper and/or cellulose derivatives applications. Finally, the ash content was found to be around 4.5%, which is much higher than that of wood (E. globulus 0.6%) and similar to non-wood plant such as Arundo donax (4.8%) (The elemental analysis of the raw material showed high proportion of carbon (42.89%), oxygen (42.21%) and hydrogen (5.76%). The analysis reveals a noticeable similarity with those found in the literature of lignocellulosic materials (The raw material was characterised by thermogravimetric analysis (TGA) as illustrated in The TGA curve displayed three main stages of thermal degradation. The initial drop in weight was observed from ambient temperature to approximately 120 °C and a weight loss of 5.5% was occurred. It was due to the release of water related to the humidity observed on the surface of hydrophilic lignocellulosic structure (). Then, the highest weight loss (88%) was occurred from 120 °C to 500 °C and attributed to the degradation of the lignocellulosic fractions of biomass (hemicellulose, cellulose and a part of lignin) (). Finally, there was little loss of mass due to the degradation of remaining lignin, over 500 °C.The pulping of S. pungens was carried out at different experimental conditions of cooking. Several relevant data were obtained, as illustrated by The delignification stage of S. pungens was carried out according to the experimental conditions used for A. armatus (). The concentration of sodium hydroxide and anthraquinone was 5% and 0.1% with respect to o.d. raw material, respectively.At 120 °C, for a yield close to 42%, the obtained pulp contains few uncooked materials and at 140 °C, the yield and kappa number were found to be 43% and 10.8, respectively. The uncooked materials were disappeared.We selected the following optimal cooking conditions: NaOH and anthraquinone concentrations of 5% and 0.1% with respect to o.d. raw material, respectively; cooking temperature of 140 °C; and cooking time of 120 min.The morphological properties of the pulp fibres obtained are presented in . These results show that the fibre width is close to that of fibres obtained from some other annual plants, while their length (0.606 mm) is significantly lower. The content in fine elements (expressed as the ratio of their length by the total length of the elements present in the suspension) is relatively high (24.3% in length).The WRV value for S. pungens fibres (120%) is significantly higher than that of unrefined pulp from softwood hardwood (90–100%). This property allows us to predict a high level of flexibility of the fibres, thus predicting good mechanical properties of the obtained paper. The degree of polymerisation (DP) of the pulp, around 939, is suitable for papermaking applications and it indicates that proper pulping conditions have been selected.Finally, the Schopper Riegler degree of the produced pulps was around 10 °SR. This value is similar to that of Posidonia oceanica () and very lower than that of other non-wood and annual plants like A. armatus (). Such a value predicts good drainage of the paper suspension even if it's in contradiction with fine content. However, the morphology of the fibre was characterised by high fine element (% in length) and low fine element (% in surface) which were 24.3% and 8.7%, respectively. This seems to be the cause of the reduced number and the high thickness of fibres which explains the very high drainability.Paper sheets from S. pungens were observed by SEM and EDS, as presented in . The physical properties are reported in ) show that the paper sample are quite homogeneous and resemble to those obtained from classical wood fibres. Then, the EDS analysis () showed that, as expected, the predominant elements in paper obtained were C and O. Lesser amounts of mineral elements of Si and Ca were also observed.Concerning the physical properties of the papers, the bulk value (1.86 cm3/g) is relatively good. This value is similar to that of A. donax L. reed pulp (The mechanical properties of the handsheets could be considered as very promising. Thus, the tensile test, the values of the breaking length (2.69 km), Young modulus (2.28 GPa) and specific energy (209.22 mJ/g) are quite good for handsheets prepared from an unbleached and unbeaten pulp. Even the elongation, burst and tear indexes present good values (close to 1.13%, 1.94 kPa m2/g and 4.48 mN m2/g, respectively), which were similar to that of date palm rachis pulp (Finally, the strength of the fibres (assessed from dry and wet zero-span breaking length) is about 10 km and 7.5 km respectively, which is quite significant and witnesses about the appropriate conditions of cooking. These data are somewhat similar to those known for other crops from annual plants such as A. armatus pulp (S. pungens, growing in large quantities in Tunisia, is a suitable source of lignocellulosic fibres for pulps and papermaking. After this study, determination of the chemical composition and optimisation of the soda-anthraquinone cooking process (5% NaOH, T |
= 140 °C, t |
= 120 min), the cooking of this residue gave lignocellulosic fibres with relatively good yield (43%). The isolated fibres were used to produce paper samples, which exhibited good mechanical properties. Therefore, S. pungens presently the unexploited residue, can be considered as a serious alternative raw material to be cooked, in order to prepare fibres for papermaking.A review paper on mechanical properties of flexural and impact test on textile reinforced engineered cementitious compositesThe following review paper thoroughly enables the readers to emulate the flexural behavior and impact test method for textile reinforced cementitious composites (TRCCs). The flexural parameters of ECCs corresponding to type of fiber material, fiber volume fraction, fiber volume length and strength parameters were discussed. For the textile reinforcement, the parameters like type of weaving pattern, mesh size, roving tex, fiber volume fraction and number of layers were discussed. Based on the tensile strength of textile fiber material, carbon, AR-glass, aramid and basalt fiber mesh were used for strengthening, repair works, plastering and reinforcement for thin elements. And also, the impact test performance on cementitious concrete composites is addressed as well as methods adopted for finding impact performance are discussed from the numerous literatures. This paper provides summary about the flexural and impact performance of textile fibers introduced into the concrete composites in different mesh sizes and fiber volume fractions. Generally, inclusion of textile fibers gives much more performance in tensile, toughness, flexural, energy absorption capacity and ductility in ECC composites. Many of the researchers have tried with different fiber materials for producing ECC and TRC composites separately, but only few of them used both have been discussed. Finally, from the literatures mostly artificial fibers are used for making textile reinforced composites.The advantageous such as workability, durability and mechanical strength of concrete make it an appropriate material among various building materials The bond between the concrete and reinforcing bars determines the nature of concrete The FRPCs are experiencing a rapid growth due to their numerous beneficial properties such as light weight, corrosion resistance and toughness and easier processing Glass Fiber Reinforced Polymer (GFRP), one of the artificial polymers emerges as an alternative to counteract the durability issues that arises with the use of conventional steel bars in the reinforced concrete The terminology “textile concrete” is a innovative building material used widely in the situation where this method is a replacement of reinforced concrete with textile Basalt can be formed as mesh with excellent chemical and mechanical properties than glass fiber This paper reviews about the type of fibers used in textile reinforced concrete and their properties. It also synopsizes the flexural and impact performance of the TRC.The various mechanical properties such as tensile splitting test, residual and flexural strength of concrete can be enhanced by adding the various types of fibers The organic fibers such as acrylic, polypropylene, polyolefin, polyethylene, polyvinyl alcohol etc., and inorganic fibers such as alkali resistant glass fibers are widely the used as reinforcement in synthetic fibers The current introduction in making fiber concrete is the use of glass fiber. The glass fiber has the excellent strengthening and toughness matrix and also extends the service life of the system. Glass Fiber Reinforced Concrete (GFRC) finds its application in the manufacture of façade plates, external claddings and other elements in which strengthening is necessary at the time of construction Introducing carbon fiber reinforcement in the concrete advances plenty of benefits in the construction when compared to conventional steel reinforcement. Carbon fiber in concrete improves higher tensile strength, lower specific weight and corrosion resistance in contrast to steel bars are the remarkable merits which enables the construction of thinner structures without compromising the load bearing structures and durability. Carbon filaments embedded in thermoplastic or thermosetting matrix are the carbon fiber reinforcement existing at present Basalt fibers are chemically stable and have extraordinary mechanical and thermal properties The increased effect on tensile behavior can be enhanced by the combination of steel fibers Polypropylene comes under the category of polyolefin group. The high heat resistant and hardness is observed in this fiber. The low tensile strength and modulus of elasticity was observed in polypropylene fiber and have high toughness The most widely used aramid fibers are kevalar-29 and kevalar-49 while the other varieties experience low compressive to tensile strength ratio which makes the fiber unsuitable when susceptible to cyclic loading The fiber reinforced sheets can be produced by the asbestos fibers in the Hatschek process which shows the good performance in variety of styles and forms in building material The impact on the environment due to high carbon emission and high cost of artificial fibers led to the use of natural fibers. Different sources of natural fibers are recycled and used in the construction as building materials which are eco-friendly, non-toxic, cost effective and renewable Nylon fiber reinforced concrete and Polypropylene reinforced concrete has its implementation in the non-structural and non-primary load-bearing applications The appearances of various fibers, classification of fibers and the properties of fibers are given in Ibrahim G.Shaaban et al, cast sixteen reinforced concrete beams of size 100 mm X 200 mm X 2000 mm with a span of 1800 mm. They were tested at mid span until failure under a single concentrated load. These 16 beams, were categorized into four groups (A, B, G, F) as mentioned in Square welded wire steel mesh (WWM), Expanded Metal Mesh (EMM), Fiber Glass Mesh (FGM) were used as reinforcement apart from steel as reinforcement. All sixteen beams were examined for three-point loading. Linear Variable Differential Transformer (LVDT) was placed at mid span to monitor the deflection at application of load. The results stated that the ductility index was higher for ferrocement beams with EMM than the beams with WWM. As the number of mesh layers was increased, ultimate load was high. Whereas the increase in the mesh reinforcement amount does not increase the ductility index. Ferro cements beams with EFC gave low ductility index due top the lower density of and strength of EFC. The ductility was affected by the type of mesh reinforcement. Ferrocement light weight beams strengthened with EMM may be a good substitute for traditional beams The effects of PVA fibers on concrete increases toughness, compressive and flexural strength of ECC composites slabs. Based on varying fiber length and volume fractions, the test results are showing better performance in flexural, better ductility value and energy absorption capacity for slabs Jamal Shannag et al, investigated the effect of short discontinuous fiber reinforced with varying layers of steel meshes in ferrocement thin laminates. The addition of brass coated steel fibers to the matrix increased its flexural strength capacity, energy absorption to failure, and greatly reduced average crack spacing and width, as well as preventing mortar cover spalling at ultimate load Shaik. Jeelani et al, carried out the capabilities of thin shells reinforced with textile reinforced concrete by both experimental and analytical method for flexure under four-point bending test for the specimen of size 1200 mm x250mm x250 mm reinforced with S-glass textile reinforced concrete with wholly replaced steel reinforcement. By changing the thickness of textile layers shows much better results in load and deflection values Qiao-chu Yang et al, studied the flexural capabilities of fine concrete sheets reinforced with short fibers and basalt textile fabric mesh of size 10 mm × 10 mm, the test results showed that use of proper amount of AR-glass fiber and carbonate whisker augmented the compressive, flexure and splitting of fine concrete Hussein et al, cast 6 beams and tested under four-point loading. The size of RC beams was 150mmX200mmX2000mm. LVDT was used to monitor the deflection. After testing and analyzing the results, it was shown that Textile Reinforced Mortar (TRM) is indecisive in augmenting flexural strength of RC beams and more efficacious in deflection. The Finite Element (FE) investigation was conducted out through LS-DYNA, a general-purpose finite element system. The quantity of layers and final state of TRM laminates were investigated for their effect on the flexural capabilities of reinforced beams. Based on this analysis, the enhancement of flexural capacity up to bond-stiffness coefficient of 225, the TRM composites are effective. Ultimate textile stress as well as deflection ductility were appreciably lowered to TRM end debonding beyond the bond stiffness co-efficient of 225. The addition of TRM layers were ineffective in augmenting flexural capacity for cases having bond stiffness co-efficient exceeding 290. TRM-strengthened beams could fail in flexure due to TRM end debonding or textile rupture, according to both experimental and numerical analysis. In contrast to cementitious mortar, polymer-modified cementitious mortar provided a strong bond between TRM layers and concrete substrate The flexural behaviour of geopolymer composite specimens made of Textile Reinforced Geopolymer mortar was investigated by Hiep Le Chi and Petr Louda. Differential aperture sizes of Basalt mesh and carbon mesh were used and applied for two layers. Specimens with carbon mesh have appreciable impact of flexural strength and deflection when compared with specimens with basalt mesh. The flexural strength of composite specimens with smaller mesh sizes were higher than that of specimens with larger mesh sizes. This was due to the fact that more yarns in the same thickness range contributed Using high strength reinforcing fabrics of glass and carbon roving, two samples were made. The textile reinforced concrete specimens of size 56 mm X 200 mm X 20 mm were cast and tested for flexural strength under three-point loading. The flexural strength for specimen assembled with AR-glass roving increased 1.55 times whereas, the specimens with carbon roving the increment was about 1.81 times. The result showed clearly that there were no appreciable differences between the specimens of carbon roving and glass roving. Hence, high residual load bearing capacity is one of the advantages of using TRC Dharmesh Bhagat carried out flexural test under two-point loading. The beams had the dimensions of 700 mm X 150 mm X 150 mm. Fourteen beams were cast in which two beams were tested as reference beams and the remaining twelve beams were strengthened with polyester and polypropylene with single, double and triple layers. The load carrying capacity of the beam strengthened with polypropylene was increased while the deflection decreased. The beam strengthened with polypropylene of single and double layer roughly had the same load carrying capacity. Both polypropylene and polyester beams strengthened with double layer had high load carrying capacity with low deflection. The increase in stiffness due to increased fabric layers caused this difference. It was concluded that, beams of fabrics with single and double layer had good flexural performance. But, double layer of fabric was the best option being an alternative for strengthening of beam by retrofitting Ultra high performance concrete reinforced with PVA fibers and 2D textile reinforcement was carried out to produce thin plates gives better deflection rate which was measured by computer based softwareNahum Lior contrasted the flexural capability of carbon textile reinforced concrete (TRC) beams to that of traditional steel reinforced concrete (SRC) beams. Five concrete beams were cast with one reference beam made of conventional SRC reinforced concrete and the remaining four beams made of TRC with carbon fabric. The behaviour of flexure was carried out under four-point loading and LVDT was used to monitor the deflection. The flexural performance and mode of failure were dominated by shear reinforcement configuration. The U-shaped fabric TRC beams evinced high load bearing values without delamination while, the steel cage rebars evinced low load bearing values with intense delamination. Since the U-shaped fabric was strongly connected with mortar, this difference occurred. Hence, the mechanical anchoring of fabric with mortar also play a major role in the mechanical (flexural) performance of TRC The one-way slab with dimension (1500 X 500 X 50) mm were examined for the effect of removing the weft yarns on the flexural performance under four-point loading. The carbon fabric of eight layers with different removing percentage of weft yarns (50%, 67%, 75%). Four slabs were cast in which one slab is cast without removing yarns. The overall behaviour of TRC was enhanced by removing the weft yarns. The