University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Virginia Tech
|
6.3 Experimental Design
The study presented in this paper examined thirty six PPRVs with varying plug thicknesses and
fill level across six temperatures. Barometric pressure and ambient room temperature were also
captured as covariates to ensure that these environmental conditions did not significantly affect
the experiment. Relative humidity was not captured because the experiment was executed in a
laboratory in which the relative humidity was kept constant at 40%. The varying PMCH fill
levels were applied to the sources to determine the effect of PMCH depletion over time assuming
that a failure of the plug or the shell had not occurred. In order to determine the effect of
temperature on release rate, six different temperatures were applied in random time order over
the course of the study. This experiment was completed in two main parts: assembling the
PMCH release sources and determining the mass flow rate of each release source. The following
section details the experimental parameters from preparing the PPRVs to executing the
experiment. The PMCH release source components and designs will first be discussed.
The main body of the release vessel is comprised of an aluminum cylinder that is 6.35 cm (2.50
in) in length with an outside diameter (OD) of 0.747 cm (0.294 in). The cylinder itself is
comprised of a hollow shell with one end sealed and the other end open to atmosphere. The shell
has a wall thickness of 0.0356 cm (0.0140 in) that is consistent throughout the body of the
cylinder, which gives an inside length of 6.31 cm (2.49 in) and an inside diameter (ID) of 0.711
cm (0.280 in). A schematic of the aluminum shell is displayed in Figure 6.3.
118
|
Virginia Tech
|
Four PMCH fill levels, 0.100 mL, 0.275 mL, 0.425 mL, and 0.600 mL, and four plug
thicknesses, 0.635 cm (0.25 in), 1.067 cm (0.42 in), 1.473 cm (0.58 in), and 1.905cm (0.75 in),
were assigned according to a full factorial design. The fill level and plug thickness increments
were divided equally between the two extremes. In previous preliminary studies, a fill level of
0.500 mL was used. A maximum fill level of 0.600 mL was chosen to determine if there was a
significant difference in release rate with a greater amount of PMCH. The minimum fill level of
0.100 mL was constrained by the lowest accurate measurement that the chosen liquid tight
syringe could achieve. The minimum and maximum plug thicknesses were chosen to represent a
range that could be consistently cut and press fitted with high precision.
Two duplicates were created for each unique set of factors (e.g. two PPRVs are created with
0.100 mL of PMCH and a plug thickness of 0.25 in). The only exception was made for the 0.600
mL PPRVs. Three sources were assembled instead of two for the 0.600 mL PPRVs. All 36
release sources were assembled at one time without any personnel substitutions to ensure
consistency. A detailed listing of the PPRVs can be found in Table 6.1.
120
|
Virginia Tech
|
The PPRVs were then randomly assigned to one of two 500 mL borosilicate beakers filled with
sand resulting in 18 PPRVs per beaker. The sand served as a temperature distribution medium to
ensure that all of the sources maintained a uniform temperature as each temperature treatment
was applied. These beakers were then placed inside a 20 L water bath capable of maintaining
temperatures from ambient room temperature to 100°C with a uniformity of ± 0.2°C, a stability
of ± 0.25°C, and a precision of ± 0.1°C. The six chosen temperatures were applied randomly in
time over the course of approximately 109 days as shown in Table 6.2.
Table 6.2. Temperature treatments.
Time Temperature
Order (°C)
1 35
2 45
3 35
4 25
5 45
6 25
7 30
8 40
9 50
10 50
11 40
12 30
The PPRVs were given 48 hours to equilibrate with each new temperature treatment. After the
48 hour period, the mass of each source was recorded once per day over a period of five days. At
the conclusion of the monitoring period, the change in mass per unit time was determined for the
final analysis.
The experiment uses a strip-plot design model with two factors and two covariates. The first
factor was the combination of plug thicknesses and PMCH fill levels. This factor was applied
randomly to the first whole-plot unit of the PPRVs. Each of the 16 treatment combination levels
was given two replicates. The one exception was the 0.600 mL PMCH fill level which was
assigned three replicates. A similar fill level had already been successfully applied in previous
122
|
Virginia Tech
|
studies. As a result the 0.600 mL treatment was assigned the extra replicates in order to provide a
highly precise base for comparing the effect of the other fill levels.
The second factor was temperature, which was assigned randomly to the second whole-plot unit
of time in weeks. Two replicates were used for each of the six temperature treatment levels. The
response release rate (i.e. change in PPRV mass per unit time) was obtained using the split-plots.
The split-plots were determined using the intersection of the whole-plots. The two covariates, or
variables that were present but not directly measured that may impact the response, were ambient
room temperature and mean sea level barometric pressure. All factor and covariate terms were
assumed to be fixed effects.
Table 6.3 displays a subset of the experiment design. The first and second rows represent PPRV
number 10 and PPRV number 8 respectively with their randomly assigned combination of plug
thickness and PMCH level. The first and second columns display the temperature randomization,
35°C and 45°C, over the course of Weeks 1 and 2 respectively. The remaining portion of the
experiment was a continuation of this table encompassing the remaining sources and treatments
listed previously.
Each temperature treatment was applied uniformly across all the plug thickness - PMCH fill
level combinations. For example, the water bath was set to 35°C during Week 1 for all sources.
No modifications were made to the sources in terms of plug thickness and PMCH fill level as the
new temperature treatment of 45°C was applied during Week 2. Thus, each factor is held
constant across all levels of the other factor defining a strip plot design. The plug thickness -
PMCH fill level combinations were held constant to decrease amount of PMCH needed for the
experiment. The application of a more traditional randomized design, such as a split plot design,
would have required the manufacture of new PPRVs for each temperature treatment to maintain
independent experimental units (EU). This type of experimental design would not have been cost
or labor effective according to the constraints of this study. In order to improve the efficiency of
time utilization, the temperature treatments were applied in a uniform manner across all PPRVs
thus reducing the number of overall trials as well.
123
|
Virginia Tech
|
6.5 Discussion and Conclusions
The results of the PPRV study presented in this paper was successfully able to derive an equation
to predict release rate as a function of temperature and plug thickness. Figure 6.5 through
Figure 6.9 present an overall graphical summary of the PMCH release over the course of the
experiment. The graphs clearly show a distinct difference in the PMCH release rate, which is
represented by the slope of data points, between difference temperatures. A significant difference
can also be seen in Figure 6.9 between the release rates of sources with different plug
thicknesses. These large release rate differences can also be seen in Table 6.5.
Table 6.4 displays the % RSDs of the release rates between PPRVs of the same design
parameters across all temperature treatments. The precision of the release rates varied from 4%
to 25%. The higher % RSDs appear to be concentrated at the two larger plug thicknesses. These
slightly elevated % RSDs may indicate that these release sources were assembled inconsistently.
However, the low % RSDs reflected by the two thinner plug thicknesses contradicts this source
of error considering all of the PPRVs were assembled at once. The sources of the increased
variation from the larger plug thicknesses are more likely derived either from the allotted
equilibrium time or the time interval between mass change measurements.
The PPRVs were given 48 hours to equilibrate with each new temperature treatment before the
mass changes were recorded. If the larger plug thickness sources had not fully equilibrated with
the new temperature, the release rate would have been slightly skewed between the beginning
and end of the measurement periods. This issue would have been intensified if significant
thermal expansion had occurred at the higher temperature. The expansion of the plug along its
longitudinal axis would have caused increased compression across the plug’s cross-section. Any
substantial compression would decrease permeability while increasing equilibrium time.
Similarly, the time interval between measurements, about 24 hours, may not have been sufficient
to produce a mass difference within the range of sensitivity of the analytical balance. Despite the
higher % RSDs, the precision of the PPRVs was still acceptable given the magnitude of the mass
changes in question.
132
|
Virginia Tech
|
Table 6.8 displays the difference in the release rate between the 0.600 mL sources and the
sources with the same plug thicknesses across the different fill levels at each temperature
treatment. The 0.600 mL fill level was proven in past studies to produce a reliable release rate.
However, the effect of PMCH depletion over time was not analyzed. The release rate is expected
to naturally deviate once the vapor volume within the PPRV is reduced below a critical level at
which time the vapor pressure will be insufficient to maintain a stable rate. In order to identify
this critical point, different fill levels were used to represent different time periods within the
lifetime of the release source. Table 6.9 displays the estimated lifetimes of the PPRVs up to an
80% to 95% depletion of its original PMCH volume. Any significant difference between the
release rate of the 0.600 mL sources and the release rate produced by the other fill levels would
indicate the limit of viability for the PPRVs. As can be seen in Table 6.8, the majority of the
release rates vary by less than 10% from the release rate of the 0.600 mL PPRVs. Only a few
exceptions exist in the table. Based on the inconsistent appearance of the higher deviations, the
fill level itself is an unlikely source for this difference. The lack of any repeating pattern suggests
that the logging of the mass change at these points contained an error that artificially increased
the difference in release rate. As such, the comparison shown in Table 6.8 indicates that the
internal PMCH does not affect the release rate within the constraints of the experiment.
The ANOVA analysis presented in Table 6.10 indicates with an alpha level of 0.05 that the plug
thickness, temperature, and the interaction between these two variables had a significant effect
on the PMCH release rate of the PPRVs. The trend between the different plug thicknesses across
temperature treatments in Figure 6.12 shows a rough linear relationship between temperature
and release rate. The release rate was found to be directly proportional to temperature and
indirectly proportional to plug thickness. Some relatively weak interaction effects, which is
represented by the slightly skewed character of the plot, can also be seen in this figure. The weak
interaction effects shown in Figure 6.12 slightly contradicts the conclusion of strong interaction
given by the p-value of the effects test in Table 6.10. This inconsistency is due to the large
sample size coupled with a smaller relative variance between the response from the plug
thicknesses and the temperature. This characteristic gives the effects test a higher power to detect
differences thus resulting in the conclusion of a more significant interaction. The ANOVA
133
|
Virginia Tech
|
results regarding the significance of temperature and plug thickness agree with the results
presented in Table 6.5.
With an alpha level of 0.05, the two covariates, ambient room temperature and barometric
pressure, did not have a significant impact on the release rate. The largest concern between these
two covariates was barometric pressure. If barometric pressure was found to be a significant
effect, it would have needed to be included in the final regression. Theoretically, normal ranges
of barometric pressure should not have a significant impact on vapor pressure, the driving force
of the PMCH release, which was confirmed by this result. The insignificance of the two
monitored covariates confirms that the major environmental factors in the laboratory did not
affect the release rate.
PMCH fill level was also found not to have a significant impact on the release rate. The
insignificance of the PMCH fill level supports the low variance between PPRVs displayed in
Table 6.8. This result strongly suggests that the release rate of the sources over time will not be
affected by the depletion of PMCH given that the silicone plug and aluminum shell maintains
their integrity. The ANOVA analysis supports the conclusion gained from the data in Table 6.4.
The PPRVs were thus free of any major experimental errors.