removal of weft yarns not only enhanced flexural behaviour of TRC but also compression concrete strain due to the improvement of matrix performance by the enhancement of layer openings through penetration. Volumetric ratio was absolutely reduced by the removal of weft yarns. From the results, it was concluded that the flexural strength of slab with high removal percentage of weft yarns had great value when compared with other slabs The flexural performance of TRC and Textile Reinforced Geopolymer (TRG) with Polyvinyl alcohol (PVA) reinforced with hybrid AR-glass fiber was presented by Faiz Uddin Ahmed. Three plate specimens for each series of size (15 X 40 X 400) mm were tested under three-point loading. The flexural strength of AR-glass TRG was greater than TRC. The inclusion of PVA fiber appreciably enhanced the absorption of energy during deflection Four-point bending test was done by Mana Halvaei et al., to study the flexural behaviour of the specimen. The dimensions were (500 X 100 X 25) mm. Different sizes of mesh such as 0 mm, 2 mm, 4 mm, 6 mm, 8 mm, 10 mm and 20 mm were used. Flexural strength and toughness of the fabric strengthened concretes (TRC) drastically elevated through lowering the fabric mesh sizes from 20 mm to 2 mm. The quantity content material of the reinforcing fabric is more vital than the fabric size Doo-Yeol Yoo et al, studied about the influence by type of textile reinforcement in characteristic flexural response of TRC beams under static and impact loads. The usage of polymer coated textile reinforcement in beams gives much more flexural strength, toughness and load carrying capacity. The test was conducted based on ASTM standards with four-point bending test as shown in . When comparing with the T-2D textile reinforcement gives higher toughness and residual load carrying capacity The flexural performance of steel textile reinforced concrete was investigated by Jungbhin You et al., by producing one reinforced concrete specimen and six TRC. Textile layout form, textile mesh type, textile reinforcement amount and textile hook type were the four different variables used for manufacturing TRC specimen. Size of the specimen to carry out flexural strength under four-point loading was (1500 X 150 X 20) mm. LVDT was used to monitor the deflection. When the mesh of the fabric is made turned into smaller, the number of contacts with inside the warp and weft roving of the fabric increased, which avoided slippage among textiles Using Charpy impact test system, the impact behaviour of the specimens with composites such as AR glass mesh, Basalt mesh, PVA weave textiles and polymer modified cementitious matrices was studied by Esma Gisem Daskiran. The laminated specimens with 4,6 and 8 (different) textile layers are called composites. Three specimens on each reinforced composite series were tested. The specimens were immersed in alkaline condition for curing of 28 days. The quantity of textile layers is severe factor in the view of absorbed impact energy. Specimens with basalt series had not increased which was due to the immediate increase in the sample thickness. As the amount of textile layer increased, the specific impact strength also increased. Specimens with PVA series had the maximum impact strength when compared with the result values of other series. This was because PVA is a ductile textile with plain weave structure with multifilament yarn A new type of testing methods was adopted under impact load to find energy absorption of textile mesh reinforced concrete slabs in which the specimen size of 610 × 610 × 30 mm was cast and inserted a fabric in the middle of the specimen at a depth of 15 mm. Quad axial fabric were laid under the slab at a 15 mm depth. Impact test were carried using the steel frame and drop weight of 8.4 kg heavy impactor was used. As measurement technique, a stereo high-speed camera which is connected to an software system GOM ARAMIS used for measuring velocities of impact Doo-Yeol Yoo et al, investigated the impact of fiber orientation on the flexural behaviour of UHPFRC under quasi-static and impact loadings. As per ASTM standards, the quasi-static tests were performed under which the beams with better fiber content in the direction of the tensile load had greater flexural strength, optimized deflection capability, and toughness. The drop weight impact test machine as shown in were used for finding energy absorption capacity Dong Z et al, investigated high zirconium alkali-resistant glass fiber combined with macro polypropylene and steel fibers were used in the concrete slab to obtain impact performance with varying volume fractions using drop hammer test as shown in which gives better results of 0.75% volume fraction tested under second impact A three-pointed bending set up was used to investigate the mechanical interaction of textile reinforced aerated concrete sandwich panels under static and low velocity dynamic loading. The core material consisted of two types of aerated concrete: Autoclaved Aerated Concrete (AAC) and Fiber Reinforced Aerated Concrete (FRAC). Two layers of AR-glass textiles and a cementitious binder made up of skin coat. Energy absorption ability of ductile skin brittle core (TRC-AAC) and ductile skin ductile core (TRC-FRAC) composites were evaluated. The dimension of the specimens to be tested under drop weight impact test setup was (50 X 50 X 250) mm. At different drop heights, the influence of impact energy on material properties were calculated. TRC-AAC’s and TRC-FRAC’s energy absorption improved by up to 20 to 50 times as compared to plain AAC and FRAC core materials due to skin effect. Under both loading modes, the outwardly bonded textile layers greatly enhanced the mechanical properties of the light weight low-strength aerated concrete core. The strength under dynamic loading was four times higher than the strength under static loading A novel hybrid that was created by mixing glass and basalt textiles to increase impact strength over individually reinforced counterpart was discussed by Smitha Gopinath et al, TRC slab specimens had measurements of (450 X 450 X 50) mm. Instrumented impact machine was used to analyze the impact behavior. Because of high energy absorption capacity, TRC slab with glass textile suffered the least damage which was followed by basalt TRC. In case of hybrid TRC, the spalling of textile from the binder was more. Hybrid TRC slabs had the maximum impact resistance. As opposed to the use of single textiles, the combination of two textiles demonstrated the highest impact resistance regardless of energy levels. When compared with the basalt reinforced slabs more energy is absorbed by the glass textile reinforced concrete slab Hence, from the results obtained, the form of textile used in TRC had a major effect on the failure modes of TRC slabs under impact loading Hakan Nuri Atahan carried out experiments to see how the fiber volume fraction and matrix properties influenced the mechanical response of 15 mm thick short cut PVA fiber reinforced cementitious composites. Different fiber volume ratio (0.5%, 1%, 1.5%, 2%) with two different water cement ratio (0.25 and 0.35) the beam samples of eight different mixtures had the dimension of (350 X 50 X 15) mm was used for Charpy impact test to measure the impact resistance. From the results obtained, he concluded that the impact resistance of composites were influenced by fiber content and matrix strength. In comparison to specific fractures under static loading, combined effect of water cement ratio and fiber volume fraction on impact resistance was more drastic. Higher values of water cement ratio and increased fiber content greatly improved the energy absorption ability of these composites when subjected to impact loads Xiangming Zhou et al., investigated fracture properties and impact studies of short discrete Jute Fiber Reinforced Cementitious Composites (JFRCC) with various matrixes in order to build low-cost natural fiber reinforced concrete and mortars for construction. Three concrete mixes such that JFRCC with GGBS/PC matrix, JFRCC with PFA/PC matrix and plain concrete with GGBS/PC matrix were prepared for mortar panels of size of (200 X 200 X 20) mm. During impact test, JFRCC mortar panels had not crumbled into pieces whereas, PC mortar panels crumbled into pieces since, they were brittle. Hence, JFRCC mortar panels made of PFA/PC matrix acquired maximum impact resistance The dynamic behavior of steel fiber reinforced concrete plates under impact loading in terms of displacement, velocity, and acceleration was investigated with the assistance of a drop weight impact test. Eighteen specimens with dimension (600 X 600) mm having three varying thickness such that 20 mm, 25 mm and 30 mm with three varying fiber quantity such that 0.25%, 0.75% and 1%. It was discovered that with a decrease in fiber content from 0.5 to 0.75 percent, there is a slight improvement in energy absorption when the aspect ratio of fibers is between 50 and 75. When the aspect ratio of the fiber is100, there is a significant improvement in absorption of energy with a steel fiber content of 1% Impact test on plain and fiber reinforced oil palm shell concrete (FROPSC) panels using a drop hammer were presented by Kim Hung Mo. Different steel contents (0.75%, 0.9%, 1%) and polypropylene fiber (0.1%, 0.5%, 1%) with crushed and uncrushed OPS were taken as variables. The panels of size (600 X 600 X 50) mm were cast and tested under drop hammer impact test. The hybrid-OPSC specimen with 0.9 percent steel and 0.1 percent polypropylene (PP) produced excellent impact energy of about 17 kJ, which was 60 times higher than the plain OPSC specimen. From the results it was concluded that the impact ductility of concrete can be determined by the effect of fiber bridging on post-crack impact energy absorption Efrat Butnariu studied the impact behaviour of fabric-cement based composites tested under A drop weight three-point bending testing machine. The use of short fibers like PVA, polypropylene, polyethylene and AR glass fibers in producing cement-based composites. In addition to this 2D fabrics were reinforced inside the composites to obtain impact test results. Introduction of fabrics shows high loading under impact behaviour To examine the behavior of textile reinforcement with carbon fiber for strengthening reinforced concrete(RC) slabs under repeated impact loads, five RC slabs were cast in the size of (1500 X 1500 X200)mm. Out of the five specimens, three specimens were reinforced with three distinct carbon textile reinforcements embedded in a 2 cm fine grained concrete layer and subjected to same striking velocity impact loads and the balance two specimens were left unstrengthened and tested under various impact velocities to better understand the failure mechanisms of RC slabs under impact loading. With increase in carbon textile reinforcement ratios, the carbon textile reinforcements got upper hand, especially for the subsequent impact test stages. Finally, the increased striker velocity resulted in a noticeable difference in specimen damage level. The test results demonstrated that the carbon textile reinforcement is extremely effective at enhancing the specimens' impact capacities Numerous studies have stated that, upon the addition of fabric or textile layers the strength of the reinforced specimens increased significantly. According to Ombres for specimens with one, two, and three layers, the load capacity increased up to 16%, 33% and 40% respectively The type of fabric or textile used in the concrete also enhance the strength of TRC members. In comparison to carbon-FRCM specimens. Ebeadet al. found that PBO-FRCM specimens have higher ductility and energy absorption According to the investigation by Yin et al., there was a significant impact on the structural performance of carbon-TRC-strengthened RC beams due to corrosion exposure. The fatigue life of reinforced beams shortened due to corrosion TRC anchoring helps in enhancing the performance by keeping away or holding up the debonding mode of failure. In end U-anchorage, the TRC strengthening system has U-shaped fabrics provided as an supporting layer at both the ends and the failure occurred due to slippage While increasing the bond length, the bond capacity of TRC specimens increased. According to D'Ambrisiet al. The practicality and efficiency of TRC were demonstrated in previous studies are discussed in this paper. TRC has arose as a modern composite with numerous potential applications in non-structural, structural materials that include thin and slender elements repair and strengthening of original structural member especially through seismic retro fitting. Comparing with Reinforced Cement Concrete (RCC) the cracking control techniques are far more superior in TRC along with that the flexural capacity and shear capacity of RCC members are generally shown good improvement while applying fabric coating. Based on the fabric type the strength of the concrete member may vary. It is found that the addition of fabric (fiber) improved the strength of concrete members by bridging the gap upon the application of load on the member. Synthetic fibers exhibited more strength when compared with natural fibers. When the number of layers of TRC is increased, the strength gain is not proportional. The textile reinforced concrete consumes less concrete, increase the life of old concrete and it is less costly compared to RCC structures. It is also highly corrosion resistance and develops light weight structures. Due to these uses they are applied in the bridges, load bearing shell structures, sandwich walls. Overall, textile reinforced concrete provide strength in the existing structure and retrofits the old structure such that promising a sustainable future in concrete.The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Porosity, hydrogen and bifilm content in Al alloy castings► The presence of bifilms in aluminium can be quantified by reduced pressure test. ► Bifilm index can discriminatingly assess aluminium melt quality. ► Number of bifilms is directly related with the number of the pores. ► Metal quality can be quantified by using both bifilm index and number of pores. ► It is predicted that very high ductility will be achieved in the absence of bifilms.In Al alloy castings, it is often not microstructure, but entrained defects in the form of bifilms that dominate mechanical properties. To assess their effect, this study targets a new quantification of metal quality, measuring the lengths and number of bifilm defects as determined from the microstructural section of a reduced pressure test (RPT) sample.In recent years, it has been shown that the reliability of mechanical properties of aluminium alloy castings can be significantly increased by simple alterations to the design of their filling systems The presence of bifilm cracks in the metal has numerous consequences: the unbonded interface can be opened by the inward diffusion of gas, thus initiating gas porosity. Alternatively, they can be pulled apart by shrinkage stresses to initiate shrinkage porosity or hot tearing. Finally, in the solidified casting they act to greatly reduce mechanical properties, particularly ductility and fatigue.Because the bifilms are not easily seen, sometimes as thin as a few tens of nanometres, in contrast with their extensive areas, sometimes measured in millimetres or even centimetres, their presence has not generally been suspected. This lack of awareness of their presence is despite the fact that they can be easily seen when opened up in a reduced pressure test (RPT). In the Al casting industry the test is widely used. It is made the central technique of this study. The test involves the solidification of a small sample of melt under a partial vacuum (100 mbar). Although the test has been traditionally used only as an indication of hydrogen content (by measuring the density and correlating it to the hydrogen level), it has become clear that hydrogen porosity cannot nucleate (either homogeneously or heterogeneously) without the presence of bifilms Recently, the authors have developed the RPT as a quality test that is capable of identifying and classifying bifilms, subject to some assumptions that are made clear in the paper The total length of bifilms is clearly an important parameter. However, an additional parameter of similar potential importance is the total number of bifilms. These two independent parameters are considered in this study. Naturally, the total length of bifilms divided by the total number gives the important parameter the average length of bifilms in a particular melt. As will be seen, this value can vary widely.The previous studies of the authors were re-evaluated. The details can be found in the following Refs. ) weighing about 6–7 kg were melted in an induction furnace at 700 °C (2250 Hz, 15 kW, 500 V, open air, clay-graphite crucible, melted within 10–15 min) and the temperature was gradually increased to 800 °C and 900 °C.) were produced by casting with two different devices. One device could deliver the melt quiescently (with relatively little surface turbulence) while the other was a turbulent pour. These ingots were melted using standardised conditions in an induction furnace at 700 °C (2250 Hz, 15 kW, 500 V, open air, clay-graphite crucible, melted within 10–15 min).), 75 kg of two charges from same batch were melted in a resistance furnace at 750 °C. A rotary degassing unit (Foseco XSR Rotor Ø135, at 500 rpm) was used to control the melt hydrogen level. Three levels of hydrogen were selected: 1, 2 and 4 L/kg Al. One of the melts was upgassed to 4 and then gradually degassed to 2 and 1 levels. The other melt was degassed first and then upgassed.In all the experiments mentioned above, 80 g of melt was poured into a sand mould () that was placed in the chamber of the RPT unit and allowed to solidify at 100 mbar. These samples were then used for bifilm index measurement. In addition, 10 cylindrical bars were cast for mechanical testing (The pore morphology change of RPT samples with increasing temperature is illustrated in . The figure also contains the information such as shape factor, bifilm index and hydrogen level.The change in the shape factor with the increasing temperature is given in The interiors of the pores from the RPT samples were studied by SEM (The comparison of metal quality of two different casting devices which was used to produce ingots of A380 is given in The bifilm index change with hydrogen level at different treatment sequences of A356 melt is given in shows the relationship between bifilm index and the number of the pores for all three alloy systems.The relationship between the mechanical properties and bifilm index of A356 castings is given in One of the particularly intriguing observations was the pore morphology of sectioned surfaces of the RPT samples. As seen in , as the melt temperature was increased, the pore morphology was altered from a round to a thin and elongated shape. This can also be seen in where the average shape factor was decreased from 0.5 to 0.3 as the temperature was increased. A similar result was found by Laslaz and Laty Once a bifilm is introduced into a melt, because of the strong bulk turbulence within the melt, often as a result of a pouring action, the bifilm will usually be crumpled into a compact form. This bifilm (which is just a few nanometres thick) may at this stage be relatively harmless because of its small overall size and would not be easily detected. In the reduced pressure test, when the vacuum is applied, the volume of gas trapped inside the bifilm will tend to increase and this would drive the tendency of the bifilms to unravel, straightening with time.The diffusion of hydrogen into the gap between the bifilm is a second factor driving the growth and straightening of pores. If the pore consisted of an inert gas in a pure liquid with no surface film, pore growth would only cause an increase in size towards a spherical geometry, which would be perfectly reversible if the hydrogen level were to subsequently fall. However, in the ratcheting action (a), instead of growing spherically, the bifilm tends to open and at the same time extend its area a little. Such an extended area is oxidized immediately by the residual air between the bifilms. The process is, of course, irreversible. Thus, as the bifilm circulates in the liquid, its expansion towards the top of the melt, followed by its contraction towards the bottom of the melt, effectively constitutes a breathing, or panting action (c. This is thought to be the mechanism of the elongated pore formation that is observed in the sectioned surfaces with the increase in temperature and time.The stabilisation of the crack-like form by high temperature, despite the varying hydrogen content, possibly suggests that the bifilm diffusion bonds together to some extent, consuming its entrained oxygen and possibly its nitrogen to form bonds of oxide and nitride between the original oxide films as was shown by Raiszadeh and Grifitths , the hydrogen levels are quite different.SEM studies of the pore interiors support the mechanism of their growth by the unravelling of bifilms. In a crack-like pore appears to show a bifilm that is only partially opened in between the dendrites. Only traces of the originating bifilm were found; the remainder was assumed to have been sucked deep into the dendrite mesh.The observations of crack-lined oxides suggest that the use of density alone as an attempt to quantify quality by RPT could be misleading One of the earlier examples of studies using the bifilm index was the comparison of different casting devices , the more turbulently the melt was transferred, the higher the bifilm index. Another similar observation was found during the rotary degassing experiments Besides the bifilm index, the number of pores is clearly also a dominant factor affecting metal quality.Since homogeneous or heterogeneous nucleation of pores is difficult or impossible in aluminium, the obvious non-nucleation process will be the opening of bifilms. Thus, the number of the pores has to directly correspond with the number of bifilms. Therefore a plot of bifilm length versus number of the pores () reveals the population of data for the alloys A319 and A380 to have a linear lower limit corresponding to 0.5 mm average pore size. A line corresponding to an approximate maximum size 2.0 mm can also be defined, suggesting an average size in the region of 1.0 mm, plus or minus a factor of 2. On the other hand, the particular batch of A356 provides significantly different data indicating an average pore length corresponding approximately to 0.10 mm.This nearly ten fold difference in the size of bifilms in these two alloy types is not easily explained at this time, and may be more strongly influenced by the different processing of the melts by the alloy manufacturer, although in this case all alloys were sourced from the same provider. It might be supposed that the smaller bifilms in the A356 alloy were a fundamental feature of this particular alloy, but a study by Fox and Campbell has interesting extreme categories as follows:It appears that scenario I is probably not common. All the data appear to lie close to the linear relationship. This is an important conclusion suggesting that all pores on the sectioned surface of RPT samples have almost same size.From the degassing and upgassing studies with A356, it was interesting to observe that as the hydrogen content was increased by the upgassing process the bifilm length had increased (a). At first, it may be assumed that as the hydrogen was increased, pore sizes would increase and thereby leading to a higher bifilm index. However, the same scenario applied to the degassed melt. As the melt hydrogen level was decreased, the bifilm index was increased (b). The only common parameter with the increased bifilm index was the melt treatment sequence which was the indication of entrained defects during rotary treatment. It is important to note that the average pore sizes were similar (around 165 μm) for all cases which indicated that only the number of pores was increased. shows how the quality field might be divided into target zones. A high quality melt will target the box in the bottom left hand corner of the figure, with separate limits on numbers and lengths of bifilms. A melt required to be ‘gassy’ with well-distributed fine pores might chose a zone as illustrated by the boxes in the centre of the graph. During the early development of the Cosworth Process one of the authors used RPT conditions defined by only 3 pores of maximum size each 1 mm (a box of dimensions 3 by 3 in A work on an alternative control chart was also carried out where mechanical properties would be included. As seen in , it was aimed to establish different regimes on the chart that would correspond to certain ductility levels.The comparison of mechanical properties of A356 is seen in . Clearly, the tensile strength falls with increasing bifilm content as would be expected. On the other hand, although elongation also decreases, there was high scatter of elongation at low bifilm index values. The results indicate that the potential for high elongation exists with cleaner metal, but the parameter is highly sensitive to the presence of even relatively few bifilms. It may be that the orientation of bifilms in a tensile test bar with respect to the stress may affect the ductility significantly.Determination of metal quality by using both (i) bifilm index and (ii) number of pores appears to be a potentially useful quantification technique.It is predicted that very high ductility will be achieved in the absence of bifilms.Experimental and numerical investigations on large angle high strength steel columnsThe stability of columns made of large angle profiles in high strength steel and subjected to centric and eccentric compression loads is investigated through experimental tests, numerical simulations and normative approaches. The aim is to widen the knowledge on such columns and so to complement previous experimental studies that focusing on smaller angle profiles and lower steel grades. The experimental campaign consists of twelve tests. Accompanying numerical studies have been carried out, considering relevant geometrical imperfections as well as geometrical and material non-linearities. Finally, an assessment of the design member resistance according to Eurocodes has been performed. It is further discussed.Angles profiles have been used since the very beginning of steel construction due to their easy transportation and on-site erection. They are extensively used in lattice towers and masts for telecommunication purposes or electric power transmission, as well as in a wide range of civil engineering applications such as buildings and bridges. They are also used to strengthen existing structures. Recent developments in transmission towers, like the installation of new 380-kV lines or high-voltage direct current (HVDC) lines, lead to the need for a wider application of large angle sections made of high strength steel. In Europe, lattice towers are designed to different standards. Generally, steel lattice towers are designed to Eurocodes and in particular to EN 1993-3-1 [] providing general rules and EN 1993-1-8 [] providing rules for connections. But in the specific field of overhead electrical lines, lattice towers are designed to CENELEC standard EN 50341-1 [] which provides rules sometimes diverging from those proposed in the Eurocodes. However, in all preceding codes, rules concern angles subjected to pure compression and bending effects due to eccentric connection in one angle, being considered through a modification of the buckling length. Eurocode 3 proposes a so-called “general method for lateral and lateral torsional buckling of structural components” which could possibly be applied to angle columns under compression and bending. In contrast, American codes [] have issued design specifications for single angle members subjected to general loading conditions including compression and bending.Compression tests on large angle sections ranging from L125 × 125 × 8 to L200 × 200 × 14 in high strength steel (HSS) S420 were conducted in Tsinghua University in Beijing []. The tests were carried out on axially loaded pin-ended columns in order to define global-local buckling interactions since cross-sections were in many cases class-4 ones. Tests on L70 × 70 × 7 profiles were performed at NTUA in Athens [], where the effects of eccentric loading were studied. Compression tests on L80 × 80 × 8 and L120 × 120 × 12 profiles were carried out at ] in which the boundary conditions were varying from clamped supports to supports allowing in-plane or in- and out-of-plane rotation. A tests series on L50 × 50 × 5 profiles were carried out at the Technical University of Braunschweig [] with various specimen lengths and end support conditions, while the load was introduced eccentrically through one bolt M12 in one leg. A series of compression tests was performed on equal angle profiles loaded through two bolts in one leg at the University of Windsor, Canada []. The cross-sections were L64 × 64 × 4.8, L64 × 64 × 6.4 and L76 × 76 × 6.4 with steel material S300. The bolts were fastened snug-tight or preloaded. Further tests were performed at the same University, where the angles were loaded through one bolt []. The cross-sections were equal angle profiles L51 × 51 in three thicknesses, 4.8, 6.4 and 7.9, L64 × 64 in two thicknesses, 6.4, 7.9 and L76 × 76 × 6.4. The material was also steel S300, as in the previous tests []. All nominal dimensions of the above-mentioned cross-sections of tested columns were in mm.In order to study the stability behaviour of steel columns from HSS (S460M) angle cross-sections subjected to compression and bending, twelve (12) buckling tests on such columns have been performed. The experiments have been limited to high strength steel only, given the fact that a number of compression tests on angles with lower steel grades were already available. The test campaign has been realized at the “Laboratoire de Mécanique des Matériaux et Structures” at Liège University. The selection of the specimens, the details about the experimental campaign such as measurements before and during the tests, as well as the test results, are presented and discussed in this article. The tests have been accompanied by numerical simulations performed with the full non-linear finite element software FINELG, considering relevant imperfections as well as geometrical and material non-linearities. The numerical results have been compared and validated with the experimental ones. Finally, a comparison between the experimental ultimate member resistances and the Eurocode predictions for centrally and eccentrically loaded columns is presented and discussed. The experimental campaign and the numerical simulations are part of the European project ANGELHY, an RFCS-supported research project dealing with lattice telecommunication and transmission towers and lattice girders from angle sections.For the experimental program, two profiles from large angle cross-sections made of S460M steel grade have been selected. For each profile, six (6) column tests have been performed with three (3) different lengths per profile and two (2) positions of load application for each length. The selected points are (see ) the centre of gravity (G), which corresponds to pure compression in the angle cross-section and the intersection point of minor principal axis v-v with the middle line of the leg thickness (P2), which represents the position of the connecting bolt for angles in structures. summarizes all the details about the specimens. The name of each specimen consists of two numbers Sp## (e.g. Sp12):the first number indicates the profile: 1 for L150 × 150 × 18 and 2 for 200 × 200 × 16;the second one is the serial number of the specimen (1 to 6 per profile).For all tests, constant dimensions have been selected for the end plates welded at the extremities of the angle members, in order to simplify the placement procedure of the specimen in the test rig. Therefore, the position of the applied load is always the same for the machine and the eccentricity is introduced by moving the profile on the end plates. The steel grade of all end plates is S355 and not S460M as for the profiles. The welds have been designed according to EN 1993-1-8 []. For all specimens, the minimum required weld thickness is 6 mm, except for specimens Sp11 and Sp21 which require a minimum thickness of 8 mm. shows the details of such end plates on which the specimens have been welded.The actual geometrical dimensions of each angle section – the width (bi) and the thickness (ti) of each leg – have been measured at 3 points along the member: at 1/4, 1/2 and 3/4 of the angle member length (L). The mean values of the measurements are reported in . The length and the load eccentricity of each specimen has been also measured and reported in Two displacement measurements (M1 & M2) on each external face (Face A & Face B) and along the column length have been performed so as to evaluate the initial imperfections of the specimens. shows the details of the set-up. Due to the end plates and the measurement system itself, it was not possible to take measurements quite close to the ends of the specimens. As a result, all the measurements start 140 mm after the top end plate and finish 140 mm before the bottom one. A measurement has been taken every 50 mm along the column. It has been quite reasonably assumed that the columns are straight close to the end plates (140 mm).As the chariot supporting the displacement transducers was moving onto a horizontal guiding bar, a small rotation of the metric system was occurring; this one has been measured with an inclinometer, so allowing correcting the measurements accordingly. In addition, the column was not perfectly parallel to the set-up. For those reasons, the first correction to achieve concerned the non-parallelism and the rotation of the metric system; it is based always on the position of a reference cable. Then, a second correction has been done in order to have zero imperfection at the extremities of the column. Finally, a horizontal movement of the curve has been achieved so that the first measurement is well reported at 140 mm of the end plate. This procedure has been followed for face A and face B. All the results from the measured geometrical imperfections can be found in Ref. []. An example of the initial measurements and the evaluated geometrical imperfections of Specimen 15 are presented in An accurate comparison between the actual measured imperfections of the specimens and those assumed in the Eurocode is quite difficult to do. The European norm EN 1090-2 [] prescribes that the deviation from straightness should be Δ ≤ L[mm]/750 while in prEN 1993-1-14 [] it is stated that 80% of the geometric fabrication tolerances given in Ref. [] should be applied. This leads to an initial bow imperfection of magnitude approximately equal to L[mm]/1000 and usually a deformation shape similar to the first member instability mode is assumed, however in reality the shape is more complex. For this reason, only a rough comparison can be done at this level (see ) through the evaluation of an experimental estimated value |Max|imperf obtained by taking into account the maximum value [M1CA, M1CB, M2CA, M2CB] and by assuming that it is the same in both faces:|Max|imperf=max{M1CA, M2CA, M1CB, M2CB}·2M1CA is the M1 maximum final corrected measurement on face A for specimen i;M2CA is the M2 maximum final corrected measurement on face A for specimen i;M1CB is the M1 maximum final corrected measurement on face B for specimen i;M2CB is the M2 maximum final corrected measurement on face B for specimen i.Then, it can be concluded that measured imperfections are smaller for all specimens than the geometrical tolerances prescribed in European regulations.Coupon tests have been performed in accordance with ISO 6892-1:2016 []. The samples for the tensile tests have been extracted from one of the extremities of the angle member (see ) after the buckling tests, based on ISO 377:1997 [ shows the strain-stress curves obtained from few tensile tests and provides the characteristic values for all. The yield stress fy (engineering stress) is determined by the