Using the significant effects, a regression equation was generated to predict the PMCH release
rate as a function of temperature and plug thickness. The coefficient of determination, or model
𝑅2, is 0.94 for fitting the release rate with these two effects and their interaction. In order to
improve the accuracy of the prediction, some second order terms were added in this regression
equation. Other models involving higher level terms were tested as well. However, the added
benefits of the additional terms did not compensate for the increased complexity of the model
and the coefficients. The resulting regression equations are provided as follows where R is the
release rage in milligrams per day (mg/day), A is the plug thickness in centimeters (cm), and T is
the temperature in °C.
134
|
Virginia Tech
|
𝐴 A
R = 0.6681−2.3973( )+0.0383∙T−0.0976∙( )∙T
2.54 2.54
( 6.1 )
𝐴 2
+0.0006∙𝑇2 +3.7104∙( )
2.54
In Equation ( 6.1 ), a day is defined as 24 hours. The plug thickness range is from 0.635 cm (0.25
in) to 1.905 cm (0.75 in) and the temperature ranges from 25°C to 50°C. The model 𝑅2 is 0.94,
indicating that this equation will accurately predict the release rate of PPRVs manufactured and
utilized within the constraints of this study. Although this equation may be extrapolated beyond
the plug thicknesses and temperatures defined in this paper, the accuracy of the outputted release
rate cannot be guaranteed.
Based on the ANOVA analysis in Table 6.10 regarding the covariates, Equation ( 6.1 ) is
accurate through normal variations of barometric pressure such as those created by altitude
changes, weather systems, and ventilation flows. Similarly the insignificant impact of PMCH fill
level suggests that the PPRVs are viable through 80% to 95% of their estimated lifetimes. Figure
6.10 and Figure 6.11 shows that PMCH level does greatly affect the life of the source. Since any
loss of silicone plug or aluminum shell integrity will compromise the release rate, the fill level
should be tailored to fit the conditions present at the intended release area. For example, certain
environmental conditions, such as high ultraviolet exposure, may cause the PPRV to degrade
faster. Although this study was able to determine the response of the PPRV based on controlled
laboratory conditions, an actual field deployment may elicit a different response. As a result, an
underground mine field evaluation of the PPRV will be conducted to determine the performance
of the sources in the field.
135
|
Virginia Tech
|
Chapter 7: Field test of a perfluoromethylcyclohexane
(PMCH) permeation plug release vessel (PPRV) using a
dual tracer deployment in an underground longwall mine
ABSTRACT: Perfluoromethylcyclohexane (PMCH) has shown to be a viable alternative to the
widely used tracer gas sulfur hexafluoride (SF ). PMCH and SF were released in a Midwestern
6 6
underground longwall mine. The operators of this mine graciously allowed full access to an
active longwall panel during two stages of its advance, designated as Phase I and Phase II, to
perform this study. This paper presents a study designed to determine the feasibility of deploying
a PPRV in an underground environment for tracer gas studies. SF was also released in parallel
6
in a full ventilation characterization of the longwall panel to examine the ability of PMCH to
compliment SF . The results of this study not only showed the PPRV to be a feasible tracer
6
release system in an underground environment but also highlighted the advantages of a dual
tracer release.
141
|
Virginia Tech
|
7.1 Introduction
Sulfur hexafluoride (SF ) has been the predominant tracer gas used in underground mine
6
ventilation studies for over 30 years (Thimons, Bielicki, and Kissell 1974). However, the ability
of SF to function as the sole tracer is being hindered by two main issues: the increasing scale
6
and complexity of mine ventilation systems along with the steadily growing background
concentration of SF in the atmosphere (Levin et al. 2010, Geller et al. 1997, Ravishankara et al.
6
1993, Maiss et al. 1996). In order to mitigate these issues, recent studies have identified the
compound perfluoromethylcyclohexane (PMCH) as a viable supplement and compliment for
SF .
6
PMCH is a perfluorinated cyclic hydrocarbon and is categorized as a perfluorocarbon tracer
(PFT) due to its chemical inertness, low toxicity, and trace level background presence in the
environment. These properties make PMCH suitable for use as a tracer gas (Dietz 1991, Watson
et al. 2007). Compounds of this type have been widely implemented in heating, ventilation, and
air conditioning (HVAC) as well as atmospheric monitoring studies (Dietz 1991) but not yet in
the field of underground mine ventilation. PMCH exists as volatile liquids at room temperature
and pressure (National Institute of Standards and Technology 2011). This physical property
prevents PMCH from being released using conventional techniques designed for gases. Previous
studies have presented a permeation plug release vessel (PPRV) designed to convert PMCH from
a liquid to a vapor and release it into a flow stream in a controlled manner. Although many
laboratory studies have been conducted using the PPRV, it has not yet been utilized in an
underground mine. This paper presents the use of a simultaneous, steady-state release of SF and
6
PMCH in a Midwestern underground longwall mine. The purpose of this study was to determine
the feasibility of the PPRV as a reliable system to release PMCH and also to examine its ability
to compliment SF .
6
142
|
Virginia Tech
|
7.2 Background
PMCH is a perfluorinated cyclic hydrocarbon whose chemical structure is composed of
perfluoroalkanes (Watson et al. 2007). Compounds of this type are biologically inert, chemically
inert, and thermally stable (F2 Chemicals Ltd. 2011). The inert, non-reactive, and non-toxic
nature of PMCH makes it an ideal choice as a tracer gas. PMCH is comprised of seven carbon
atoms and fourteen fluorine atoms, which gives it a chemical formula of C F . The molecular
7 14
structure of PMCH is displayed in Figure 7.1.
Figure 7.1. Molecular structure of PMCH.
The PMCH molecule is composed of two main parts, the cyclohexane ring and the methyl group
bonded off to the side. This fully fluorinated molecule has a molecular weight of 350 g/mol and a
boiling point of 76°C (169°F). The volatility of PMCH allows it to vaporize even at relatively
lower temperatures. Once in a vapor state, PMCH, will remain a vapor even through cooler
temperatures (National Institute of Standards and Technology 2011). Another advantage of
PMCH is its detectability by a GC even at low concentrations. This ability stems from PMCH’s
low ambient background in the atmosphere with concentrations in the parts per quadrillion
(PPQ) (Cooke et al. 2001, Simmonds et al. 2002, Watson et al. 2007) and its high detection
sensitivity derived from its molecular electronegativity (Simmonds et al. 2002).
The basic concept of the PPRV used in this field study was initially introduced by Brookhaven
National Laboratory (BNL). The basic PPRV design consists of a hollow aluminum cylinder
with one end of the cylinder opened to the atmosphere and the other end closed. Liquid PMCH is
143
|
Virginia Tech
|
The high flow resistance caused by the silicone plug produces a pseudo-closed system allows the
PMCH to reach dynamic equilibrium within the source. This equilibrium produces a steady
pressure differential between the inside and the outside of the vessel equivalent to the vapor
pressure produced at the ambient temperature. This differential causes the vapor PMCH to
steadily diffuse through the silicone plug. The release rate will remain consistent as a function of
temperature as long as the integrity is maintained and the compression of the plug is not
excessive (Jordan and Koros 1990).
Vapor pressure is affected most significantly by ambient temperature and is independent of
atmospheric pressure. Thus the release rate can be predicted as a function of temperature. The
thickness of the silicone plug also greatly affects PMCH diffusion as flow resistance increases
with a thicker medium. As a result, a previous long-term study of the PPRV displayed in Figure
7.3 produced the following equation to compute the release rate of PMCH as a function of
temperature and plug thicknesses.
𝐴 A
R = 0.6681−2.3973( )+0.0383∙T−0.0976∙( )∙T
2.54 2.54
( 7.1 )
𝐴 2
+0.0006∙𝑇2 +3.7104∙( )
2.54
This paper outlines a study utilizing the PPRV presented in Figure 7.3 in conjunction with
Equation ( 7.1 ) to determine the feasibility of deploying this system in an underground
environment for tracer gas studies. SF was also released as a part of this study for the full
6
ventilation characterization of the target longwall panel in conjunction with PMCH to examine
the ability of PMCH to compliment SF . A dual tracer release of this nature in an underground
6
mine has not been attempted prior to this study.
145
|
Virginia Tech
|
7.3 Longwall Mine Overview
PMCH and SF were released in a Midwestern underground longwall mine. The operators of this
6
mine graciously allowed full access to an active longwall panel during two stages of its advance,
designated as Phase I and Phase II for this study. The ventilation around the panel is facilitated
using a three entry headgate, a single entry tailgate, a three entry bleeder, and a fringe ventilation
path. This mine utilizes a hybridized bleederless ventilation system to provide fresh air to the
panel. As such, the tailgate consists of a single return entry that is not connected to the bleeders
at the rear of the panel. The fringe ventilation system is another unique aspect of this type of
ventilation system. A small amount of intake air is provided to the outside edges of the gob to
ventilate any accumulation of gases. This ventilation stream only flows around the outside of the
gob and is not designed to penetrate into the gob itself to limit the potential for spontaneous
combustion. The fringe ventilation is directed to the rear of the panel where it joins the bleeders
through small pipes located in the headgate and the tailgate.
Intake air is delivered through three entries in the headgate. The three intake branches combine
to carry fresh air both across the longwall face and into the fringe. The ventilation flow across
the longwall face is carried into the tailgate and then exits the mine through the main return. A
basic schematic diagram of the ventilation system is displayed in Figure 7.4.
146
|
Virginia Tech
|
7.4 Experimental Design
As previously introduced, the tracer study of this panel was accomplished using a simultaneous
steady-state release of SF and PMCH. The tracer study was completed in two separate phases,
6
Phases I and II, representing the ventilation system near the start of the panel and near the end of
the panel. Phases I and II were executed consecutively with a delay of approximately two months
(50 days) between studies. During this delay, the longwall advanced an additional 802 m (2,630
ft) from the Phase I position. The tracer flow around the panel was monitored at nine locations.
These nine locations represented all of the primary ventilation branches serving the longwall
panel. A description of the release points (RP) and sampling points (SP) are presented in Table
7.1.
Table 7.1. Release point and sampling point descriptions.
Station Name Description
RP1 Bleeder fan intake entry inby the last open crosscut (SF )
6
RP2 Longwall shield adjacent to the tailgate (PMCH)
SP1 Main fan intake entry just outby the face
SP2 Belt entry outby the face
SP3 Fringe entry
SP4 Center of the face
SP5 Headgate bleeder tap
SP6 Tailgate bleeder tap
SP7 Fresh air entry in the bleeders
Consists of the SP8A, SP8B, and SP8C sampling points
SP8
representing the three tube bundles in the tailgate
SP9 Tailgate entry outby the face
All of the sampling points were located in open airways except for SP8. SP8 represents the outlet
of a tube bundle system consisting of three tubes extending various distances into the gob. These
tubes were placed in order to detect any flow communication between the longwall face and the
gob. Both phases of the study utilized the same relative locations points located at different
absolute locations based on the position of the longwall. The general locations of the release
point used for both tracer study phases are graphically represented in Figure 7.5. Detailed views
of the release and sampling points are presented in Figure 7.6 and Figure 7.7.
148
|
Virginia Tech
|
Figure 7.7. Detailed view of the SP5, SP6, and SP7 sampling points.
Figure 7.6 shows that SF was released in the intake entry supplied by the bleeder fan just inby
6
the last open cross-cut while PMCH was released near the tailgate from the last longwall shield.
PMCH was released at this location in order to isolate the PMCH release to the single tailgate
entry. These two separate release points were chosen not only to determine the performance of
the PPRV but to also examine the multi-zone analysis potential of a dual tracer release.
From the assigned intake entry, SF was deployed using a mass flow controller that provided a
6
steady stream of the tracer at 200 standard cubic centimeters per minute (SCCM). A SCC of gas
is defined as the mass of the gas that occupies a cubic centimeter of volume at standard
temperature and pressure (STP). Assuming ideal gas behavior and a SF density of 6.0380 g/L at
6
a STP of 25°C and 14.696 psia as defined by the manufacturer, the 200 SCCM mass flow is
equivalent to 1.21 g/min.
100 PPRVs filled with 0.600 mL of PMCH and press-fitted with 0.635 cm (0.25 in) plugs were
placed at the last longwall shield. This number of PPRVs were deployed to provide an adequate
concentration of PMCH in the ventilation flow stream to be detected using a GC. For Phase I, the
150
|
Virginia Tech
|
PPRV bundle was set on an elevated flat area adjacent to the shearer’s electrical track. For Phase
II, the PPRV bundle was suspended from one of the longwall shield’s hydraulic arms. Since the
vapor diffusion of PMCH from the PPRV is a function of ambient temperature, the release rate
could not be actively controlled. As previously introduced, the release rate is dictated by the
following equation.
𝐴 A
R = 0.6681−2.3973( )+0.0383∙T−0.0976∙( )∙T
2.54 2.54
( 7.1 )
𝐴 2
+0.0006∙𝑇2 +3.7104∙( )
2.54
As a result, the ambient temperature was monitored throughout the study to provide an average
release rate. Barometric pressure and relative humidity were also monitored to record any drastic
shifts. However, these two variables are not expected to significantly impact the PPRV.
After the tracer gases were released, air samples were taken at regular intervals for several hours.
For Phase I, air samples were taken at 15 min intervals for a six hour period. Although the tracers
should have rapidly achieved a homogeneous distribution in the open entries, this time period
was allotted to capture SF at SP5 and SP6 in the bleeders. The high resistance, low flow design
6
of the fringe ventilation branch was expected to delay the travel of SF thus increasing the time
6
required to reach a steady-state concentration. As a result, the six hour period was selected in
order to accommodate a reasonable amount of tracer travel time. The SP8 tube bundles in this
study extended approximately 300 ft, 400 ft, and 500 ft into the gob from the tailgate. The
longwall was idled for the entire duration of the Phase I study.
For Phase II, air samples were taken at 30 min or 60 min intervals, depending on the sampling
location, for a seven hour period. The original study design called for a 12 hour sampling period
to account for the increased linear travel distance in the fringe of approximately 802 m (2,630 ft).
However, logistical complications decreased the available time to seven hours. The scheduled
study period intersected with the activation of the longwall near the end of allotted time. As a
result, Phase II represented six hours of idled longwall ventilation and one hour of active
151
|
Virginia Tech
|
longwall ventilation. The SP8 tube bundles for this phase extended 200 ft, 300 ft, and 400 ft into
the gob from the tailgate.
7.5 Experimental Results
Due to the large amount of data collected for this study, the results are organized in subsections
based on study phase and tracer gas.
7.5.1 Phase I SF Results
6
The Phase I vacutainer samples were analyzed using a gas chromatograph (GC) equipped with
an electron capture detector (ECD) for SF . PMCH was analyzed using a slightly modified
6
approach that will be discussed in later sections. The GC was installed with a 30 m porous layer
open tubular (PLOT) column coated with sodium sulfate deactivated alumina oxide. The column
has an internal diameter (ID) of 0.25 mm and a film thickness of 5 µm. Table 7.2 displays the
method parameters used to analyze SF .
6
Table 7.2. GC analytical method for Phase I.
Parameter Description
Sample Injection Size 100 µL
Carrier Gas He
Injector Temperature 150°C
Split Ratio 30:1
Linear Velocity 35 cm/s
Isothermal Column Temperature 65°C
Detector Temperature 200°C
Make-up Flow N at 30 mL/min
2
Total Program Runtime 2.5 min
In order to determine the steady-state concentrations at the sampling points, a calibration curve
was generated within the range of the data. The calibration curve is represented by the equation
𝐴 = 205.53𝐶 +4414.48 where A is the GC area count response in Vmin and C is the SF
6
concentration in parts per billion by volume (PPBV). This equation was produced from a set of
152
|
Virginia Tech
|
As expected, the sampling points located in open airways showed the presence of SF . The
6
presence of the tracer at these locations confirms that the flow paths represented by sampling
points SP1 – SP4, SP5, SP6, and SP9 are directly connected to RP1. These results also confirm
that the air is flowing toward these sampling points from this release point. The processed data is
displayed in Table 7.3.
Table 7.3. Phase I summary of results.
SF
6
Sampling Concentration
Point (PPBV) % RSD
SP1 159.09 11.45
SP2 168.23 7.91
SP3 167.13 7.18
SP4 173.98 13.43
SP5 152.52 9.57
SP6 173.76 20.83
SP7 0.00 N/A
SP8-1 90.96 125.92
SP8-2 58.85 63.15
SP8-3 67.30 48.68
SP9 140.44 18.38
The % RSD, or percent relative standard deviation, shown in Table 7.3 is a measure of analysis
precision between subsamples at each sampling point. All individual samples were analyzed in
triplicate in order to determine % RSD to indicate GC analysis precision. The % RSD between
the different sampling points, excluding the SP8 tube bundle, showed an acceptable level of
consistency given the concentrations at which the tracer was present. The overall % RSD for all
points except the tube bundles was 7%, demonstrating that all locations were sampled
consistently and that a similar steady-state tracer concentration was achieved. The results of the
tracer analysis are also displayed graphically in Figure 7.10 through Figure 7.12.
154
|
Virginia Tech
|
7.5.2 Phase I PMCH Results
The average temperature recorded at SP4 was used in Equation ( 7.1 ) to determine the average
release rate of PMCH. Based on an average temperature of 24°C (75.4°F) and a PPRV plug
thickness of 0.635 cm (0.25 in), the release source bundle was expected to produce an average
release rate of 7.01·10-5 g/min. Using the surveyed quantity of 1,780 m3/mi (63,000 CFM) from
the tailgate, the steady-state concentration of PMCH at SP9 is expected to be 3.35 PPTV.
Due to the magnitude of the concentration, GC-ECD provided an inadequate detection sensitivity
for the samples. As a result, a slightly different analysis technique was developed to quantify
PMCH. The vacutainer samples were analyzed using a gas chromatograph (GC) equipped with a
single quadrapole mass spectrometer (MS) modified for negative ion chemical ionization (NCI).
The chemical ionization gas was methane (CH ). The GCMS was modified for NCI because
4
traditional electron impact (EI) ionization provided an inadequate detection sensitivity. PMCH
has a high electron affinity thereby facilitating the formation of negative ions. Given the soft
ionization mechanism of NCI, the preservation of PMCH’s negative molecular ion provided an
exceptional detection sensitivity.
The NCI-GCMS, identical to the GC-ECD systems described in Section 7.5.1 and Section 7.5.3,
was installed with a 30 m porous layer open tubular (PLOT) column coated with sodium sulfate
deactivated alumina oxide. The column has an internal diameter (ID) of 0.25 mm and a film
thickness of 5 µm. Table 7.5 and Table 7.6 display the method parameters used to analyze
PMCH.
157
|
Virginia Tech
|
Table 7.5. GC method used for PMCH analysis.
Parameter Description
Sample Injection Size 150 µL
Carrier Gas He
Injector Temperature 150°C
Split Ratio Program Splitless for 0.50 min and then split at
30:1 to sweep injector port.
Linear Velocity 30 cm/s
Column Temperature Program 50°C initial temperature is held for
0.10 min. Temperature then increases
at 40°C/min to 170°C. 170°C is held for
1.00 min. Temperature then decreases
at 60°C/min to 120°C. 120°C final
temperature is held for 8.07 min.
Total Program Runtime 13.00 min
Table 7.6. NCI-MS method used for PMCH analysis.
Parameter Description
Interface Temperature 185°C
Ion Source Temperature 195°C
Threshold 260
MS Scan Mode SIM
SIM Target m/z 350
Scan Time 11.51 min to 13.00 min with
an event time of 0.42 sec
Total Program Runtime 13.00 min
The added complexity of the splitless NCI-GCMS method compared to the GC-ECD method
was necessary to achieve three primary goals: focusing of the PMCH aliquot within the column,
separating the PMCH peak from contaminants with similar molecular weights, and removing
background noise through selected ion monitoring (SIM). A chromatogram produced by this
method is displayed in Figure 7.13.
158
|
Virginia Tech
|
7.5.3 Phase II SF Results
6
The SF vacutainer samples were analyzed using a GC-ECD for Phase II. PMCH was analyzed
6
using a slightly modified approach that will be discussed in later sections. The GC was installed
with a 30 m porous layer open tubular (PLOT) column coated with sodium sulfate deactivated
alumina oxide. The column has an internal diameter (ID) of 0.25 mm and a film thickness of 5
µm. Table 7.9 displays the method parameters used to analyze SF .
6
Table 7.9. GC analytical method for Phase II.
Parameter Description
Sample Injection Size 100 µL
Carrier Gas He
Injector Temperature 150°C
Split Ratio 30:1
Linear Velocity 30 cm/s
Isothermal Column Temperature 50°C
Detector Temperature 200°C
Make-up Flow N at 30 mL/min
2
Total Program Runtime 2.5 min
In order to determine the steady-state concentrations at the sampling points, a calibration curve
was generated within the range of the data. The calibration curve is represented by the equation
𝐴 = 596.52𝐶 −7558.77 where A is the GC area count response in Vmin and C is the SF
6
concentration in PPBV. This equation was interpolated from a set of laboratory mixed standards
with a regression (R2) value of 0.99. The calibration curve is displayed in Figure 7.1.
162
|
Virginia Tech
|
The volumetric flow of SF at RP1 was determined using the ambient atmospheric conditions
6
recorded at SP1 with a 200 SCCM mass flow (SF ) assuming ideal gas behavior.
6
7.5.4 Phase II PMCH Results
The average temperature recorded at SP1 was used in Equation ( 7.1 ) to determine the average
release rate of PMCH. Based on an average temperature of 21.6°C (70.8°F) and a PPRV plug
thickness of 0.635 cm (0.25 in), the release source bundle was expected to produce an average
release rate of 6.20·10-5 g/min. Using the surveyed quantity of 1,900 m3/min (67,100 CFM) from
the tailgate, the steady-state concentration of PMCH at SP9 is expected to be 3.60 PPTV.