value of the yield plateau in the curves and defers from the upper value yield stress ReH. It may be also observed that while the actual ultimate stress was for all specimens above the nominal values, this was not true for the yield stress.The tests have been carried out in an Amsler 500 testing machine, with a compression capacity of 5000 kN. The specimens are pin-ended in the testing rig, since the rotations about the minor and major axes can develop freely, but no twist or warping is able to occur at the extremities. During the tests, the following displacements illustrated in the vertical displacement C1 (using two transducers, one at the front and one at the back side of the specimen);four horizontal displacements C2, C3, C4 and C5 at the mid cross-section (1st position);four horizontal displacements C6, C7, C8 and C9 at the lower cross-section (2nd position).All the displacement transducers have been placed 30 mm from the edges/corner of all cross-sections and profiles. The set-up allowing the record of those displacements is illustrated in . In addition, four strain gauges (I1 to I4) have been placed at the mid-height cross-section of each column, in order to check local yielding. The strain gauges have been positioned as close as possible to the edges of the cross-section, taken into account the curvature of the latter.The displacements of the corner of the angle (O point) as well as the twist of the cross-section, that are reported in the graphs of section 4.1, have been evaluated using the following formulae (for the definition of the axes and symbols, see θ=12(atan(C3−C2d)+atan(C5−C4d))·1000[mrad]where d = 90 or 140 [mm] for L150 × 150 × 18 or L200 × 200 × 16. The formulae are given for the middle cross-section, but they may also be used for the lower one, by replacing C2, C3, C4 and C5 by C6, C7, C8 and C9 respectively.To transform the displacements from the geometrical axes to the principal ones (see right), the following equations have been used, for an angle equal to 45°.The axial deformation of the specimen has been evaluated as the mean value of both vertical transducers.The results of the experimental tests are presented below through graphs and tables. Numerical simulations of the tests have been performed with FINELG software, taking into account the actual dimensions, length, imperfection and material properties. A comparison between the ultimate resistances obtained experimentally and through EN 1993-1-1 predictions has also been achieved for centrally and eccentrically loaded columns. show the load-axial displacement (shortening) curves for the profiles L150 × 150 × 18 and L200 × 200 × 16 respectively. All the measurements (initial geometrical imperfections, rotations, strains and deflections) for each specimen are available in Ref. [Both figures indicate that the results obtained by experimentation are in line with the physical expectations (for instance, the influence of the member length and of the eccentricity on the member stiffness and resistance properties). This seems a priori to validate the initial selection of the different parameters in the test campaign.Amongst the specimens without nominal loading eccentricity, three (Sp11, Sp15, Sp25) showed nearly zero and three (Sp13, Sp21, Sp23) very small deflections transverse to the weak axis (see ). This may indicate that only three specimens were almost loaded without any eccentricity, while the other three had probably some unintentional eccentricity resulting from installation tolerances as explained later. Nevertheless, for all the tests, the deflections along weak axis increased significantly with the load until failure was reached by weak axis buckling; towards the heel of the cross section for Specimens Sp13, Sp15, Sp25 and in the opposite direction for Sp11, Sp21 and Sp23. Concluding, specimens Sp11, Sp13 and Sp15 failed in a pure flexural buckling mode (see (a)), while in specimens Sp21, Sp23 and Sp25 twist rotations were recorded (see ), and in addition to weak axis deflections indicating a flexural torsional buckling mode.The eccentrically loaded specimens were initially subjected to compression and strong axis bending. At low load levels, the deflections transverse to the strong axis were high and very small in the other principal direction (see ). However, this was opposite to the tendency of the angles to fail by weak axis buckling. At higher load levels, deflections transverse to the weak axis grew quickly and prevailed at failure (in , the red cross-section corresponds to the ultimate load and the blue cross-section is after buckling). In the specimens Sp22, Sp24 and Sp26 these deflections were accompanied by significant twist rotations indicating clearly failure with a flexural-torsional buckling mode (see (b)). On the contrary, twist rotations were small for specimens Sp12, Sp14 and Sp16 indicating a mixed mode between flexural and flexural torsional buckling (see (c)). The most stressed mid-height cross section was subjected to compression and bi-axial bending. In fact, strong axis bending was primarily due to the eccentric loading and weak axis bending due to second order effects.The absence of visible local buckling in all specimens should be also mentioned. presents for each specimen the measured failure load (Nexp) as well as the deflections about minor (vO) and major (uO) principal axes and the twist at the mid-height cross-section, all at the failure load. The sign of the deflections and the twist is in accordance with Subsequently, numerical simulations considering relevant imperfections as well as geometrical and material non-linearities were performed and compared with the results of the experimental tests. The numerical analyses were performed with FINELG non-linear finite element software [] using beam elements. The choice of beam elements is acceptable and justified from the fact that no local buckling took place during the tests. The FINELG software was selected due to its wide and successful use in the past for such a type of analyses []. Only the column has been modelled while the end plates at the extremities have been considered indirectly. Each column has been meshed in twenty beam finite elements along the member length. This is an optimal mesh as the difference of member’s response (ultimate load and deflections) is less than 1% ether the member is meshed in fifteen elements or thirty. The length (L) of each column has been increased by 107 mm, what corresponds to the thickness of the end plates of the specimens as well as the connection plate, so as to simulate the actual buckling length (Lcrit) of the column (length between the zero moment levels in ). The columns were assumed as pin-end members with free rotations at their extremities, except the rotation that leads to torsion along the length axis, which was blocked. All the other DOF at the extremities were blocked, except ux of the node of the applied load. Therefore, the experimental boundary conditions were rather well represented by the model.The FINELG finite element analyses adopting the GMNIA method were performed considering:an initial member imperfection (shape and magnitude in accordance with the measured ones);residual stresses resulting from the hot-rolling procedure. The selected pattern (] in which appropriate measurements had been realized; It has to be taken into account that the residual stresses in hot rolled steel sections are independent of the steel grade and therefore a magnitude of 0,15·fy, that is approximately equal to 70 MPa for steel grade S460, is used. The selected pattern is applied automatically by the software to each beam element along the member length.a material law in accordance with the measured one (see A tolerance on the position of the applied load from up to 2,0 mm has been adopted for the numerical simulations in order to calibrate the results. It has been found that even a small eccentricity could affect the ultimate resistance and the stiffness of the member in comparison to the perfectly “no loading eccentricity” case. shows that an eccentricity equal to 1,5 mm (for the angle section L200 × 200 × 16) is able to change the ultimate resistance by approximately 6%. The influence of this small eccentricity of the applied load on the stiffness and the ultimate resistance has been also observed in Ref. []. The eccentricity has been applied in u direction as it has been found through numerical simulations that the influence on the response of an eccentricity in the v direction is negligible (see for the definition of the axes). This tolerance could be explained by the two following reasons:the nominal position of the load has been designed to coincide with the centre of the end plates and accordingly with the centre of gravity of the cross section. In reality, due to small differences of the cross-section geometry, the real centre of gravity does not coincide exactly with the loading point;the positioning of the specimen in the testing rig may also induce a small and unexpected eccentricity. show the axial deformations (shortening) of the specimens versus the load for both experimental tests (solid lines) and numerical simulations (dotted lines). summarizes and compares the ultimate experimental and numerical resistances. One can see from the graphs and the table that there is a very good agreement between numerical simulations and experimental tests in terms of axial stiffness and load carrying capacity. The mean value of the ratio Nexp/NFEM is equal to 0,98 with a COV of 1%. show some characteristic load-deflections curves of the specimens for both experimental tests and numerical simulations.The numerical results provide similar responses for most of the specimens without nominal loading eccentricity (same as Specimen Sp15 in ), as well as a good correspondence with the tests. However, for two specimens (Sp21, Sp23), the numerical response appears to be more flexible than the test, at least in the first part of the test; but close to the failure, a rather good agreement with the experiment is contemplated.Similar results, in terms of lateral flexibility of the columns, have been observed for eccentrically loaded specimens (see ). This additional flexibility could be explained by the fact that the end plates at the extremities of the angle members have been modelled indirectly (additional length and restraints at the extremities of the member), as well as by the consideration of the unintentional eccentricities. The prevailing of weak axis buckling near the failure load has been observed also through the numerical simulations for the eccentrically loaded specimens.In the following, the test results have been compared with the relevant Eurocode 3 predictions. Since many provisions have been changed during the development of the 2nd ongoing version of the Eurocodes, the analyses have been based on EN 1993-1-1: 2005, as well as the latest version of the Code available to the authors, namely prEN 1993-1-1:2019 []. Hereafter, a unique reference to “EN 1993-1-1” means that there is no difference between the two versions.According to EN 1993-1-1, the first profile (L 150 × 150 × 18) is classified as Class 1 and the second one (L 200 × 200 × 16) as Class 4. The procedure of EN 1993-1-5 [] has been followed to evaluate the effective cross-section of the Class 4 profile, even if no plate buckling has been reported during the tests.According to EN 1993-1-1, the buckling resistance for centrally loaded angle members is calculated by the equation:is the buckling reduction factor which should be determined as a function of the relative slenderness λ‾ of the compression member;is the ratio ρ = Aeff/A, where the effective area of the cross-section is defined in Ref. [The non-dimensional slenderness λ‾ is equal to:where Ncr is the elastic critical load for the relevant buckling mode (i.e. the minimum eigenvalue amongst all flexural and flexural-torsional buckling modes) obtained by an elastic instability analysis considering actual material (Young modulus), gross cross-section properties and buckling length, using FINELG software. For all centrally loaded specimens, the calculated eigenmode was a flexural buckling one. A pure torsional mode cannot be obtained for a centrally loaded angle column as explained in Ref. [The only difference of both aforementioned norms concerning the buckling resistance of centrally loaded members is in the selection of the buckling curves. EN 1993-1-1(2005) indicates curve b for all hot-rolled angle cross-sections independently of the buckling axis, while prEN 1993-1-1(2019) recommends curve b for steel grades S235 to S420 and curve a for S460 up to S700. includes the results of the calculations, with the assumption that specimens Sp11, Sp13 and Sp15 are S420 (based on the actual yield stress – buckling curve b) and Sp21, Sp23 and Sp25 are S460 (buckling curve a). In as in the rest of the paper, the subscript EC3 refers to EN 1993-1-1 (2005) and EC3’ to prEN 1993-1-1 (2019). illustrates the experimental results (only for centrally loaded specimens) compared with those obtained through the recommendations of EN 1993-1-1; reference buckling curves a0, a and b are reported too. The buckling reduction factor χexp of specimens has been evaluated by the equation:. The results from the calculations presented in In EN 1993-1-1(2005), the buckling curve b has been selected for axially loaded equal angle columns (solid line in ). It has been found that the experimental results are in line with this curve or above for specimens Sp1#, but are much higher for specimens Sp2#. In addition, it can be easily observed that the actual ultimate resistance of all centrally loaded columns is higher than the predictions of Eurocode; the later seems to provide safe evaluations, especially for specimens Sp2# where the detrimental effect of local buckling are possibly overestimated. Based on prEN 1993-1-1(2019), buckling curve b has been again selected for specimens Sp1# and the results are the same as before. However, when buckling curve a is used for Sp2#, the difference between the experimental and design resistance according to EC3 is smaller, as shown in The general method of Eurocode 3 can be applied to evaluate the resistance of the eccentrically loaded angle members subjected to axial force and bending moments. In this case, the out-of-plane buckling resistance of the member is sufficient if the following equation satisfies:χop is the reduction factor corresponding to the non-dimensional slenderness λop‾ and aimed at accounting for lateral and lateral torsional buckling, i.e. χop=min{χb,χLT};ault,k is the minimum load amplifier of the design loads to reach the characteristic resistance of the most critical cross-section of the structural component considering its in plane behaviour without taking lateral or lateral torsional buckling into account however accounting for all effects due to in plane geometrical deformation and imperfections, global and local, where relevant. It can be derived from eq. the first term relates to the stress under pure compression;the second, to the second order maximum stress resulting from the amplification of the first order moment NEd·e0 (e0 is the equivalent imperfection as defined in Eurocode 3), i.e. the moment NEd·e0[1/(1-NEd/Ncr,u)];the third one relates to the second order maximum stress resulting from the amplification of the first order moment NEd·ev (ev is the load eccentricity), which can be estimated as NEd·ev[1/(1-NEd/Ncr,u)].The load eccentricity (ev) resulted in strong axis bending, while geometrical equivalent imperfections (e0) around both axes. However, the latter were smaller than the former, so that finally the effects of load eccentricity prevailed.The global relative slenderness λop‾ for the structural component should be determined from eq. , in which the term αcr,op is the minimum load amplifier for the in-plane design loads to reach the elastic critical load of the structural component associated to lateral or lateral torsional buckling without accounting for in-plane flexural buckling.Regarding both versions of EN 1993-1-1, the difference is again in the selection of the buckling curves, but also here in the value of the equivalent imperfection e0. The calculations according to EN 1993-1-1(2005)-§6.3.4 and prEN 1993-1-1(2019)-§8.3.4 have been done and their results are reported in respectively. Factors and parameters that are the same for both calculations are presented only in . Additionally, a recalculation has been done for each specimen, so as to evaluate the analytical load for which eq. satisfies (χop·αult,k = 1,0). The obtained analytical estimations of the failure loads have then been compared to the experimental ones; the results are represent in Based on the results, one may conclude that the general method of Eurocode 3 is over-conservative even in its 2019 revised version.From the present study involving experimental, numerical and analytical aspects, the following conclusions may be drawn.The centrically loaded specimens with class 1 cross-section (Sp11, Sp13, Sp15) and the eccentrically loaded specimens with class 4 cross-section (Sp22, Sp24, Sp26) failed very clearly in a pure weak axis flexural buckling mode and correspondingly flexural torsional buckling mode.The centrically loaded specimens with class 4 cross-section (Sp21, Sp23, Sp25) and the eccentrically loaded specimens with class 1 cross-section (Sp12, Sp14, Sp16) failed mostly in a flexural torsional buckling mode, which was more pronounced in the former.Local buckling was not visibly observed in any specimen, although some of them were categorised as class 4 according to EN 1993-1-1.A very good agreement between the numerical GMNIA simulations and the experimental results in terms of axial stiffness and ultimate resistances has been achieved, through the consideration of an unintentional small eccentricity.A small eccentricity of the position of the applying load can affect the ultimate resistance of the member in comparison with the perfect “no loading eccentricity” case. For the current study, an eccentricity equals to 1,5 mm may reduce the ultimate resistance by about 6%.The design resistance of the concentrically specimens based on EN 