The vacutainer samples were analyzed using a NCI-GCMS with the same method outlined in
Section 7.5.2. The results of the NCI-GCMS analysis of PMCH are displayed in Table 7.12 and
Figure 7.21. The SP8 tube bundle data are not presented because the concentration of PMCH
was below the LOD. The following data was produced using the same calibration curve outlined
in Section 7.5.2.
Table 7.12. Phase II PMCH results.
PMCH
Sample Time Area Concentration
Number (HH:MM) Counts (PPTV)
SP9-7 2:30 992.3 13.22
SP9-8 3:00 932.0 12.37
SP9-9 3:30 591.3 7.63
SP9-10 4:00 858.7 11.35
SP9-11 4:30 515.0 6.56
SP9-12 5:00 468.3 5.91
SP9-13 5:30 345.0 4.19
SP9-14 6:00 356.0 4.34
SP9-15 6:30 314.3 3.76
SP9-16 7:00 306.7 3.66
SP9-17 7:30 282.0 3.31
SP9-18 8:00 340.0 4.12
SP9-19 8:30 1241.0 16.68
SP9-20 9:00 1531.7 20.73
167
|
Virginia Tech
|
7.6 Discussion and Conclusions
7.6.1 SF Tracer Characterization
6
The front of the SF stream was immediately apparent at the majority of the sampling points, for
6
both Phases I and II, which can be seen in Figure 7.10 and Figure 7.17 respectively. The only
exceptions were SP7, which was located in an isolated airway, SP5 and SP6, which were located
at the bleeder taps, and the SP8 tube bundles. A discussion of these points is provided later in
this section. At the sampling points showing immediate tracer presence, SP1 – SP4 and SP9, the
concentration distribution shows that the SF required approximately 30 min to become evenly
6
distributed in the Phase I and II flow streams. This rapid equilibration time demonstrates that the
airflow was fully turbulent and traversing the connecting distances quickly. Although air
quantities could not be derived from the tracer concentrations at any point other than SP1, the
presence of SF coupled with their uniform concentrations at the sampling points demonstrates
6
that the tracer samples were taken in fully developed turbulent flows for both phases of the
study. In addition, no significant differences were found between the flow patterns of Phases I
and II at the locations represented by SP1 – SP4 and SP9.
Using the concentration of SF at SP1 and the volumetric flow rate of SF from RP1, the airflow
6 6
quantity at SP1 is calculated to be 55 kcfm for Phase I and 130 kcfm for Phase II. For Phase I,
the calculated value was approximately 26 kcfm less than the surveyed airflow quantity of 81
kcfm. For Phase II, the calculated value was approximately 46 kcfm more than the surveyed
airflow quantity of 84 kcfm. This discrepancy between the surveyed value and the calculated
value of both phases may have resulted from one of three primary reasons: the barometric
pressure and temperature readings did not have adequate accuracy, the ventilation survey data
contained a discrepancy, or the vacutainer samples were taken at a point in the entry’s cross-
section with a particularly low SF concentration. Based on the proximity of the sampling point
6
to the air direction change, the physical air sample was most likely taken in an area of layered
SF concentrations in the entry for both phases. This conclusion can be inferred from the erratic
6
concentration changes displayed in Figure 7.10 and Figure 7.17 for Phases I and II respectively.
169
|
Virginia Tech
|
The SP1 flow pathway separates into the SP2, SP3, and SP4 flow pathways. From SP1, the
average tracer concentration reflects a volumetric SF flow of 0.013 cfm for Phase I and 0.007
6
cfm for Phase II being delivered from SP1 to the aforementioned branches. These volumetric
flows were computed assuming ideal gas behavior of the 200 SCCM mass flow at the recorded
environmental conditions. Based on flow conservation, the volumetric tracer flow from SP1
should equal the sum of the volumetric tracer flows from SP2, SP3, and SP4. However, the total
flow of the three branches did not balance in terms of quantity for either phase. The Phase I flow
is approximately 0.004 cfm greater than the measured tracer flow at SP1. Similarly, the Phase II
flow is approximately 0.001 cfm greater than the measured tracer flow at SP1. In order to
determine the tracer flow at SP4, the airflow at SP9 was used due to the absence of survey data
on the longwall face. These discrepancies equate to an error of 33% and 16% for Phases I and II
respectively. The flow discrepancy indicates that SF was added to one of the airflow streams.
6
The non-conservation of tracer quantity may be the result of leakage from the No. 2 entry to the
No. 3 entry that occurred between SP1 and SP3, of an unknown recirculating event at one of the
sampling points, or of an error in sampling. Incidentally, a stopping door near SP1 was left open
during both studies. This door was located between SP2 and SP3. A short circuit between the
No. 2 and No. 3 entries was likely created as a result of the open door. Some of the tracer gas
from RP1 would have been diverted into the No. 3 entry prior to reaching SP2. Leakage is thus
the most likely explanation of this abnormality. The loss of tracer to SP3 in the airflow stream
would have occurred prior to reaching SP2 thus causing an artificially inflated value in the sum
of tracer flows from SP2, SP3, and SP4. However, given the low magnitude of the tracer
concentration, these errors may not be significant for either phase.
The lack of tracer presence at SP7 for Phases I and II was expected and confirms that this
location did not interact significantly with any of the airways directly downwind from the release
point. SP7 was the only sampling location at which SF was not detected. Both SP5 and SP6
6
showed a gradual increase of tracer gas over time. These points were located at the headgate and
tailgate bleeder taps respectively. Air is delivered to these taps through the edges of the gob (i.e.
the fringe) from the SP3 intake branch. As can be seen in Figure 7.11, SP7 and SP8 required
approximately four hours and six hours respectively to reach an equilibrium tracer concentration
170
|
Virginia Tech
|
for Phase I. The time required to reach a steady-state concentration was not captured for Phase II
due to logistical complications. Based on the trend in Figure 7.18, SP5 and SP6 would have
required more than seven hours to reach an equilibrium tracer concentration for Phase II.
The slow increase of tracer concentration over time reflects the low flow, high resistance design
of the fringe ventilation system. The arrival time of the tracer front between Phases I and II was
approximately 1.5 hours longer for Phase II than for Phase I. This comparison is displayed in
Figure 7.20. The longer travel time is expected due to the fact that the longwall face had
advanced further down the panel thus resulting in a longer distance from the bleeders.
The samples collected from the Phase I 300 ft, 400 ft, and 500 ft tube bundles at SP8 showed the
presence of SF but this data is inconclusive. The SF concentrations from the tube bundle are
6 6
presented in Figure 7.12. The random appearance of SF at various magnitudes in the tube
6
bundles suggest the presence of a leak in the tube system. This leak may have been in the tube
itself, in the sampling system, or a combination of both. Any one of these three scenarios would
have compromised the tracer samples thus contaminating these samples. As such, the results of
this study could not identify and inter-zonal interaction between the face ventilation and the gob
for the Phase I study.
The samples from the Phase II 200 ft, 300 ft, and 400 ft tailgate tube bundles at SP8 are
displayed in Figure 7.19. The SF concentrations from the tube bundles does indicate some
6
interaction between the longwall face ventilation and the gob at the 200 ft (SP8A) and 300 ft
(SP8B) tube bundles. Although Figure 7.19 shows that a low concentration of SF was
6
immediately present at each of the tube bundles, a leak in the tube system was not likely given
the low SF magnitude and the consistency of the concentration over time. This behavior was
6
reflected by SP8B and SP8C. The double-sided needles used to take the tube bundle samples
have a small amount of headspace present in its internal volume. The consistent, low
concentration of SF can be attributed to the internal volume of the needle and can be considered
6
as zero SF presence.
6
171
|
Virginia Tech
|
The SP8A tube bundle showed a gradual increase of SF over time with the tracer first appearing
6
between two to three hours after the initial release. The increase in tracer concentration and
apparent achievement of steady state concentration suggests that the face ventilation did have
some flow 61 m (200 ft) into the gob. The lack of significant SF presence at SP8B and SP8C
6
suggests that the interaction with the gob is restricted to a boundary located between SP8A and
SP8B. SF does not penetrate deeply enough into the gob to affect SP8C located 122 m (400 ft)
6
into the gob.
A sharp increase in tracer concentration was found at SP8B at the end of the sampling period.
This sudden buildup of SF may have been caused by two primary reasons: a collapse in the gob
6
opened a free volume between SP8A and SP8B thus producing a direct flow path between the
inlets of these two tube bundles or a failure in the tube bundle/sampling system had occurred.
This sharp rise in tracer concentration coincided with the start of the longwall for the morning
shift. As such, either scenario was probable.
The SF portion of the tracer study for both studies was free of major errors as reflected by the
6
precision of the data and analysis technique. The qualitative tracer gas data showed that the
ventilation flow streams did travel from RP1 to the expected branches. The lack of detectable
tracer presence at SP7 confirmed that this point was located in an isolated airway with no
communication from the flow streams downwind of RP1. The time required to achieve a steady-
state tracer concentration at the sampling points indicated that SP1, SP2, SP3, SP4, and SP9 were
located in open, unobstructed airways while SP5 and SP6 were located in areas with low,
restricted airflow. Some interaction between the gob and the longwall face ventilation was found
for Phase II but not for Phase I. This interaction during Phase II was derived from the SP8 tube
bundle system.
7.6.2 PMCH PPRV Evaluation
The PPRVs successfully released a detectable level of PMCH for both Phases I and II. The
PMCH analysis results are displayed in Section 7.5.2 and Section 7.5.4. The Phase I results did
not provide enough useable quantitative data for a thorough evaluation of the PPRVs. The
172
|
Virginia Tech
|
majority of the NCI-GCMS data did not provide a sufficient signal to noise ratio for
quantification. As a result, only the quantifiable vacutainer samples are presented. However, the
presence of PMCH was detected in all of the SP9 tailgate samples. The positive presence of
PMCH at an unquantifiable concentration suggests that the PPRVs were located in area of the
longwall face that did not allow for homogeneous mixing of the tracer, the SP9 samples had been
compromised, or a combination of both. Either event may have occurred given the placement of
the sources at a relatively complex area adjacent to the belt conveyor combined with the fact that
the samples were not analyzed for several weeks due to equipment complications. Despite the
lack of quantitative data provided by Phase I, the presence of PMCH does show that the PPRVs
were releasing PMCH.
The Phase II results did successfully provide quantitative data derived from the PPRVs. The
NCI-GCMS analysis results are presented in Table 7.12. The Phase II PMCH data is also
presented graphically in Figure 7.21. The distribution of PMCH over time demonstrates that
concentration did reach steady-state prior to the activation of the longwall shearer. Figure 7.21
shows that PMCH required approximately three hours to reach an equilibrium concentration.
This increased equilibrium time as compared to SF may have been due to the PPRVs
6
acclimating to the new environmental conditions, the high molecular weight of PMCH, or the
manner in which the PPRVs were deployed.