1993-1-1 (2005) is on the safe side, especially for the second profile for which the local buckling reduction effects seem to be overestimated by Eurocode. The results conform better with the application of the 2nd generation of the Eurocode 3, namely prEN 1993-1-1:2019.The tests and the numerical simulations indicate, that it is more reasonable to calculate the member resistance of rolled angles by using the slenderness for flexural buckling only and not based on the “relevant buckling mode” which includes torsional effects, as EN 1993-1-1 prescribes.For eccentrically loaded columns, the general method recommended by Eurocode 3 appears as over-conservative.It should be noted that the conclusions may are limited to the tested specimens, but a wider study is in progress to investigate the classification and design formulae, covering this time the whole domain of application, including different profiles, width-to-thickness ratios, steel grades and member lengths.Bezas Marios-Zois: Writing - original draft, Investigation, Software, Validation Demonceau Jean-François: Methodology, Investigation, Formal analysis Vayas Ioannis: Conceptualization, Supervision, Writing - review & editing Jaspart Jean-Pierre: Conceptualization, Supervision, Writing - review & editingThe authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Pushover analysis of confined masonry walls using a 3D macro-modelling approachThis paper shows a novelty way to simulate the nonlinear behaviour of confined masonry walls subjected to in-plane lateral loading by using a 3D macro-modelling approach. For this purpose, the finite elements method implemented in ABAQUS software was used. All the 3D solid finite elements were modelled as a single part, which allowed avoiding modelling the contact interfaces between concrete and masonry elements. The nonlinear behaviour of the concrete and masonry were governed by two main types of failures: crushing and cracking, which were properly represented by the Concrete Damage Plasticity (CDP) model. Steel rebars were modelled as elastic–plastic with hardening and were assumed to have a perfect adhesion with the surrounding concrete by means of the embedded constraint. Prior to the modelling process, experiments were carried out whose results were used as patterns to validate the proposed model. A calibration process of the tensile properties of masonry was conducted for properly fitting the experimental patterns. As a result, there were good agreements between the numerical and experimental outcomes in terms of capacity curves and cracking patterns.Confined masonry buildings are the most common type of construction for dwellings in Peru and other South American countries. The major issue with many of these constructions is their informality: According to shows a common case of masonry dwellings in Lima of up to five stories.Peruvian seismic events have revealed the poor quality of these informal masonry dwellings, which has been responsible for human and material losses. However, this is not only related to the informality of the masonry constructions, but also to the lack of knowledge of the nonlinear behaviour of the masonry. For this reason, much research around the world has been devoted to experimental studies of the nonlinear behaviour of masonry walls, either for in-plane or out-plane loads In literature, many alternatives can be found for assessing the seismic vulnerability of masonry constructions in terms of the seismic hazard and the seismic response of the masonry walls during an earthquake Different numerical studies have been conducted to assess the behaviour of masonry walls subjected to in-plane lateral loads As the different studies have demonstrated, the nonlinear behaviour of the masonry walls is of great interest mainly where they are part of the building’s structural system. In the present paper, an easy way of assessing structural parameters of masonry walls is presented by using the macro-modelling approach. For this purpose, full-scale confined masonry walls were modeled to be subjected under monotonic lateral displacements. Finally, the proposed model was carefully calibrated by comparing capacity curves and cracking patterns with those recorded in a experimental campaign conducted by Manchego and Pari The previous work was carried out by Manchego and Pari shows the typical assemblage of the tested walls. Note the toothed connection between the confining columns and the masonry panel. Different studies, such as the one conducted by Singhal and Durgesh shows the typical assemblage that was used to carry out the cyclic tests. The lateral displacements were imposed by means of a dynamic actuator, which was controlled by displacements on a computer. This actuator was intended to be fixed to a reaction frame rigid enough to avoid distorted lateral displacements. On the other hand, one hydraulic jack was located at each end of the foundation to prevent it from being overturned. In addition, one hydraulic jack was located at one of the ends of the foundation to prevent its sliding horizontally in one direction. In the other direction, the foundation was intended to react against the rigid reaction frame. The vertical load, where it was applied, was imposed by an additional hydraulic jack, which in turn was connected to two rigid steel beams, in order to distribute as much as possible the vertical load over the wall’s confining beam.In order to compare the pushover analysis conducted in this paper with the cyclic experimental test conducted by shows these envelope curves for both walls with and without vertical load. Note that the shear stresses shown on the secondary vertical axis do not correspond to the real stresses, but to nominal stresses computed as the ratio between the lateral forces and the cross-sectional area of the walls.Experimental tests on small samples were also carried out in order to characterize the material properties involved in the confined masonry walls with and without vertical load. For instance, uniaxial compressive tests of brick prisms were conducted to get the compressive strength and Young’s modulus of the masonry. In addition, uniaxial diagonal compressive tests were conducted over small square masonry walls in order to get their tensile strength. Typical compressive tests of cylindrical specimens were carried out for each concrete element (foundation, column and beam), in order to get their compressive strength. It should be noted that every small sample was taken from each kind of tested wall (with and without vertical load) by considering they were built on different dates. The experimental results of these tests are shown in In this research, the macro-modelling approach has been adopted. In this way, the bricks, mortar, concrete, and their contact interfaces were not modelled separately. Rather, all components were treated as homogeneous and isotropic materials. The modelling process was carried out in the commercial software package ABAQUS. The foundation and wall were intended to be a single part, which meant that each contact between different materials was assumed to be monolithic. This assumption was made because of the constructive typology of the confined masonry. Namely, in this kind of constructions, the masonry panel is built prior to the casting of the confining elements. This fact added to the toothed connection often used between masonry and columns make all the components work as monolithic.Except for the steel reinforcement, all the components were modelled as continuum three-dimensional elements with 8 nodes with reduced integration (C3D8R). The steel rebars were modelled as truss three-dimensional elements with two nodes (T3D2). The interaction between the steel rebars and the surrounding concrete was considered as perfect adhesion, implemented by means of an embedded constraint. This means that no slip was taken into account between these two materials.As boundary conditions, each wall was assumed to be over an analytical rigid surface which represents the reaction slab shown in . The hydraulic jack and reaction frame intended to prevent the horizontal sliding of the foundation, as well as the dynamic actuator, were also modelled as analytical rigid surfaces. The hydraulic jacks which were intended to prevent the foundation from overturning were modelled as pin supports where only the vertical component was restricted. However, these hydraulic jacks had an unknown initial pressure prior to the cyclic testing, which in turn was increasing while the lateral displacements were increasing. This fact was intended to be modelled by assuming that a certain area below the foundation did not suffer vertical displacements like its corresponding restricted top area. For this purpose, it was assumed that the transmission of pressure between these hydraulic jacks and the reaction slab had a slope of 1:2, as shown in A variant of classical plasticity theory with the introduction of damage concepts is commonly used with Concrete Damage Plasticity to simulate the nonlinear behaviour of quasi-brittle materials. However, its accuracy is questionable, due to its tensile behaviour formulation Concrete Damage Plasticity (CDP) is a continuum plasticity-based damage model for concrete and other quasi-brittle materials in any type of structure. It is assumed that the failure of a material is governed by two main mechanisms: tensile cracking and compressive crushing. The evolution of its yield surface, which is defined by Eq. , is controlled by two hardening variables, ε̃tpl and ε̃cpl, which are the tensile and compressive equivalent plastic strain, respectively.F=11-αq¯-3αp¯+β(ε̃pl)〈σ¯̂max〉-γ〈-σ¯̂max〉-σ¯c(ε̃cpl)⩽0where p¯ is the effective hydrostatic pressure, q¯ is the von Mises equivalent effective stress, σ¯̂max is the maximum eigenvalue of σ¯, and β(ε̃pl) is the function defined bywhere σ¯t and σ¯c are the tensile and compression effective stress, respectively. The parameter α can be obtained experimentally with the following expression:where σb0 and σc0 are the failure stress for biaxial and uniaxial conditions, respectively. The parameter γ is defined aswhere Kc is a constant that can be obtained experimentally through triaxial tests where ψ is the dilation angle measured in the p-q plane with a high level of confinement pressure, σt0 is the uniaxial tensile strength, and e is an eccentricity that defines the rate at which the function reaches the asymptote. A typical yield surface for plane stress conditions is shown in . The intersection between the yield boundary and principal axes represents both the compressive and tensile uniaxial strength of the material. As is characteristic of quasi-brittle materials, a reduced biaxial tension and increased biaxial compression are also illustrated in All the parameters involved in CDP can be obtained from uniaxial, biaxial and triaxial tests, as described by Jankowiak and Lodygowski According to the configuration of the tested walls, 5 different materials were considered for modelling: (1) the foundation’s concrete, (2) the column’s concrete, (3) the beam’s concrete, (4) the masonry, and (5) the rebar’s steel. shows their mechanical properties, which were used for modelling purposes, where E is the Young’s modulus, ν is the Poisson’s ratio, fc′ is the compressive strength, ft is the tensile strength, fy is the yield strength, Gch is the crushing energy, and Gf is the fracture energy of the materials.Regarding the constitutive laws of concrete and masonry, it is known that their behaviour lies between ideal brittle and ductile. In fact, they are closer to a brittle behaviour than ductile, therefore, both are considered as quasi-brittle materials The compressive behaviour of concrete was represented by three main parts: (1) linear, (2) hardening and (3) softening. The linear part was taken to last up to a compressive stress equivalent of 0.4fcm′. The second part was characterized by a parabolic hardening, in compliance with CEB-FIP σc(2)=Eciεcfcm-εcεcm21+Eciεcmfcm-2εcεcmfcmγc=π2fcmεcm2Gchleq-0.5fcmεcm(1-b)+bfcmE02 shows the stress–strain curves considered for concrete in compression.Regarding the tensile behaviour, it was assumed to be governed by a first linear elastic part up to its tensile strength, at which point tensile failure begins. Thereafter, a post-failure behaviour of cracked concrete was intended to be defined in strain-softening terms. For this purpose, the formulation given by Hordijk defines the post-failure tensile behaviour in terms of the crack opening w [mm]. However, in the case of reinforced concrete, the post-failure relation is usually expressed in terms of strains. In this way, it is intended to avoid any dependence of the results on the mesh size. For this purpose, the cracking strain was expressed as εck=w/leq. shows the adopted post-failure stress–strain curves for the tensile behaviour of the concrete. It should be noted that a residual stress σr=ftm/50 was used to avoid kinetic instabilities.The compressive behaviour of the masonry was represented by three main parts: (1) parabolic hardening, (2) linear softening and (3) residual, according to the constitutive model proposed by . It is worth highlighting that, a residual stress of 0.1fcm, unlike , was taken into account to avoid kinetic instabilities As with the concrete, the tensile behaviour was assumed to be governed at first by a linear-elastic part followed by nonlinear behaviour. In this case, a post-failure exponential softening was assumed shows the adopted curves for the masonry.In cases where the behaviour of the reinforced concrete is what dominates, it is important to consider the bond slip effect between the steel reinforcement and the concrete. This effect is related to the fact that given a particular deformation of an RC element, the steel reinforcement and the concrete have different strains due to the difference in their material properties (e.g. the Young and Poisson modules). This effect can be taken into account by modifying the constitutive law of the steel reinforcement On the other hand, the steel reinforcement was modelled as elastic–plastic with a hardening of 2%E slope between the strains related to the yield and ultimate stresses, εy and εu, respectively. The proposed constitutive law that takes into account the effect of bond-slip compares a typical experimental curve for the steel reinforcement and the numerical curve adopted as the constitutive law.The damage parameters for both the compressive and tensile behaviour of the concrete were computed by following the formulation proposed by Alfarah et al. It is worth mentioning that even when a unit value of a damage parameter means a total failure of the material, which in turn means that the material can not carry more stress, this could not be applied to the model. This is related to the formulation of CDP Experimentally, the cyclic tests of the walls were carried out slowly, as a quasi-static event, in order to avoid kinematic effects. For modelling quasi-static phenomena, ABAQUS offers two powerful solvers: implicit and explicit. The implicit solver involves the solution of static equilibrium equations by enforcing an equilibrium between the internal and external forces. If the implicit solver does not find a convergence between these forces in a specific time increment, it adds certain corrections through the Newton–Raphson method. The process continues until the difference between the internal and external forces is less than a small value, called the convergence criterion. However, to solve the equilibrium equations, the implicit solver needs to invert the stiffness matrix, which involves a high computational cost, depending on the number of degrees of freedom.On the other hand, the explicit solver involves the solution of dynamic equilibrium equations. Unlike the implicit solver, the explicit solver does not enforce an equilibrium between the internal and external forces, which means that there is no convergence criterion. Moreover, the explicit solver needs to invert the mass matrix instead of the stiffness matrix, which turns out to be much cheaper computationally. Indeed, inverting an uncoupled diagonal mass matrix is less expensive than inverting a fully coupled stiffness matrix. Once the mass matrix has been inverted, the acceleration in a specific time increment is calculated. Thereupon, the velocity and displacements are calculated by means of the central difference method.In general terms, an explicit solution moves away from the real solution if each time step is divided into only a few time increments. That is why the ABAQUS Explicit Solver efficiently implements a large number of time increments in order to obtain reliable results Taking into account that each wall was assumed to be a single part, a single mesh size would affect the entire part. Therefore, in order to find an appropriate mesh size, one that would require less computational cost without losing accuracy, a sensitivity analysis was carried out.The Young’s modulus of the masonry shown in was computed experimentally by uniaxial compressive tests conducted perpendicularly to the bed joints. However, this value can not be used directly for modelling since it would lead to stiffer responses, as has also been observed by others shows the numerical curves obtained from iterating the Young’s modulus, where it is possible to note that an equivalent Young’s modulus E∗=65%E0 allowed to properly fit the initial stiffness of the experimental results. It has to be noted that regardless the percentage of the experimental Young’s modulus, the variation of the initial stiffness of the entire wall is not quite and it also does not affect too much the nonlinear behaviour of the wall. For instance, whether a value E∗=100%E0 had been used instead of E∗=65%E0, only a mistake of 14% would have been committed when capturing the initial experimental stiffness of the entire wall. Nevertheless, it must be taken into account that the initial stiffness of the entire walls is due to the contribution of both concrete frames and masonry panel, therefore, the small variation of 14% is due to the fact that only one source of stiffness is being affected. On the other hand, it is worth