Prior to the study, the PPRVs were transported underground approximately 12 hours prior to the
study to allow for temperature acclimation. However, the PPRVs were not placed at the release
point. As a result, the temperatures between the two locations may have been sufficiently
different to require additional time to reach a steady release rate. In contrast to Phase I, the
PPRVs were hung from one of the rear hydraulic arms of the longwall shield for Phase II. The
PPRVs were set up in this manner because the Phase II study period intersected with the start of
the longwall shearer. As a result, the PPRVs had to be situated in a manner that allowed the
bundle to move with the advance of the shield. The rear of the longwall shield did not have a
large amount of airflow at the time of the release. The inadequate flow coupled with the high
molecular weight of PMCH may have also caused the higher relative equilibration time.
173
|
Virginia Tech
|
The large spike in tracer concentration occurred just after the start of the longwall shearer. The
increase may be been caused by a rise in temperature, a change in ventilation flow, or the
opening of a cavity that had been accumulating with PMCH prior to the movement of the shield.
Due to the lack of observations at the PMCH release point, the actual cause for the sharp increase
is unknown. Despite this unexpected occurrence, the PMCH concentration did achieve and
maintain steady-state for approximately three hours.
Based on the surveyed air quantity of 1,900 m3/min (67,000 cfm), an average temperature of
21.6°C (70.8°F), and an average absolute barometric pressure of 63,400 Pa, the steady-state
PMCH concentration was expected to be 3.60 PPTV. The average steady-state concentration
measured from the SP9 vacutainer samples was 3.90 PPTV. The difference between the expected
and measured concentration equates to an error of 8%. At the PPTV concentration level, an 8%
deviation effectively denotes a zero difference between the expected and observed values. Using
the detected steady-state concentration of PMCH, the air quantity is calculated to be 1,755
m3/min (62,000 CFM), which is validated by the results of the ventilation survey.
As such, the PPRVs not only performed according to their design specifications but also
supported the ability of Equation ( 7.1 ) to predict the release rate of the PPRV. The results of
Phase II suggest that the PPRVs will perform as expected in field conditions when placed in an
area of adequate turbulent flow. The results of this study also demonstrate the potential of PMCH
to supplement SF in tracer gas studies. The location of the SF release point prevented any
6 6
useful quantitative data to be derived from the tailgate. The location of the PPRVs remediate this
problem by providing a secondary release point. Additionally, since SF and PMCH do not
6
interfere with each other when using a GC, these two tracers can be simultaneously analyzed and
sampled using the same medium. Given the simplicity of the PPRV, several benefits were
realized during the execution of the study. Two of the most prominent advantages were rapid
setup and potential for deployment in inaccessible areas, such as the gob. These two advantages
were afforded by the simplicity of the PPRV coupled with it passive release mechanism. The
PPRV is thus shown to be a feasible system within the parameter of this study for the release of
PMCH in underground mine ventilation studies.
174
|
Virginia Tech
|
Chapter 8: Conclusions and Future Work
The study presented in this paper sought to develop a PPRV system that was not only flexible
but that allowed for the controlled release of PMCH in an underground mine environment. In
order to complete this objective, extensive laboratory and field studies were designed to bring the
PPRV from conception to reality. The PPRV development and evaluation study successfully
developed a PMCH calibration curve preparation technique for GC, completed an extensive
evaluation of potential PPRV designs, deployed a potential PPRV in a small-scale turbulent
environment, and completed a field study that validated the final PPRV system. A detailed
explanation of each major topic was provided in the preceding chapters.
The multi-dilution calibration curve technique was found to balance level of difficulty with
precision thereby producing a viable method for producing PMCH standards. This calibration
curve technique proved to be repeatable in the PPRV evaluation studies that followed. The
preliminary PPRV designs were found to produce a reliable PMCH release rate and to be free of
manufacturing defects. This preliminary study found that plug thickness and temperature
significantly affected the release rate. From this data, a comprehensive strip-plot experiment was
designed to derive an equation to calculate PMCH release rate as a function of plug thickness
and ambient temperature. The results of the final PPRV development study successfully derived
an equation to predict release rate as a function of temperature and plug thickness. Additionally,
this study confirmed that release rate was independent of both barometric pressure and internal
PMCH volume. The PPRV will thus perform consistently across a wide range of elevations and
maintain its release rate throughout the majority of its expected lifetime.
In a parallel study, the newly developed PPRVs were evaluated in a controlled turbulent
environment. The result of the turbulence experiment showed that the PPRVs were highly
precise over a range of flow quantities. Air flows within the transitional characteristic zone were
found to cause the PPRV to behave unreliably. Since transitional type flows contain both laminar
and turbulent elements, this erratic behavior agrees with the fact that PMCH has a high layering
potential from its high molecular weight. Although laminar flows were not included in the
evaluation, poor PPRV performance is also expected in the laminar zone due to the inadequate
177
|
Virginia Tech
|
mixing of PMCH in the flow stream. The final study presented in this paper evaluated the
performance of the PPRV in a Midwestern underground longwall mine.
In this field study, the PPRVs performed not only according to their design specifications but
also validated the equation derived in the final development study to predict the release rate of
the PPRVs. This study also represented the first time that a dual PMCH – SF release has been
6
conducted in an underground mine. The execution of the field study proved to be relatively
straightforward when using the PPRVs when compared to SF . The simplicity of the PPRV
6
allowed rapid setup of the system. The PPRVs’ passive release mechanism also allowed the
potential for deployment in inaccessible areas of the mine such as the gob. The results suggest
that the PPRVs will perform as expected in an underground mine if the sources are given
sufficient time to equilibrate with the environmental conditions. The release sources must also be
placed in an area of adequate turbulent flow and be deployed in adequate numbers to satisfy the
LOD.
The operating principles of the PPRV show great potential for adaptation to release other similar
perfluorocarbon tracers in underground mines. Based on the combined results of this overall
study, the PMCH PPRV developed in the preceding chapters was found to be a feasible release
system for underground tracer gas studies. This feasibility is, however, limited to the parameters
of each of the aforementioned studies. The PPRVs can be enhanced with additional studies
designed to develop standard operating procedures (SOP), to examine the effects of release
source diameter and plug compression, to examine the effects of sub-ambient temperatures, to
investigate the impacts of different underground environments on the PPRV, to explore different
techniques for dual tracer releases, as well as to improve trace-level analytical techniques.
The current iteration of the PPRVs would benefit from a comprehensive development of SOPs
for tracer studies. The study presented in this paper introduced some recommendations for the
use of PPRVs such as suspending the PPRVs in turbulent flow. However, these suggestions were
anecdotal in nature and were not formally part of the overall study. A formal study for SOP
development would examine the impact of a variety of variables including, but not limited to,
different flow path geometries, proximity to ventilation controls, and underground release
178
|
Virginia Tech
|
locations on the performance of the PPRV. The product of such a study would be to ultimately
provide a user’s manual for the PPRV along with recommendations on the ideal use of the PPRV
based on underground mining conditions.
Given the deployment flexibility of the PPRV, the available release options are indefinite. As a
result, a formal SOP study may benefit from being separated into two parts, a broad spectrum
exploration of release scenarios to determine significant factors along with an in-depth study of
these significant factors. Although numerous experimental designs are available for such a study,
any future experiments should at least include performance as a function of PPRV placement at
different locations in an entry’s cross-section, of PPRV placement in different mining areas such
as various longwall faces, continuous miner sections, smooth entries, rough entries, overcasts,
etc., and of PPRV proximity to abrupt flow path changes such as a turn, a contraction, an
expansion, etc. The interactions of the different placement options should also be included for a
more robust study. SOP development should include a discussion regarding the limitations of the
PPRVs based on the LOD of various analytical techniques, the flow quantity, and the type of
flow. Additionally, a cost-benefit analysis of the PPRV vs. the traditional SF release should be
6
given in order to provide recommendations for when the PPRV may be used in place of SF and
6
when the PPRV may be used in conjunction with SF .
6
One of the limitations of the PPRV is the relatively low release rate when compared to traditional
release systems. In order to provide an adequate PMCH concentration with typical underground
flow quantities, the PPRV introduced in this study should either be deployed in parallel with
other PPRVs or be analyzed using a technique with a sufficient LOD. An increase in PPRV
diameter may, however, be another manner in which the PMCH release can be increased. Future
studies should examine the effect of this design variable as well as its interaction with plug
thickness and temperature on the PPRV’s release rate. Increasing the PPRV diameter would
result in a sympathetic increase in the exposed surface area of the plug. Greater surface area at a
given plug thickness is expected to increase the release rate because of the higher volumetric
flow potential. In conjunction with PPRV diameter, the range of temperatures can be expanded
to include sub-ambient temperatures to interpolate a wider range of operating conditions. Due to
the unknown effect of diameter on the PMCH release rate, close attention should be given to the
179
|
Virginia Tech
|
data produced by PPRVs with a high diameter to plug thickness ratio. As the release rate
increases, a critical point will be reached at which the loss of PMCH through the plug exceeds
the rate at which PMCH vaporizes. If this critical point is reached, a vapor pressure equilibrium
cannot be maintained and would result in an unsteady release rate.
Further development of the PPRV would benefit from an investigation regarding the impact of
different environmental conditions on the integrity of the PPRV. The study presented in this
paper showed that the PPRV’s release rate would remain consistent throughout at least 90% of
the PPRV’s estimated lifetime. This long-term viability can only be maintained if the integrity of
both the silicone plug and aluminum shell is not breached. The environmental conditions in
underground mines can be extremely harsh with dripping water, suspended dust, and nebulized
industrial chemicals. A formal study on the long-term effects of such environmental condition on
the overall integrity of the PPRV would provide further insight into the operating constraints of
the PPRV.
The field evaluation of the PPRV presented in this paper represents the first time that both
PMCH and SF have been simultaneously released in an underground mine. As a result, the
6
application of multiple tracers as a mine ventilation characterization tool has not yet been
extensively explored. The multi-zone analysis ability afforded by multiple tracers has great
potential to enhance the analysis of mine ventilation systems. In addition to the steady state dual
release presented in this paper, numerous other quantitation techniques also exist, such as the
pulse releases and tracer decay release methods. Although these additional quantitation methods
have already been extensively studied in HVAC, further research is needed to translate these
models for use in underground mines.
The detection of PMCH and SF in this study was accomplished by sampling the tracer remotely
6
and analyzing the sample later using a GC. This manner of tracer gas analysis has been used
numerous times in analog studies and has a well-established protocol. However, this method has
two main drawbacks, increased probability of sample contamination and delayed results
production. Although vacutainers have shown high sample stability, contamination can still
occur through a reduction of stopper integrity through exposure to ultraviolet light and
180
|
Virginia Tech
|
Resources 2015, 4 940
Keywords: Mining; Coal Dust; Respirable; Occupational Health; Particulate Composition;
Dust Characterization; SEM-EDX
1. Introduction
A key consideration for responsible development of mineral and energy resources is the well-being
of workers. Respirable dust in mining environments represents a serious concern for occupational health.
Coal mine dust, in particular, has long been linked to various lung diseases like coal workers
pneumoconiosis (CWP) and silicosis [1,2]. Implementation of dust regulations in the US beginning in
the late 1960s has significantly decreased overall incidence of such diseases over the past several
decades [2–4], but analysis of long-term surveillance data appears to show a recent and unexpected
uptick in disease amongst some miners in particular geographic regions like Central Appalachia [3,5–7].