mentioning that the experimental area showed in , turned out from the area enclosed by the experimental curves showed in . Once calibrated the Young’s modulus, masonry’s tensile strength was the next parameter to be iterated, as is shown in In terms of fitting the nonlinear part of the experimental capacity zone (), it has to be noted the effect of decreasing the tensile strength of masonry, which is related to an earlier cracking of the masonry panel which in turns results in a quicker decreasing of the wall elastic modulus. According to , a value of tf=1.20 MPa was taken into account for the iteration of the next parameter. Regarding the fracture energy, shows the numerical curves obtained from iterating this parameter. Like tensile strength, the effect of decreasing this parameter is associated to a quicker cracking process and indeed to a quicker stiffness degradation. These results allowed to conclude that both parameters are linked and govern the nonlinear behaviour of the wall. According to the experimental zone, the value Gf=0.10N/mm was chosen for showing the best fitting.Once calibrated all the material parameters, it turned out important to compare the cracking pattern of both experimental cyclic test and numerical pushover analysis. shows the cracking pattern for both experimental and numerical tests, according to different performance levels related to: (1) end of linear behaviour, (2) yielding beginning and (3) maximum load capacity.It has to be noted that there was a good agreement in terms of cracking pattern between half an experimental wall and numerical results by showing bending cracks which occurred first on the confinement column and grew up to the bottom center of the wall. According to the evolution of cracks, in both cases it can be seen how horizontal cracks are propagated in the height of the confinement column whereas the lateral displacement increase. Regarding the beam foundation, it was possible to capture an unexpected bending failure which shows the potential of the proposed model for reproducing all the experimental effects. As it was aforementioned, the foundation could be assumed to be rigid enough in comparison with masonry wall, which implies that all the failure should be concentrated in the wall. However, the lack of bending stiffness of the RC beam foundation led to its unexpected failure which could be properly captured by the proposed model. In addition, it should be noted that the experimental cracking pattern shows additional cracks that can not be captured by a pushover analysis since they are related to the cyclic behaviour of materials. Namely, additional cracks are intended to appear when there are excursions among compressive and tensile states.The calibrated material properties for the case of walls without vertical load, were intended to be used for showing the reliability of the proposed model for fitting the experimental results of another testing setup, which corresponded to consider a vertical load prior the application of lateral loading. This vertical load had a value of 170 kN which tried to represent the weight of 3 stories over a wall located at the first floor. It is worth mentioning that this vertical load was controlled manually by an operator who noted an oscillating variation of the vertical load conforming the lateral displacements were increasing. In fact, this variation became significant from a displacement level of 7.70 mm onwards. That was why the first test with vertical load was stopped for this displacement level (). For the next tests, this vertical load was intended to keep close to the 170 kN as much as possible. However, it was not possible to control which meant a complication for the modelling process to capture the nonlinear behaviour of the walls from the displacement level aforementioned. This oscillating effect of the vertical load caused an increment in the load capacity of the walls which was considered as unreal because it was caused by an uncontrollable boundary condition during their tests.As it was mentioned before, the calibrated material properties from walls without vertical load were used here. However, it was noted that an equivalent Young’s modulus of E∗=65%E0 did not capture the initial stiffness showed by the experimental results. On contrary, it had to be used a value of E∗=100%E0 to properly fit the initial stiffness of the experimental results (). This variation in percentage is attributed to the fact that each wall, from its construction up to its test, has enough time to develop some micro-cracks which are related to the shrinkage of concrete and mortar, or to the fact that some high stresses can take place during the lifting of the walls. It is known that the presence of cracks is linked to the reduction of the Young’s modulus of the materials. However, the presence of vertical load helps to the closing of these cracks which is related to a recovery of this parameter. Anyway, within the showed range 65–100%, it can be highlighted that the variation of Young’s modulus does not affect too much the nonlinear response of the wall, as it was noted in It is important to mention that a proper application of the vertical load should have lead to a behaviour more close to the numerical curve. In fact, a constant vertical load would be more appropriate to represent the load condition of a wall located on the first floor in a real building. To support this idea, it can be seen the results of the experimental tests conducted by Perez et al. shows the cracking pattern for both experimental and numerical tests, according to different performance levels related to: (1) end of linear behaviour, (2) maximum load capacity and (3) ultimate state. It should be noted there was a good agreement between experimental and numerical cracking pattern until the second analyzed performance level. In the ultimate state, differences are evidenced by the presence of additional cracks, which in fact were produced by the effect of the experimental cyclic loading and the uncontrollable vertical load from second point onwards. Regarding the beam foundation, numerical results showed also a bending failure of the beam foundation, which was not observed experimentally. This effect is entirely attributed to the uncontrollable increment of the applied vertical load which offered a major restriction to the beam foundation against bending.As mentioned before, the quasi-static problem was intended to be solved by means of a purely dynamic explicit solver. For this purpose, the loads were applied by defining smooth step amplitudes, which has the advantage of having zero velocity in the application of the loading at the beginning and ending of the load step. In addition, these smooth steps allow gradually increasing the application of the loading, which helps to minimize kinematic effects. Subsequently, in order to be sure that the numerical results resulted mainly from quasi-static effects, i.e. that kinematic effects were not dominant, the kinetic and internal energy of the whole model (ALLKE and ALLIE, respectively) were compared over the entire time step (). Note that the kinetic energy curves are very close to the axle time. In fact, they were less than 1% of the internal energy over the largest part of the time step, which allowed being sure that the numerical results were not influenced by kinetic effects.A 3D finite element model based on the macro-modelling technique was presented to simulate the nonlinear behaviour of confined masonry walls subjected to in-plane lateral loading. For this purpose, all the solid elements were modelled as single parts, which means that no contact surfaces between the different materials were physically modelled. In addition, due to the assumptions made by Concrete Damage Plasticity (CDP) of treating quasi-brittle materials as isotropic, it was not possible to use the material properties directly obtained from small sample tests. On the contrary, these materials properties needed to be calibrated. After many iterations, varying the material properties of the concrete and masonry, it was concluded that the main parameters that controlled the nonlinear behaviour of the walls were the Young’s modulus, tensile strength and fracture energy of the masonry. Therefore, in a parametric study, these parameters were iteratively varied until reaching a good fit of the experimental results.Recalling that Young’s modulus, E0, of masonry was obtained experimentally by means of the well-known compressive tests of brick prisms, it could be seen that values of 65%E0 and 100%E0 properly fit the initial stiffness of walls without and with vertical load, respectively. In fact, Young’s modulus is related to the presence of cracks prior to testing, but anyway the impact of fitting the initial stiffness with a more or a little less precision does not affect too much the nonlinear response of the confined masonry walls.Regarding the tensile strength and fracture energy of the masonry, it is worth mentioning that both parameters controlled the cracking pattern and the nonlinear behaviour of the confined masonry walls. Therefore, they were iterated together so as to obtain a reduced value of tf=1.20 MPa, instead of the tf=1.40 MPa obtained from the well-known diagonal compressive test of square masonry samples, and a fracture energy of Gf=0.10 N/mm, which allowed properly fitting both the cracking pattern and the nonlinear part of the capacity curves.Finally, the proposed model achieved good precision in capturing the nonlinear response of confined masonry walls as well as their cracking pattern. Therefore, taking into account the efficiency and the simplicity of the application of the model herein proposed, it can be concluded that it can be used to help laboratory tests and designing codes in case it is important to predict the cracking patterns, maximum load capacity, and the ultimate displacements of confined masonry walls.The authors declared that there is no conflict of interest.Cold-formed steel thin-walled members are omnipresent in the modern building industry due to their inherent enhanced characteristics over conventional thicker hot-rolled sections. Their usage in building construction has grown rapidly, particularly in the last few decades due to the availability of more innovative cold-formed steel products, which allows direct replacement of masonry and timber products with cold-formed steel sections. These cold-formed steel products are also faster to fabricate and construct compared to the traditional building products. Versatility of the different shapes and sizes of cold-formed steel sections that are currently available allow them to be used effectively as floor joists, roof trusses, roof purlins and partition walls.Design equations in the current cold-formed steel design standards for web crippling are based on experimental data obtained from many research studies conducted in the past. However, comparisons and reviews reveal that current design equations are often inaccurate and limited to a range of sections controlled by geometric and material properties of experimental data used to calibrate the design equations. Inaccuracy in the design capacity predictions is mainly due to the inconsistency in the test set-up and specimen length used in the past experimental studies.Web crippling failures are defined into four categories depending on the location of applied load or reaction force. They are End-Two-Flange (ETF), Interior-Two-Flange (ITF), End-One-Flange (EOF) and Interior-One-Flange (IOF). Many experimental web crippling studies of cold-formed steel sections conducted in the past had variations among their test arrangement and specimen lengths. To overcome this issue, a standard test method with details of suitable web crippling test set-ups and procedures was published by the American Iron and Steel in AISI S909 ϴ = angle between the web plane and bearing surface plane, 90° ≥ ϴ ≥ 45°Cw = Web slenderness coefficient (d1/tw)Cr = Inside bent radius coefficient (ri/tw)Rb=(k1k2k36.66-dwtw641+0.01ℓbtwtw2fy)/γM1Rb=(k3k4k521.0-dw/tw16.31+0.0013ℓbtwtw2fy)/γM1k1=1.33-0.33kk2=1.15-0.15ri/tw 0.50 ≤ k2 ≤ 1.0k4=1.22-0.22kk5=1.06-0.06ri/tw, but k5 ≤ 1.0γm1 = Partial safety factor for member resistance = 1.00 show the details of the SupaCee test specimens used in this experimental study, which included 21 ETF load tests (7 SupaCee sections) and 15 ITF load tests (5 SupaCee sections). They include all the measured geometric and mechanical properties including the important parameters such as web thickness (tw), inside bent radius (ri), section depth (d), and average web yield stress (fy) of each tested SupaCee section. Sections were defined based on their Australian manufacturer’s approach of using the nominal section depth and thickness of the section, for example, SC15010 means 150 mm section depth and 1.0 mm thickness. The yield stresses in were obtained from tensile coupon tests, which indicate that the use of Australian steel grades G450, G500 and G550 guarantee minimum yield stresses of 450, 500 and 550 MPa, respectively. show the web crippling test set-up used in this experimental study for SupaCee sections under ETF and ITF load cases. The minimum specimen lengths specified in the AISI standard test method (AISI S909) An Instron testing machine and associated data acquisition system in the university laboratory were used in conducting all the web crippling tests. Three different bearing plates (50, 100 and 150 mm) used in this study provided three types of bearing length loading conditions for both ETF and ITF load cases. The applied load and associated test specimen displacements were measured during the tests. A SupaCee test specimen was located between the bearing plates, the measuring system in the Instron testing machine was then initialized and the loading was commenced following a 50 N load application. Test specimens were loaded using a displacement control method (0.7 mm/minute). Each test specimen was loaded beyond reaching the ultimate load until a failure mechanism was formed.The web crippling failure modes of ETF-SC15010 and ITF-SC15010 with 50 mm, 100 mm and 150 mm bearing plates are shown in (b) and (c) show the tested section in the initial and ultimate failure stages. (a)–(c) show the applied load versus vertical and lateral deflection (mid-height of web) curves, initial and ultimate failure stages of ITF-SC15015 section with 100 mm bearing plate under ITF load case. give the ultimate web crippling capacities obtained from the ETF and ITF tests. These test capacities are compared with the predicted capacities from the current web crippling design equations in AS/NZS 4600 (Eq. . For ETF load case, the mean value of the ratios of experimental and predicted web crippling capacities of SupaCee section by AS/NZS 4600 is 0.65 and the coefficient of variation (COV) is 0.09. For ITF load case, these values are 2.41 and 0.39. Hence, AS/NZS 4600 This section presents the details of the development of finite element models of SupaCee channel sections subject to web crippling using ABAQUS Version 6.14 . The measured dimensions of SupaCee sections were converted to centreline dimensions in order to accurately represent the section in ABAQUS using middle surface shell offset definition. SupaCee sections were created using 3D deformable shell elements while loading and support bearing plates were modelled with discrete rigid elements. All the shell elements used for SupaCee sections were of type S4R, which is a linear four-node reduced integration shell element with finite strains. (a) and (b) show the developed finite element models of SupaCee sections under ETF and ITF load cases.The material property of developed finite element model was defined based on the tensile coupon tests of samples taken in the longitudinal direction of the web. The average yield stresses determined from the tensile coupon tests were used for this purpose (). The elastic modulus and Poisson’s ratio of steel were considered as 200,000 MPa and 0.3, respectively.In this study a stress–strain model with strain-hardening effects was simulated using the measured average yield and ultimate stresses to replicate the stress–strain curve obtained from the tensile coupon tests. The steel density is also needed in the quasi-static analysis. It was taken as 7850 kg/m3 (7.85 × 10−9 tonne/mm3 in ABAQUS) in this study.A single reference point can be assigned to represent the rigid body elements in ABAQUS. Boundary conditions were assigned to the reference points of loading and support bearing plates. In this study, surface-to-surface contact was assigned between shell finite element model representing SupaCee sections and rigid plates at the top and bottom representing bearing plates. To make the behaviour of the local contact area more realistic, separations were allowed between the contact surfaces after the initial contact. The friction between contact surfaces was set to 0.4 and the contact surfaces were assigned to hard surfaces using hard contact pressure overclosure.Quasi-static analytical option was chosen in this study. Kaitila The finite-element models developed for SupaCee sections were validated by comparing their ultimate capacities with those obtained from the web crippling tests. The web crippling capacities obtained from finite element analyses of SupaCee sections under ETF and ITF load cases are compared with the capacities from experimental studies in . For ETF load case, the mean value of the ratios of test and FEA web crippling capacities of SupaCee sections is 0.93 while the associated coefficient of variation (COV) is 0.07. For ITF load case, these values are 0.99 and 0.08, respectively. Therefore, it can be concluded that using quasi-static finite element analyses based on the measured geometric and mechanical properties show a very good agreement with experimental results for ETF and ITF load cases.The failure modes observed in experiments and FEA for SupaCee sections and their load–deflection curves are compared in shows a comparison of the applied load versus vertical deflections curves and failure modes of ETF-SC15012 section with 50 mm bearing length as observed in experiments and FEA. Comparison