Such trends are alarming considering that most coal mines currently operate below regulatory limits on
respirable dust (i.e., particulates with aerodynamic diameter <10 μm), which generally pertain to total
mass concentration and crystalline silica content. These trends may suggest that other exposure factors,
including specific dust characteristics such as particle composition, size, and shape distributions, may
be important in the occupational health context.
While MSHA’s new dust rule issued in April 2014 targets further reductions in respirable dust
concentrations, it is unclear if or how the lowered limits will affect health outcomes for miners in
locations where causal factors for disease are not well understood. The “new dust rule”' was first
proposed by the US Mine Safety and Health Administration (MSHA) on 19 October 2010 and was
finally issued under the title Lowering Miners’ Exposure to Respirable Coal Mine Dust Including
Continuous Personal Dust Monitors on 23 April 2014 [8]. The rule makes a number of changes to
previous regulations on dust limits and sampling in underground coal mines, and specifically will reduce
the permissible respirable dust concentration from 2.0 to 1.5 mg/m3. It will also require use of continuous
personal dust monitors (CPDMs) by mine operators, and require that citations be issued in any instances
where MSHA-collected samples for single, full shifts exceed the new 1.5 mg/m3 limit.
Indeed, more comprehensive characterization of coal mine dust is necessary to fully explore these
factors. Currently, a standard methodology for comprehensive, particle-level characterization of coal
mine dusts does not exist. This paper describes such a methodology, which uses scanning electron
microscopy equipped with energy dispersive X-ray (SEM-EDX). Although not commonly applied to
respirable mine dust samples, electron microscopy with EDX has proven useful in a variety of
environmental and mineral processing/metallurgical applications for fine particulate analysis. It is
increasingly being used to specifically understand chemistry and morphology of airborne particulates
that represent health hazards—in occupational or ambient environments. For example, methodologies
for analysis of nano-sized particulates in around active welding have recently been described [9,10].
A major objective of the method development included optimization of manual analytical efforts—
i.e., minimizing the required SEM user time for each sample, while maximizing the range of valuable
raw data types to be collected. The developed method includes particle-level analysis of composition,
size and shape, from which mass and volume can also be estimated. Construction of automated
|
Virginia Tech
|
Resources 2015, 4 941
spreadsheet program for computational analysis is also described here, as well as preliminary verification
of the dust characterization method using three samples collected in the field.
2. Description of Developed Dust Characterization Method
The following sections provide a detailed discussion of the particle characteristics that are included
in the developed dust characterization method, as well as a description of procedures used for dust
sample collection and preparation, and selection and analysis of specific particles by SEM-EDX.
Additionally, computation via an automated analysis program is described for easy analysis of raw
data inputs.
2.1. Particle Characteristics of Interest
To fully characterize particles, specific properties are of interest. Particle composition, dimensions,
and shape are values which are determined with SEM-EDX, and volume and mass are calculated as a
result of the analysis. These particle characteristics provide an abundance of data and information
regarding respirable dust samples and aid in the comprehensive analysis of coal mine dust.
2.1.1. Composition
Classification of the dust particles is based on their EDX spectra, which provides a graphical
representation of the elements associated with the particle surface. The spectra are generated by detection
of X-ray emissions from the particle, caused by interaction of the SEM electron beam with its surface;
each element on the particle surface produces a characteristic X-ray when excited by the impinging
electrons. Each peak of a spectrum, thus, represents a specific element, and relationships between peak
heights can provide some indication of the elemental composition (i.e., minerals can be identified by
their atomic stoichiometry). For relatively small particles, such as respirable dust particulates, the
electrons may penetrate deep enough into the particle (e.g., to a depth of about 1 μm) to provide relatively
good information about its overall composition. However, EDX analysis on small particles is also subject
to interference from the sample background (i.e., if electrons penetrate completely through the particle
or the electron beam is sufficiently close to the particle edge).
For the developed dust characterization method, considerable effort was aimed at establishing a set
of pre-determined compositional categories into which most particles in a coal mine dust sample would
be expected to fit. As a preliminary effort, lab-generated dust samples were collected using run-of-mine
(ROM) coal, consisting of coal and rock (i.e., primarily shale and sandstone) taken from an underground
coal mine in Central Appalachia. The mine is considered “low seam” based on its average coal seam
thickness of 24 inches. With an average extraction height of 40 inches, the operation is, thus, cutting
about 16 inches of roof and floor rock during coal extraction.
Dust was generated under a fume hood by pulverizing a sample split from the ROM multiple times.
For each dust sample collected, a pump was operated at a flow rate of 5 L/min to collect dust onto a
37 mm diameter polycarbonate (PC) filter (0.4 μm pore size), which was positioned near the top of the
fume hood, just below the suction fan; this arrangement was deemed appropriate to collect relatively fine
dust over short time periods (i.e., 5–10 min) without the use of a cyclone or other size classifier. A cyclone
|
Virginia Tech
|
Resources 2015, 4 942
was not used to collect the laboratory samples, since the primary objectives were simply determination
of the ROM mineralogy (i.e., such that appropriate particle composition categories could be identified),
and development of standard procedures to be used during the SEM-EDX analysis. For more in-depth
investigation of mineralogy, dust samples were also generated by pulverizing approximately pure rock
and pure coal sub-samples hand-picked from the ROM.
An FEI Quanta 600 FEG environmental scanning electron microscope (ESEM) (FEI Company:
Hillsboro, OR, USA) equipped with a Bruker Quantax 400 EDX spectroscope (Bruker Corporation:
Ewing, NJ, USA) was used. In conjunction with the SEM-EDX hardware, the FEI image analysis
software and Esprit EDX software provided imaging and graphical spectra results. The ESEM was
operated under high vacuum at 15 kV with an ideal resolution and a working distance of approximately
12–13 mm, which was observed to be optimal for this particular scope and application. To prepare
collected dust samples for SEM-EDX analysis, filters were removed with clean tweezers, and on a clean,
hard surface, a 9 mm diameter trephine (i.e., a cylindrical blade) and a clean razorblade were used to
extract the center of the filter. The center sub-section was then attached to an SEM pin-stub mount with
double-sided copper tape and sputter coated with gold/palladium (Au/Pd) to generate a thickness of
about 10–20 nm (i.e., 60 s sputtering time) and create the conductive surface layer needed for electron
microscopy analysis.
Based on detailed analysis of the lab-generated dust samples using a number of EDX parameters, it
was determined that twelve elemental peaks should be included in the developed coal mine dust
characterization methodology: carbon, oxygen, sodium, magnesium, aluminum, silicon, sulfur, potassium,
calcium, titanium, iron, and copper. Further, it was determined that most particles could be classified
into six defined categories based on the peak height ratios: “carbonaceous”, “mixed carbonaceous”,
“alumino-silicate”, “quartz”, “carbonate”, and “heavy mineral”. Although the ROM dust samples did
not contain significant carbonate particles, carbonate particles are expected to be collected in field
samples due to “rock dusting” programs in underground coal mines (i.e., applying pulverized inert
minerals, such as limestone or dolomite, to coal and rock surfaces underground in order to reduce
explosion propagation), For the relatively few particles that could not be classified into one of these six
categories, a seventh category “other” was created.
Table 1 provides examples of typical minerals associated with coal mine dust that fall into each of
these categories, and defines the rules developed for compositional classification. These rules are
fundamentally based on atomic abundance (i.e., atomic percentage equivalencies of primary minerals in
each category), which are correlated to the real-time observed peak height ratios (i.e., Cps/eV) on EDX
spectra of specific elements for each category. For the purpose of expedient decision making during
SEM-EDX use, the observed peak heights are the main parameters used for characterization.
Each of the six defined categories has one or more dominant elements (DEs), which are associated
with the mineral(s) represented that category. For a particle to be classified into a given category, the
observed DE spectral peak heights must exceed the minimums shown in Table 1. It should be noted that
the atomic percentage equivalents shown in Table 1 are operationally defined (i.e., based significant
experience of the authors and preliminary analysis of many known particle compositions), and are not
representative of stoichiometry expected in the mineral(s) in each category. This is because significant
interference from the filter background cannot be avoided for most particles in the respirable size range.
|
Virginia Tech
|
Resources 2015, 4 943
Table 1. Description of dust categories for particle classification by composition.
Parameters for Real Time Classification
Dust Category Example Mineralogy Classification (Atomic % (Raw Peak Heights
Equivalents) (Cps/eV))
Carbon ≥ 70% Carbon ≥ 80
Carbonaceous Coal
Oxygen ≤ 30% Oxygen ≤ 20
4% > Silicon ≥ 2% 20 > Silicon ≥ 10
Very thin clay minerals,
Mixed 4% > Aluminum ≥ 2% 20 > Aluminum ≥ 10
or clay minerals with
Carbonaceous Carbon > 70% Carbon ≥ 80
some carbon content
Oxygen < 20% Oxygen ≤ 20
Silicon ≥ 4% Silicon ≥ 20
Alumino-silicate Clay minerals, feldspars Aluminum ≥ 3% Aluminum ≥ 20
Oxygen > 20% Oxygen > 20
Silicon ≥ 5% Silicon ≥ 20
Quartz Crystalline silica
Oxygen > 20% Oxygen > 20
Calcium/Magnesium ≥ 5% Calcium/Magnesium ≥ 20
Carbonate Calcite, dolomite Oxygen > 20% Oxygen > 20
Carbon < 70% Carbon < 80
Iron/Titanium/Aluminum ≥ 5% Iron/Titanium/Aluminum ≥ 20
Heavy Mineral Pyrite, titanium oxides
Oxygen > 20% Oxygen > 20
Other Diesel particulates, etc. Does not fit any of the above Does not fit any of the above
Note: DE-Dominant element(s) are italicized for each defined category. For particles <1.5 μm in long
dimension, DE content can be up to 50% less than the values noted in Table 1 for all defined groups with the
exception of “carbonaceous”. It was found that the filter media increasingly influences the spectra of smaller
particles, with carbon content increasing and DE content decreasing (see below for details).
Indeed, it is well established that particles in this range often produce spectra that are influenced by
electron penetration depth and/or electron scattering [11]. Electron penetration depth is generally defined
as the depth at which the electron beam can penetrate the sample material. Thus, particles that are very
small or thin may produce X-ray spectra that are greatly affected by filter background, and since the
developed methodology for dust characterization utilizes PC filters, small particles or those with
significant penetration depth are generally observed to exhibit apparently high carbon and oxygen
abundances. For example, although crystalline silica particles (SiO ) should exhibit a silicon and oxygen
2
atomic percentages of roughly 47% to 53% based on stoichiometry, a respirable-sized particle on a PC
background may show silicon and oxygen at 10% and 30%—with the balance being attributed carbon.
Given the particle sizes in question, it is unlikely that poor liberation between materials (e.g., quartz
particles ingrained in carbonaceous dust) is playing a significant role.
To further illustrate, Figure 1 shows the spectrum for a PC filter, which has atomic percentage
equivalencies of approximately 85% carbon and 15% oxygen, and Figure 2 shows spectra and actual
SEM images of typical “carbonaceous” and “alumino-silicate” particles. The spectra of alumino-silicates
should not inherently show high abundances of carbon, but the carbon peak is observed to be very high
as an artifact of PC filter interference. The phenomenon of increasing carbon content with decreasing
particle size is applicable for all defined dust categories.
|
Virginia Tech
|
Resources 2015, 4 944
Figure 1. Example spectrum of the PC filter media. The red peak on the left side of the
spectrum is the peak associated with carbon, and the peak to the right of carbon is the oxygen
peak. The small peaks between 2 and 3 keV are the peaks from the Au/Pd sputter coating,
which should be present in all spectra when Au/Pd is used to coat the samples.
(a) (b)
Figure 2. Comparison of example spectra and images for carbonaceous (a) and
alumino-silicate (b) particles at 12,500× magnification. The spectrum for the carbonaceous
particle (L = 9.87 μm) has a relatively large carbon peak and a much smaller oxygen peak,
while the spectrum for the alumino-silicate particle (L = 11.24 μm) has relatively large
oxygen and carbon peaks, and aluminum and silicon peaks of similar height.
To understand more about the particle size at which electron penetration depth may result in
apparently enhanced carbon peaks, an experiment was conducted that examined quartz particles of
decreasing size. To investigate, a ROM lab-generated dust sample was collected onto a PC filter. Under
the SEM, the filter was scanned for quartz particles of varying sizes. Particles with a long dimension (L)
|
Virginia Tech
|
Resources 2015, 4 945
of roughly 0.7 μm, 1 μm, 1.5 μm, 2 μm, 2.5 μm, and 3.5 μm were found, and EDX spectra were observed
for each. (L is simply the longest dimension visible for the particle, see e.g., [12]). Results showed that
carbon peaks were higher for smaller particles; specifically, particles with L ≥ 1.5 μm had carbon peaks
<80 Cps/eV and silicon peaks >50 Cps/eV, while particles with L < 1.5 μm had carbon peaks
approximately 80 Cps/eV and silicon peaks approximately 20 Cps/eV. Thus, particles with L much less
than 1.5 μm may have exceedingly small DE peaks; and, the rules for classifying such particles into each
compositional category should make allowances for their larger carbon peak due to the probability of
electron penetration and/or scatter.
Understanding the carbon content of particles in the “mixed carbonaceous” category is particularly
challenging. While their EDX results indicate that these particles have both alumino-silicate and
carbonaceous characters, their identity and origin are not definitely known. Several possibilities exist.
Most likely, particles classified as “mixed carbonaceous” are actually very thin and platy alumino-silicate
particles, which are influenced ever more than other alumino-silicates by electron penetration. This
prospect is supported by other recently published work by the authors [13]. Another possibility is that
mixed carbonaceous particles may actually be alumino-silicates that are coated with ultrafine coal dust.
Finally, it cannot be ruled out that this category could include clay mineral particles with some biogenic
component, which seems possible considering the diagenesis of coal and surrounding sedimentary rock
formations such as black shales.
To determine the minimum carbon content that permits classification into the mixed carbonaceous
category, an experiment was conducted that looked at dust particles on a copper background media; the
copper tape ensured that any electron penetration would not result in an enhanced carbon peak, but rather
in copper peaks. This experiment was aimed at determining if EDX spectra from “mixed carbonaceous”
particles actually exhibited high carbon peaks due to their composition, or if such peaks are simply an
artifact of significant electron penetration. An ROM dust sample was collected on a PC filter, and then
some of the dust particles were transferred onto copper tape and prepared for SEM analysis by the usual
sputter coating routine. Particles with L > 5 μm whose EDX spectra exhibited relatively high aluminum
and silicon peaks were specifically studied. Upon analysis of 30 such particles, only four spectra were
found to have carbon peaks >80 Cps/eV. These results indicate that, in most cases, the high carbon
content in “mixed carbonaceous” particles is related to interference from the PC background.
2.1.2. Dimensions
The long (L) and intermediate (I) dimensions of any particle analyzed can be determined directly
from the SEM images using standard “line measurement” tools included in the SEM imaging software.
I is the longest dimension perpendicular to L, which was defined above, in the same plane [12].
Following direct measurement of L and I (in μm), the short or third-dimension (S) can be estimated.
Theoretically, S is the length dimension of a particle measured at a right angle to the plane in which L
and I have been found; so S essentially describes particle thickness. Since different minerals have
characteristic shapes, a unique ratio between S and I can usually be defined for a given mineral type. The
unitless S:I ratio (R) is similar to the aspect ratio generally used in the field of sedimentology (e.g., see [14]).
Alumino-silicate particles, for example, tend to be relatively flat with relatively small R values, whereas
|
Virginia Tech
|
Resources 2015, 4 946
quartz particles tend to be thicker with higher R values. Thus, based on the compositional classification
of each dust particle by EDX and its measured I value, an S value (in μm) can be estimated by Equation (1):
(cid:1845) = (cid:1844) × (cid:1835) (1)
For the particle characterization methodology developed here, the R values assigned to each of the
six defined compositional categories of interest are as follows: 0.6 for carbonaceous, 0.5 for mixed
carbonaceous, 0.4 for alumino-silicate, 0.7 for quartz, 0.7 for carbonate, and 0.7 for heavy minerals.
These constants are based on those commonly used in the field of sedimentology and extensive
experience of the authors in electron microscopy analysis of mineral particulates. The mixed carbonaceous
category R value is an average of the carbonaceous and alumino-silicate values since the identity of these
particles is not definitively known. Dust characterized as “other” cannot be assigned an R value.
2.1.3. Shape, Volume, and Mass
A variety of shape factors can also be computed for particles, including a measure of maximum
projection sphericity (Ψ ), and the cross-sectional (d ) and spherical (d ) diameters. The Ψ value can be
p c s p
determined from the L, I and S dimensions using Equation (2), which was derived by Sneed and Folk
(1958). Ψ is a dimensionless quantity and values range between 0 and 1; values that approach 1 are
p
associated with particle shapes that are increasingly spherical (i.e., L, I, and S are very similar), whereas
values that approach zero are associated with particle shapes that exhibit relatively small S dimensions
as compared to L and I [12]. The d and d values (in μm) can be computed from Equations (3) and (4),
c s
respectively. The cross-sectional diameter is the only calculated value based entirely on measured
properties of particle size and is only accurate if the particle is a perfect sphere. The spherical diameter
is more commonly used and is considered a better approximation of the particle size in aerodynamic
applications [15]. Further, the spherical volume (V) can also be computed (in μm3) from Equation (5).
By assigning approximate density values (ρ) to each compositional category, the particle masses (m) can
additionally be estimated (in μg) using Equation (6). Based on average densities for the primary minerals
expected in each of the six defined compositional categories (i.e., see [16]), the following ρ values (in
g/cm3) have been assigned: 1.4 for carbonaceous, 2.0 for mixed carbonaceous, 2.5 for alumino-silicate, 2.6
for quartz, 2.7 for carbonate, and 4.0 for heavy minerals. The mixed carbonaceous class density is an
average of the carbonaceous and alumino-silicate class densities.
(cid:1845)(cid:2870) (cid:2869)/(cid:2871)
(cid:2006) = (cid:4678) (cid:4679) (2)
(cid:3043) (cid:1838) × (cid:1835)
(cid:3013)×(cid:3010)
(cid:1856) = (3)
(cid:3030)
(cid:2870)
(cid:1856) = (cid:2006) × (cid:1838) (4)
(cid:3046) (cid:3043)
4 (cid:1856) (cid:2871)
(cid:1848) = × (cid:2024) × (cid:3436) (cid:3046) (cid:3440) (5)
3 2
(cid:1865) = (cid:1848) × (cid:2025) × 10(cid:2879)(cid:2874) (6)
In addition to the shape factors noted above, particle angularity might also be considered. Angularity
is an effective measure of the sharpness of the edges of a particle and, in the context of coal mine dusts,
|
Virginia Tech
|
Resources 2015, 4 947
may be important in controlling interactions between respired particles and lung tissue. Angularity can
be rigorously determined by measuring the observed angles of particles on SEM images; however,
particularly for small particles (i.e., with L ≤ 5 μm), such analysis would require significant time. Given
that a stated goal of the dust characterization method developed here was to efficiently collect data, it
was, therefore, decided that a qualitative evaluation of angularity should be employed; practically, this
allows for collection of some potentially valuable information without requiring excessive analytical
time. This type of classification of angularity has historically been applied to particles in the micrometer
size range [17,18]. To qualitatively describe angularity, particles selected for characterization should be
classified as rounded (r), transitional (t), or angular (a) by the SEM user, see Figure 3 [19].
Figure 3. Angularity classification categories based on the qualitative analysis of the
sharpness of particle edges.
2.2. General Procedures for Dust Characterization
In order to successfully analyze samples in a methodical manner, the collection, filter preparation,
and analytical process should be sound. The following steps are outlined to provide the user with a
detailed protocol to efficiently and effectively characterize respirable dust samples.
2.2.1. Sample Collection and Filter Preparation
For collection of respirable dust samples in the field for SEM-EDX analysis, an appropriate pump
deemed permissible for use in underground coal mines must be used; at present, the MSA Escort ELF
pump is almost exclusively used for such applications because it has the capability to maintain near
constant flow rate under a variety of environmental conditions [20]. To ensure collection of only
respirable dust particles and, thus, rejection of particles above the respirable range, the pump should be
operated with a cyclone at a flow rate between about 1.7–2.2 L/min [21], such that the cyclone median
cut point is 4 μm according to the NIOSH 0600 method of sampling [22]. While compliance dust samples
used for determining respirable mass concentration are generally collected on pre-weighed PVC filters,
samples to be analyzed by SEM-EDX should be collected on PC, because they provide a suitable
substrate (i.e., background media) for electron microscopy [11,23,24]. Filter cassettes should be unassembled
two or three-piece types, such that the filters can be easily removed from the cassette for analysis.
In preparing the dust samples for SEM-EDX analysis, filter cassettes are carefully unassembled and
the filters are removed with clean tweezers. On a clean surface, a 9 mm diameter trephine and a clean
razorblade are used to extract the center of the filter. The sub-section removed for analysis represents
approximately 6% of the 37 mm filter. It is recognized that particle uniformity as a function of particle
size may be variable for these types of filters, which can result in larger particles depositing toward the
|
Virginia Tech
|
Resources 2015, 4 948
center [25]; yet deposition is fairly radially symmetric [26]. Center filter analysis has been shown to
provide reasonably precise results for field samples using two or three-piece cassettes [26]. As the main
objective here is to provide relative comparisons between center filter sub-sections, some work has been
completed to demonstrate that particles >0.5μm are uniformly distributed by number across the sub-section.