of failure modes and load–deflection curves with 150 mm bearing length for the same section are presented in . Similar comparison for ITF-SC15012 section with 50 and 150 mm bearing lengths are presented in , respectively. These figures show that the developed finite element models simulated the web crippling behaviour of SupaCee sections observed in their experiments accurately. can be used to calculate the web crippling capacities of SupaCees (Rb) with the relevant web crippling coefficients given in . However, it was not developed for sections with web stiffeners such as SupaCees. Therefore, the ultimate web crippling capacities obtained from the experimental and finite element studies are compared in . For ETF load case, the overall mean of the ratios of test or FEA capacity to the web crippling capacities predicted by AS/NZS 4600 is 0.68 while the associated COV is 0.11. These values are 2.40 and 0.38 for ITF load case. Since the observed web crippling behaviour of lipped channel and SupaCee sections was similar, all the coefficients except the overall coefficient (C) were kept the same as those presented in Sundararajah et al. shows the proposed coefficients to be used with Eq. The web crippling capacities predicted using Eq. and the proposed coefficients are also presented in . The mean and COV of the ratios of test or FEA capacity to predicted capacity using the proposed design equation under ETF load case are 1.00 and 0.12, respectively. These values for the ITF load case are 1.00 and 0.07. This comparison indicates that the web crippling capacities (Rb) of SupaCees predicted using Eq. and the modified web crippling coefficients agree well with those obtained from experiments and finite element analyses.SupaCee sections are only made of G450, G500 and G550 high strength steels depending on their thickness. They are also available in limited sizes unlike the more commonly used lipped channel sections that can also be made of low grade steels such as G300 steels. Hence a detailed parametric study was not undertaken for SupaCee sections and only the results for the available sections in were used in developing the web crippling design capacity equations.Capacity reduction factors (Фw) were also calculated for use with the proposed design equation using the mean and COV values reported in . They are 0.85 and 0.90 for ETF and ITF load cases, respectively. Hence a Фw factor of 0.85 is recommended for use with Eq. to calculate the design web crippling capacities of SupaCee sections under ETF and ITF load cases. In this study, appropriate capacity reduction factors (Фw) were determined and proposed based on the AISI S100 guidelines in Chapter F, which recommends a minimum reliability index (β) of 2.5 for cold-formed steel sections.The direct strength method (DSM) is included in both AS/NZS 4600 The critical buckling loads (Rb,cr) of SupaCee sections are important to predict the web crippling capacities using DSM. Sundararajah et al. ) to calculate the buckling coefficients for lipped channel sections under ETF and ITF load cases for use with Eq. . Using a similar approach, idealized finite element models of SupaCee sections under ETF and ITF load cases were created and analysed.kprop=Cb1-Cb,rritw1-Cb,wd1tw1+Cb,ℓℓbtw1+Cb,bbftw(a) and (b) present the buckling modes of SupaCee sections under ETF and ITF load cases, respectively. compare the proposed (kProp) and finite element analyses (kFEA) buckling coefficients. The mean value of kFEA/kProp is 1.0 for both load cases while the associated COVs are 0.05 and 0.08. These comparisons show that the predicted buckling coefficients using Eq. agree well with those obtained from the FEA of SupaCee sections under both load cases. The proposed coefficients for SupaCee sections in are identical to those proposed for lipped channel sections in Sundararajah et al. Since the failure modes observed in the experiments and finite element analyses of SupaCee sections were similar to lipped channel sections, the yield loads (Rb,y) of SupaCee sections were calculated using the yield load equation derived for lipped channel sections (Eq. ) and the yield mechanism length equations (Eqs. where Nm = yield mechanism length, rm = inside bent radius at the middle of the section, (ri + t/2) and rext = external bent radius, (ri + t).Based on the critical buckling load (Rb,cr) and the yield load (Rb,y) calculated using Eqs. and the experimental and numerical studies conducted in this research, DSM equations were developed for SupaCee sections under ETF and ITF load cases (Eqs. ). The slenderness (λ) for web crippling is defined using Eq. Forλ⩽0.72,Rb=Rb,yForλ>0.72,RbRb,y=1-0.25Rb,crRb,y0.98Rb,crRb,y0.98for ETF Load CaseForλ⩽0.84,Rb=Rb,yForλ>0.84,RbRb,y=1-0.17Rb,crRb,y0.70Rb,crRb,y0.70for ITF Load CaseTo verify the DSM based equations proposed for SupaCee sections in this study, the ultimate web crippling capacity results from both experiments and finite element analyses were compared with the proposed design equations (Eqs. for ETF and ITF load cases, respectively. A non-dimensional DSM format of Rb/Rb,y versus λ (√(Rb,y/Rb,cr) was used in these figures. They show that the DSM equations proposed in this study are capable of accurately predicting the web crippling capacities of SupaCee sections. As expected DSM based Eqs. developed for SupaCee sections are similar to those proposed for lipped channel sections in Sundararajah et al. Predictions from the DSM equations were also compared with the results from both tests and finite element analyses in . The overall mean and COV values of the ratios of test or FEA web crippling capacity to the predicted capacity of SupaCee sections are 0.99 and 0.16, respectively, for ETF load case, while they are 1.00 and 0.09 for ITF load case. This comparison indicates that the web crippling capacities of SupaCees sections predicted by the DSM equations (Eq. ) agree well with those obtained from experiments and finite element analyses.Suitable capacity reduction factors (Фw) were calculated based on AISI S100 . Detail of capacity reduction factor calculations are also reported in Sundararajah et al. to calculate the design web crippling capacities of SupaCee sections under ETF and ITF load cases, respectively.. The overall mean and COV values of the ratios of test and FEA web crippling capacities to the predicted capacity of SupaCee sections by Nguyen et al. Experimental and finite element studies of lipped channel and SupaCee sections were presented in the previous sections, which show that there are some differences between them for ITF load case. Comparison of their web crippling capacities showed that the web crippling capacity of SupaCee sections was reduced by the presence of longitudinal web stiffeners.In this section, a finite element study was undertaken to investigate and compare the web crippling behaviour of lipped channel and SupaCee sections subject to ETF and ITF load cases. As shown in , two sections having similar mechanical and geometrical properties were considered, except the lip and stiffener geometries to characterize each type of section. These sections were subjected to 100 mm bearing plate loading under ETF and ITF load cases. Following sections provide a comparison of failure modes and web crippling capacities of these sections. shows the geometric and mechanical properties of the analysed sections and their web crippling capacities under ETF load case. The influence of longitudinal stiffeners against web crippling appears to be absent under ETF load case. The web crippling capacity of SupaCee section is only marginally lower than lipped channel section. shows the failure modes observed at the ultimate failure stage under ETF load case. Despite the presence of longitudinal web stiffeners and curved lips, both sections failed in web buckling under ETF load case. The longitudinal web stiffeners did not appear to make any significant difference to the failure mode. also shows the cross sectional views of the failed sections at the loading point and confirm the above observation. Experimental and finite element studies presented in the previous sections also showed that the web crippling capacities of SupaCee sections are only slightly lower than those of lipped channel sections and therefore the effect of longitudinal web stiffeners and the curved lips on the failure capacity of SupaCee sections is minimal. shows the geometric and mechanical properties of the analysed sections and their web crippling capacities under ITF load case. shows the failure modes observed at the ultimate failure stage of each section under ITF load case. In contrast to ETF load case, the web crippling capacities show a considerable difference between them under ITF load case. The web crippling capacity of SupaCee section was reduced by 19% when compared with a similar lipped channel section. This comparison confirms the finding of about 15% reduction in the web crippling capacity for SupaCee sections based on experimental and finite element study results presented in the previous sections. shows the difference in the web crippling behaviour of the two sections, ie., localized failure along the longitudinal web stiffeners and the presence of concentrated stresses, leading to the reduced web crippling capacity of SupaCee sections.The effects of longitudinal web stiffeners on the web crippling capacities of channel sections were also investigated. This study has shown that the web crippling capacity of SupaCee sections was reduced by about 15% due to localized failures along the longitudinal web stiffeners when compared to similar lipped channel sections under ITF load case, whereas only a marginal difference was observed under ETF load case.Production and characterization of polycaprolactone- hyaluronic acid/chitosan- zein electrospun bilayer nanofibrous membrane for tissue regenerationSkin, as the largest and heaviest organ of the human body, has a total area of 1.5–2.0 m2 and a weight of 3.5–10 kg for an adult human body. This organ comprises three interconnected layers, epidermis, dermis and hypodermis that perform several functions including protection against external threats, thermoregulation, excretion, fluid homeostasis and sensorial detection To overcome these issues, researchers are now developing bioengineered skin substitutes composed of acellular or cellular three-dimensional (3D) matrices that are aimed to improve skin regeneration. According to the layer they are conceived to replace, skin substitutes are grouped in epidermal, dermal and bilayered/dermo-epidermal Therefore, electrospinning as a simple and cheap technique has been selected by different researchers for producing nanofibrous meshes that are aimed to be used as skin substitutes Chitosan (CS), Zein (ZN) and Salicylic Acid (SA) were used to manufacture the bottom layer of EBM. CS is a natural polysaccharide obtained from partial deacetylation of chitin, that promotes hemostasis and also owns bacteriostatic and fungistatic activities 3,3,3 Trifluoroethanol (TFE), HA (MW = 1.5–2.2 million Da), SA and ZN were purchased from Acros Organics (New Jersey, USA). 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) was purchased from Alfa Aesar (Ward Hill, USA). Fetal bovine serum (FBS) (free from any antibiotic) was acquired from Biochrom AG (Berlin, Germany). Glacial acetic acid (AA) was obtained from LabChem (New York, USA). Paraformaldehyde (PFA) was obtained from Merck, SA (Algés, Portugal). Normal Human Dermal Fibroblasts (NHDF) cells were purchased from PromoCell (Labclinics, S.A., Barcelona, Spain). Amphotericin B, CS (low molecular weight (LMW)), Dimethylformamide (DMF), Dulbecco’s modified Eagle’s medium (DMEM-F12), Ethylenediaminetetraacetic acid (EDTA), Gentamicin, Kanamycin, LB Broth, Phosphate-buffered saline solution (PBS), PCL, Sodium hydroxide (NaOH) and Trypsin were purchased from Sigma- Aldrich (Sintra, Portugal). Dimethyl sulfoxide (DMSO) was obtained from Thermo Fisher Scientific (Rockford, IL, USA).CS was purified and deacetylated through a method previously described by Miguel and colleagues To produce the EBM, two distinct polymeric solutions were electrospun using a conventional electrospinning apparatus. The system setup was comprised of a high voltage source (Spellman CZE1000R, 0–30 kV), a precision syringe pump (KDS-100), a plastic syringe with a stainless steel needle (21 Gauge) and an aluminum disk connected to a copper collector. In order to produce the bottom layer of the electrospun mesh, 40% ZN (w/v) and 7% CS (w/v) were dissolved in 80% ethanol (EtOH) (w/v) and 70% AA (w/v), respectively. The final mixture (9:1 w/w ratio) was then homogenized with a X10/25 Ultra-turrax®, during 15 min. Finally, 8% SA (w/v) was added to the mixture. To produce the top layer of the electrospun mesh, 7% PCL (w/v) was dissolved under stirring in 80% TFE (w/v), at 60 °C, and 1% HA (w/v) was dissolved in 4:1 NaOH (5 M)/DMF (w/v). After the complete dissolution of both polymers, a 1:9 HA/PCL (w/w) final ratio was prepared for the following step.The CS_ZN_SA solution (10 mL) was then placed in the syringe and electrospun using a constant flow rate of 3.0 mL/h, a working distance of 15 cm and an applied voltage of 28 kV (see A for further details). After the production of the first layer, 10 mL of the HA_PCL polymeric solution were electrospun on CS_ZN_SA surface at a constant flow rate of 2.5 mL/h, using a working distance of 15 cm and an applied voltage of 25 kV (B). The produced EBM was then used in different in vitro assays for evaluating its applicability as a wound dressing (To characterize the composition of the produced EBM, Attenuated Total Reflectance-Fourier Transform Infrared Spectroscopy (ATR-FTIR) analysis was performed. The spectra were acquired for the HA_PCL and CS_ZN_SA membranes using an average of 128 scans, with a spectral width ranging from 400 and 4000 cm−1, with a spectral resolution of 4 cm−1. All the samples were mounted on a diamond window, and the spectra were recorded using a Nicolet iS10 FTIR spectrophotometer (Thermo Scientific, Waltham, MA, USA). In order to compare the spectra of individual components with those of the produced membranes, all the raw components used for nanofibrous mesh production were analyzed.The tensile properties of EBM were determined using a Shimadzu AG-X Tensile Testing Machine (Tokyo, Japan) at room temperature (RT), using a protocol described elsewhere Where F is the applied force; A is the cross-sectional area; Δl is the change in length; and L is the length between the clamps.The microporosity of EBM and their individual membranes was determined by using a liquid displacement method adapted from Antunes et al. Where Ww and Wd are the wet and dry weights of the membrane, respectively. Dethanol represents the density of EtOH at RT and Vmembrane is the volume of the wet membrane.The contact angles of EBM were determined using a Data Physics Contact Angle System OCAH 200 apparatus, operating in static mode, at RT. For each sample, water drops were placed on the top and bottom layers of the EBM. The reported contact angles are the average of at least three independent measurements.The water vapor transmission rate (WVTR) through EBM was determined as described elsewhere Where Wloss is the daily weight loss of water and A is the area of the tube opening.The in vitro degradation of the EBM and of the individual membranes (CS_ZN_SA and HA_PCL) was monitored by immersing the samples in PBS (pH 5.5), under stirring (40 rpm), at 37 °C. Then, samples were removed from the solutions after 1, 3 and 7 days and were completely dried and weighed. The degradation percentage at each time point was calculated according to Eq. Where Wi corresponds to the initial weight of the sample and Wt to the weight of the sample at time t.The amount of SA released from EBM was determined according to the method previously reported by Ouimet et al. The cytotoxic profile of HA_PCL and CS_ZN_SA individual membranes and of EBM was characterized using an MTT assay that was performed according to the guidelines set by ISO10993-5 standard. Briefly, the medium was removed and 50 μL of MTT (5 mg/mL PBS) were added to each sample (n = 5), followed by their incubation for 4 h, at 37 °C, in a 5% CO2 atmosphere. Then, cells were treated with 200 μL of DMSO (0.04 N) for 30 min. A microplate reader (Bio-rad xMark microplate spectrophotometer) was used to read the absorvance at 570 nm of the samples from each well. Cells cultured without materials were used as a negative control (K−), whereas cells cultured with EtOH (90%) were used as positive control (K+).Confocal laser scanning microscopy (CLSM) was used to characterize the cell distribution on both layers of the EBM. Cells (10 × 103 cells/mL) were seeded in μ-Slide 8 well Ibidi imaging plates (Ibidi GmbH, Germany) on the bottom and top layers of EBM. Cells seeded on the surface of the Ibidi imaging plate were used as control group. After 24 h, cells were fixed with 4% PFA in PBS during 20 min and then stained with the WGA-Alexa 594® conjugate. Cells were then rinsed several times with PBS and labeled with Hoechst 33342® nuclear probe (2 μM). The 3D reconstruction and image analysis were performed using Zeiss Zen 2010 software.Staphylococcus Aureus (S. Aureus), a gram-positive bacterium obtained from the clinic, was used to characterize the bactericidal activity of EBM, HA_PCL and CS_ZN_SA membranes. Briefly, 25–50 mg of each sample were added to 10 mL of LB broth at pH 6.2, containing 1 × 105 colony forming units (CFU)/mL of an early mid-log phase culture of S. Aureus incubated at 37 °C, for 24 h. After the incubation period, serial dilutions were prepared and 100 μL of bacterial samples were transferred into LB agar plates. Following an overnight incubation at 37 °C, bacterial colonies were counted and expressed as the number of colony forming units per mL. The bacterial concentrations were monitored through optical density measurements at 600 nm A modified Kirby-Bauer technique was used to further characterize the antimicrobial properties of the membranes SEM analysis was performed to characterize the EBM morphology, cellular attachment and biofilm formation on its surface. Samples that contain cells or bacteria were fixed during 4 h with 2.5% glutaraldehyde (v/v). After, samples were washed three times with PBS, frozen using liquid nitrogen and freeze-dried for 3 h. Finally, all samples were mounted onto aluminum stubs using Araldite glue and sputter-coated with gold through a Quorum Q150R ES sputter coater. SEM images were then acquired with different magnifications, at an acceleration voltage of 20 kV, using a Hitachi S-3400N Scanning Electron Microscope.The statistical analysis of the obtained results was performed using one-way analysis of variance (ANOVA), with the Newman-Keuls post hoc test. A p value lower than 0.05 (p |
< 0.05) was considered statistically significant.A shows macroscopic images of the top and bottom view of EBM, where both layers present an opaque appearance and a uniform structure. Additionally, HA_PCL layer shows a white color which is in accordance with previously published studies Furthermore, the morphology of electrospun nanofibers and fibers diameter were analyzed by SEM (B and C). HA_PCL and CS_ZN_SA layers presented a highly porous 3D nanofiber network composed of randomly oriented fibers with diameters of 472 ± 192 nm and 530 ± 180 nm for HA_PCL and CS_ZN_SA electrospun layers, respectively. Cheng et al. developed PCL/gelatin nanofibrous scaffolds with a mean diameter of 470 nm, which is similar to the diameter displayed by the layers of EBM , it is possible to observe that both layers of EBM form an almost continuous arrangement from the bottom to the top of the EBM. Such connection between the two layers can be explained by the presence of polymers with opposite charges on both layers. Indeed, HA present in HA_PCL electrospun layer, possesses glucuronic acid residues that contain carboxyl groups that confer negative charge to this membrane layer The ATR-FTIR spectra of raw materials, HA_PCL and CS_ZN_SA membranes are presented in . The HA_PCL spectrum displays the characteristic peaks of HA at 3200–3600 cm−1 (OH and NH stretching), 1640–1690 cm−1 (CO stretching of primary amide) and 1035 cm−1 (CC stretching) and 1170 cm−1 (symmetric CFurthermore, the spectrum of the CS_ZN_SA electrospun membrane (B) displays the typical bands of CS, ZN, and SA. The peaks observed at 1374 cm−1 (O stretching of the primary alcohol group), 3200–3400 cm−1 (H stretch) and at 1650 cm−1 (acetylated amine group) are referred to CS H stretching band of secondary amide), 1645 cm−1 (CO stretching of primary amide band), and 1515 cm−1 (NThe mechanical properties of the EBM produced were performed according to a protocol described elsewhere . The Young’s modulus obtained for EBM (4.29 ± 1.46 MPa) is extremely close to values displayed by native skin (4.6–20 MPa). However, the tensile strength (2.17 ± 0.62 MPa) and breaking strain (23.09 ± 3.07) values obtained for the produced EBM are slightly lower than those displayed by native tissue. Nevertheless, the mechanical properties of EBM are compatible with its handling during surgery and are similar to that reported for other electrospun membranes produced by other authors, and that used as skin substitutes The mechanical performance displayed EBM can be attributed to PCL, a synthetic polymer well known by its mechanical strength Porosity is a crucial feature of 3D constructs that are aimed for biomedical applications. Cells are prone to adhere and proliferate in porous 3D constructs. So far, different reports have highlighted the importance of the porosity of skin substitutes. Such property is important for improving fibroblast, endothelial and stem cell infiltration, nutrient transport, neovascularization, homeostasis, exudates absorption and gas permeability In literature, it is described that a surface is considered hydrophilic if the contact angle is lower than 90°, whereas surfaces with contact angles higher than 90° are considered moderately hydrophobic. Surfaces with a contact angle between 150° and 180° are called superhydrophobic Herein, the upper layer (HA_PCL) presented a contact angle of 120.20 ± 0.85° and thus a hydrophobic character (90° < Water Contact Angle < 150°). Such result is explained by the hydrophobic character of PCL present in this layer .) was optimized according to other reports in literature, where 40% (w/v) was determined as the ideal ZN concentration for obtaining stable fibers through a fast and reproducible electrospinning process The maintenance of a moist environment at the wound site is crucial for improving the healing process. In fact, a moist environment avoids patient dehydration, enhances angiogenesis and collagen synthesis. Thereunto, WVTR was determined in order to assess the EBM capacity to maintain a moist environment at the wound site. The obtained data shows a nearly constant water weight loss for control (non sealed) and membrane sealed group. The results showed that EBM had a WVTR of 1762.91 ± 187.50 mL/m2/day, a value that is similar to the control (1539.36 ± 50.01 mL/m2/day) (Table S2). These results demonstrate that EBM does not limit water vapor exchanges. Furthermore, the recommended WVTR of wound dressings is 2000–2500 mL/m2/day, which is half of the value presented by granulation tissue (5138 ± 202 mL/m2/day) Biodegradable biomaterials act in the area of tissue engineering, as a temporary ECM. Ideally, the degradation rate of polymeric matrices must be compatible with the new tissue formation. Fast degradation rates lead to the loss of tissue integrity and functions, whereas slow degradation rates can result in mechanical mismatches such as stress shielding, that can cause the failure of the system B, HA_PCL membrane lost approximately 5% of its weight after 7 days. Such result can be explained by the presence of PCL, which exhibits a slow degradation profile Moreover, CS_ZN_SA membrane presents the highest weight loss (almost 30%) after 7 days The EBM produced herein lost almost 10% of its weight after 7 days, a value that is comprehended between the weight losses of the individual membranes. The degradation process observed for EBM may be explained by the entanglement between the two layers. A shows that the top and bottom layers of EBM suffered some structural variations along 7 days of incubation in PBS (pH 5.5). After this period, the layers displayed a lower fiber density and fibers became more irregular. This may have a positive effect on cellular internalization, nutrients diffusion and ultimately on the skin healing process The release studies revealed that approximately 16% of SA was released from EBM during 5 days (Fig. S2). Such result is explained by the degradation of EBM. During the first hour in contact with PBS at 37 °C, the EBM exhibit a burst release of SA and hereafter a sustained release. The initial amount of SA that is delivered is appropriate to eliminate bacteria at the wound site, while a subsequent slow discharge of drug may avoid biofilm formation and infection of the wound To evaluate the cytotoxic character of EBM, NHDF were used as model cells. NHDF were chosen, since they interact with the surrounding cells (keratinocytes, fat cells and mast cells) and they are involved in production of ECM, glycoproteins, adhesive molecules and various cytokines that are essential for wound healing to occur Optical microscopic images of NHDF cells in contact with membranes after 1, 3 and 7 days were acquired (see Fig. S3). The images show that NHDF cells in contact with the materials do not suffer any morphologic variation and were able to proliferate similarly those on K− (cells that were grown without being in contact with biomaterials). As expected, cells presented in K+ display a spherical shape, which is characteristic of dead cells.Additionally, cell viability was also characterized through an MTT assay during 1, 3 and 7 days. The metabolic conversion of MTT, a yellow tetrazole salt, to purple formazan crystals occurs in living cells and such process is proportional to the number of viable cells present in each well A) revealed that the produced membranes did not affect the cell viability for 7 days.On the other hand, EBM may provide a three-dimensional structure for cellular attachment, growth and migration. Therefore, the interaction between cells and nanofibers were also evaluated by SEM. In B, it can be observed that after 3 days, the cells already present various filopodia protrusions, thus revealing that NHDF cells adhered and proliferated on the surface of EBM. At day 7, cells were completely attached onto the top and bottom of EBM Furthermore, CLSM images were also acquired () and they further corroborate the results obtained by SEM analysis. The cells maintained a normal phenotype and remained biologically active at the surface of both surfaces of EBM. After 3 days of culture, the cellular adhesion and proliferation onto EBM was noticed.All these data clearly demonstrate the potential of EBM to enhance the wound healing process by promoting cell migration, adhesion and proliferation. Once fibroblasts are at the wound site, they produce and secrete ECM proteins (predominantly collagen type III) and also growth factors such as Transforming Growth Factor-β1 (TGF-β1), Fibroblast Growth Factor (FGF) and Vascular Endothelial Growth Factor (VEGF) that are required for restoring the structure of the injured tissue Bacterial contaminations are currently regarded as the most severe and devastating complications associated with the implantation of biomaterials into the human body In the present study, S. aureus strain was used as model bacteria to evaluate the bactericidal activity of EBM. This strain was selected since it is described in the literature as the most common pathogen found in skin infections, when biomaterials are used for wound treatments A) and no biofilm was noticed on EBM’s surface after 5 days (B). In comparison with the individual electrospun membranes (HA_PCL and CS_ZN_SA), EBM showed the highest bactericidal activity. Such result can be explained by a synergetic effect obtained due to the presence of SA and CS in the bottom layer of EBM, that have been already described in literature as holders of antimicrobial properties The synergistic effect between CS and SA results in a bilayered system with antimicrobial activity that is active for at least 5 days. This profile is suitable to provide an aseptic environment to the wound site, inhibiting the bacterial proliferation, and hence, facilitating the wound healing process.Every year millions of people worldwide suffer from both acute and chronic skin injuries. Among the several therapeutic approaches that are currently under development, electrospun membranes gained special attention in the area of skin tissue engineering since they present a similar morphology and structure to that of the ECM. Furthermore, the easy incorporation of drugs or bioactive molecules into nanofibrous mesh can also improve the healing process. In this study, a new bilayer electrospun membrane was produced to cover/protect the wound as well as promoting the wound healing process. The combination of natural and synthetic materials to produce this membrane resulted in a system with suitable physicochemical properties and biological properties to be applied in the pretended biomedical application. Furthermore, the encapsulation SA in EBM avoided biofilm formation at the surface of the nanofibrous matrix. The in vitro assays also revealed that the wound dressing is noncytotoxic and provides a 3D polymeric support to allow cell adhesion and proliferation.The incorporation of growth factors, vitamins, or other biomolecules to enhance the healing processes can be hypothesized in a near future to improve the performance of these membranes. Furthermore, the real-time detection of the wound bed environment parameters, including pH or temperature, through the incorporation of biosensors in the skin substitutes can allow the development smart devices capable of monitoring the healing process in real time.Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ijbiomac.2016.09.080The following is Supplementary data to this article:Resistance of steel I-sections under axial force and biaxial bendingThe plastic criteria for the verification of steel cross-sections resistance are usually based on some basic hypotheses such as the development of plastic hinges, which depend on the interaction between the internal forces and the cross-section shape; therefore, specific equations are required for each type of cross-section.This paper presents new alternative interaction criteria for the analysis of steel I-sections subjected to an axial force and biaxial bending moments, at the elastic or the plastic limit states (as long as buckling phenomena are not involved).The plastic interaction criteria are presented, in a first step, for some particular combinations of the internal forces, such as axial loading with bending about a main axis, and biaxial bending without axial loading. In these cases, they are given by exact equations (within the frame of the hypotheses adopted in this study). All these plastic interaction criteria are compared with the corresponding plastic criteria adopted in the Eurocode 3 (EC3).Afterwards, a simplified global criterion is proposed for the simultaneous combination of an axial force and bending moments about both the main axes of inertia. This new simplified plastic criterion and the corresponding plastic criterion adopted in the EC3 are compared with the exact solution, obtained by a mixed numerical and analytical integration procedure. This comparison shows that this simplified criterion usually leads to results closer to the exact solutions. Some suggestions are then presented to improve the results given by the EC3.► Exact criteria for I-sections under axial force and bending about a main axis. ► Exact criterion for I-sections under biaxial bending without axial force. ► Simplified criterion for I-sections subjected to biaxial bending with axial force. ► Modified EC 3 criterion giving better agreement with the exact solutions.The analysis of the behaviour and limit carrying capacity of a cross-section under biaxial bending is usually a complex problem, which has been studied by many researchers for a long time. A large number of publications may be found, covering the study of structural cross-sections made of different materials (such as reinforced concrete sections In the case of steel sections, a considerable amount of research has been done concerning the study of different types of cross-sections, such as H and I shapes The elastic–plastic methods are currently adopted in modern standard codes of design to estimate the ultimate resistance of some steel structures, since they allow the beneficial effects of yielding in the redistribution of stresses to be taken into account. The analysis of the limit carrying capacity of a cross-section under biaxial bending is simpler than the analysis of its behaviour along the elastic–plastic range, The research works carried out with this purpose have been based on analytical studies Although the results of some numerical models evidence a very good agreement with test results, their practical use for design purposes is limited, since most of them are not currently available and the labour required by the numerical calculations is quite important The interaction criteria between the cross-section internal forces at its plastic limit state depend on the cross-section shape. Consequently, specific analytical expressions are required for each type of cross-sections. However, these analytical expressions are not currently available for some cross-section shapes, or they are defined by means of simplified equations, which do not take in account all the possible scenarios of loading, depending on the combinations of internal forces and relevant geometrical parameters.On the other hand, the existing accurate methods are frequently complex, and difficult to apply in practice. This is often the case when biaxial bending of a cross-section is involved.The design interaction formulae used to check the safety of members and cross-sections subjected to biaxial bending and axial force are usually the result of previous research studies, which are in the origin of those formulae or were dedicated to their discussion and validation. One of these interaction criteria, indicated in Eq. where Mn, |
y and Mn, |
z are the bending moment components, about the cross-section main axes of inertia, associated to an axial load N, and Mo, |
y and Mo, |
z represent the cross-section resistance capacities in simple bending under the axial load N, when Mn, |
z |
= 0 or Mn, |
y |
= 0 respectively. Many solutions have been suggested for the evaluation of the α1 and α2 coefficients or for alterations to Eq. Yet, even if these equations give a good estimation of the cross-section resistance for a large number of practical situations, some research works have pointed out its limitations and have presented alternative solutions, namely under the form of design tables This work presents new interaction criteria for the analysis of I-shaped cross-sections subjected to a combination of an axial force and biaxial bending moments, at the elastic or plastic limit states (as long as buckling phenomena are not involved). Written in a non-dimensional form, these criteria are independent from the cross-section dimensions and steel strength, and from the Unit System used in the analysis presents a general configuration of an I-shaped cross-section under biaxial bending. The v axis is assumed to be the bending axis. Its direction is defined by the cross-section linear segment where the stresses due to biaxial bending are equal to zero.The cross-section b dimension represents its width, parallel to the y axis, and the h dimension corresponds to its height, parallel to the z axis (The values of My and Mz are supposed to be always positive; therefore, the inclination angle α of the bending axis v regarding the y axis is within the limits 0 ≤ |
α |
≤ |
π/2.In the case of uniaxial bending about the strong axis, we have Mw |
= 0, Mz |
= 0, My |
= |
Mv and α |
= 0; if the bending axis is the weak axis, we have Mw |
= 0, My |
= 0, Mz |
= |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.