This filter sub-section is then attached to an SEM pin-stub mount with double-sided tape (e.g., copper,
carbon), and sputter coated with gold/palladium (Au/Pd) to create the conductive surface layer needed
for electron microscopy analysis. It should be noted that carbon sputter coating cannot be used since this
will interfere with composition analysis by EDX of the dust particles containing carbon, but other sputter
coatings (e.g., platinum, Pt) might be considered. During development of the characterization method,
it was observed that a coating thickness of about 10–20 nm (i.e., 60 s sputtering time) was optimal for
preventing sample charging while allowing sufficient electron interaction with the dust particles to
provide high-resolution SEM images and EDX spectra.
2.2.2. Particle Selection and Analysis by SEM-EDX
Following dust sample collection and filter preparation, SEM-EDX is used for particle characterization.
Although equivalent equipment could be used, for the method outlined in this paper the same equipment
and software, described above, was utilized. The developed method utilizes images obtained from a
secondary electron (SE) detector for physical characterization of the dust particles (i.e., to measure
dimensions and qualitatively evaluate particle angularity), and EDX spectra for compositional analysis.
In order to select particles for characterization without bias, a rigorous routine was developed to
navigate the prepared 9 mm diameter filter sub-sections under the SEM. The routine was developed
using an iterative process, whereby over 700 particles in total from the lab-generated dust samples were
interrogated for elemental composition, long and intermediate dimensions and estimated shape factors
(all described in detail below). With each iteration of analysis, the routine was improved until nearly all
particles encountered could be quickly classified into one of the pre-determined compositional categories
described above using the EDX spectra, and raw size and shape data could be efficiently gathered for
later computational analysis. It is important to note that this routine was developed based on the
assumption that somewhere between 50 and 150 particles would be analyzed per dust sample, with fewer
particles limiting the statistical power of results and more particles limiting practicality due to time
requirements. During preliminary verification of the dust characterization method, a simple evaluation
of the effect of number of particles analyzed (i.e., statistical sample size) on resulting compositional
distribution was conducted (see below). Ultimately, it was determined that analyzing 100 particles per
sample provided enough information about the sample while maintaining reasonable analytical time
requirements (i.e., about 75–90 min per sample). A detailed description of the particle selection and
analysis routine follows.
First, the SEM should be focused at a magnification of 10,000×, which will allow for analysis of
particles within the desired size range (i.e., about 0.5–8 µm); a somewhat higher magnification could be
used if the particle size distribution is relatively small (i.e., there are few large particles), but significantly
lower magnification will prohibit adequate resolution for analysis of finer particles. With the line
measurement tool, two horizontal lines are then drawn 2 μm apart and spanning the entire width of the
screen, such that the space between the lines is centered on the screen (Figure 4). The SEM is then
|
Virginia Tech
|
Resources 2015, 4 949
positioned such that the dust characterization will begin in the top left-hand portion of the prepared filter
subsection, approximately three screen shifts from its outer edge and approximately 2.25 mm from the
top (i.e., one quarter of the diameter) (Figure 5). Three screen shifts from the edge of the filter prevents
analysis of any particles disturbed during the filter sub-sectioning process. Additionally, the placement
of the SEM stub inside the instrument determines the orientation of the “top” of the stub, based on the
upper border of the screen.
Figure 4. Example of particle selection and screen shifting via the joystick. The image on
the left illustrates analysis of particles intersecting between the two lines in the center of the
screen at 10,000× magnification. The image on the right, at 2500× magnification, shows four
screens, each outlined in a white dotted line, where analysis (at 10,000× magnification) will
take place consecutively.
Figure 5. Illustration of 9 mm diameter filter sub-section and navigation routing for SEM-EDX
analysis. The image on the left is the whole 37 mm diameter filter and the image on the right
depicts the sub-section removed for analysis. The box in the top, left corner of the filter
sub-section illustrates the first frame (i.e., field of view) in which particles should be selected
for characterization; the black arrows in the filter sub-section define the directions for
successive screen shifts between characterization frames. When one horizontal line of
analysis is complete (black arrow directions), the red arrows define shifting back to the left
side of the filter to continue analysis.
Once the instrument is focused and initially positioned, selection and analysis of dust particles can
begin. Moving from left to right on the screen, each particle with L > 0.5 μm that intersects the space
|
Virginia Tech
|
Resources 2015, 4 950
between the two horizontal lines and falls completely within the field of view should be selected for
analysis; if no particles in the field of view fit these criteria, the next field to the right can be examined,
and so on (see below.) Particles with L < 0.5 μm are too small to produce quality spectra results—if
analysis of smaller particles is critical, transmission electron microscopy (TEM) would be better suited
for this application [27]. In an effort to analyze more of the filter area, in regards to high dust density
samples, a maximum of 10 particles (i.e., the first 10 that meet the above criteria moving from left to
right) per field of view should be analyzed. This would allow a minimum of 10 fields of view in order
to characterize 100 particles.
At the approximate center of each particle selected for analysis, the “spot” (or analogous) analysis
function on the SEM software can be used with in conjunction with the EDX software to generate
elemental spectra. Based on the rules outlined in Table 1, the particle can be classified into one of the
seven compositional categories. Additionally, the L and I dimensions of each selected particle should be
measured using the built-in line measurement tools in the SEM software. Finally, angularity should
qualitatively be classified into one of the three categories described above (Figure 3). After recording
raw data (i.e., L, I, angularity, and composition), the user can proceed to the next particle selected for
analysis. Once all eligible particles (i.e., based on the criteria above) in the current field of view have
been analyzed, the user should proceed to the next field of view (i.e., moving to right per Figures 4 and 5)
for selection and analysis of more particles.
The above steps should be followed until analysis reaches the right-hand side of the filter subsection,
approximately three screen shifts from its edge, or until 100 particles have been analyzed, whichever
comes first. If 100 particles have not yet been analyzed, the user should navigate back to the left side of
the filter subsection (see top red arrow in Figure 5), and reposition the sample such that the field of view
is approximately three screen shifts from the outer edge of the filter subsection and approximately
4.5 mm from the top (i.e., half of the diameter). From this position, particles should again be selected
for analysis by scanning from left to right within the current field of view and adhering to the criteria
outlined above; then, analysis should proceed to the next field of view. If the user again reaches the right
side of the filter subsection before 100 particles are analyzed, the SEM can be repositioned back to the
left—this time approximately three screen shifts from the left edge of the filter subsection and
approximately 6.75 mm from the top (i.e., three-fourths of the diameter). Particle selection and analysis
should proceed as before.
2.3. Automated Analysis Program
To automate analysis of the raw data collected from SEM images and EDX spectra, a spreadsheet
program was also developed using Microsoft Excel 2010 (Microsoft, Redmond, WA, USA). For each
dust particle, the user inputs the compositional classification (i.e., per Table 1), measured dimensions
(L and I), and qualitative angularity classification (i.e., r, t, or a), and the program then computes the
following characteristic quantities based on the assigned R and ρ values for each compositional category
and Equations 1-6: short dimension (S), maximum projection sphericity (Ψ ), cross-sectional diameter (d ),
p c
spherical diameter (d), volume (V), and mass (m). Subsequently, distributions of composition, size (i.e., d),
s s
and angularity (either by particle number or mass) can be automatically generated for each dust sample.
While composition and angularity classifications are inherently categorical (i.e., each particle has been
|
Virginia Tech
|
Resources 2015, 4 951
placed into a specific composition or angularity category by the SEM-EDX user), particle size is
continuous (i.e., the computed spherical diameter is numeric quantity.) Thus, to generate distributions
of quantities based on particle dimensions, a number of size categories (or classes) was defined; for this,
a logarithmic base-2 scale was uses, which is a common approach used to classify particles based on
work done by Wentworth (1922) [28]. Here, the automated program considers a total of nine size classes
from >0.125 μm to >16 μm.
The spreadsheet program additionally includes input cells for general sample information (e.g., sample
name or number, description of collection location or conditions, total filter area and filter sub-section
area, total number of particles characterized, total linear length of filter analyzed), and provides basic
output based on that information (e.g., percent of total filter analyzed, approximated particle density on
the sub-section by mass or number). A number of graphical representations of the data results are also
generated for each sample.
3. Preliminary Verification of Developed Characterization Method
In order to provide some preliminary verification of the characterization method developed for coal
mine dust by SEM-EDX, three field samples were collected and analyzed according to the guidelines
outlined above. In particular, the objectives were to: (1) verify that analysis of 50–150 particles per
sample is sufficient to describe the compositional, size, and shape distributions on the filter sub-sections;
and (2) verify that the six defined compositional categories using the lab-generated dust samples from
ROM material, and rules for classification of particles into each category do, indeed, allow characterization
of the majority of particles from real field samples (i.e., do most particles fit into one of these categories,
or are many particles being classified as “other”?)
It should be noted that the question of particle distribution was briefly addressed in Sellaro and Sarver,
2014. In summary, particle quantification was completed on four different areas (at 2500× magnification)
of filter sub-sections from 17 field samples; this involved counting all particles with L dimensions >0.5 μm
in each of the four areas, which were each located in a different quadrants of the filter sub-section.
Particle counts were determined to be similar (i.e., based on a 95% confidence interval) between each of
the four areas for all but two samples. These specific filters had one quantification area with many
agglomerated particles, as opposed to few, separate particles, viewed on the other three quantification
areas. The agglomeration in these samples is thought to be due to humidity throughout the intake airway
of the mine, where both were collected [29].
3.1. Materials
Three dust samples used for method verification were collected from the same underground coal mine
where the ROM sample used for method development originated. An Escort ELF pump with a Dorr-Oliver
cyclone was used to collect the samples onto 37 mm PC filters, and each sample was collected over a
period of about 120 min. The first sample, “Roof Bolter”, was collected from a location adjacent to a
roof bolting machine, and thus was expected to contain relatively high proportions of alumino-silicate,
and possibly quartz particles (vs. other compositions), due to the drilling activity of the machine into
roof material. The second sample, “Belt Drive”, was collected from a location just above a belt drive,
where coal and rock were being transported below on a conveyor belt. The “Belt Drive” sample was
|
Virginia Tech
|
Resources 2015, 4 952
anticipated to include greater proportions of carbonaceous particles, and some carbonate particles were also
expected due to heavy rock dusting in the belt entries. (Rock dusting is a practice used to limit
propagation of coal dust explosions, and requires walls and floors to be covered with fine inert material
such as CaCO ). The third sample, “Intake”, was collected from a location near the working section of
3
the mine in intake air (i.e., fresh air being delivered to the mine by its ventilation system). The “Intake”
sample was expected to have relatively similar proportions of carbonaceous and alumino-silicate
particles, with some carbonate due to rock dusting in the area.
Estimated particle densities on the Roof Bolter, Belt Drive, and Intake filter sub-sections were
16,292 particles/mm2, 12,639 particles/mm2, and 1850 particles/mm2, respectively. These densities were
extrapolated from the average number of particles counted in four different areas on each sub-section;
each area was 10,404 μm2 and located in a different quadrant of the sub-section areas. Figure 6 displays
SEM images for each sample.
Figure 6. SEM images at 2500× magnification for the filter sub-sections from each
verification sample showing relative particle densities. The far left image represents the
“Roof Bolter”, followed by the “Belt Drive” image, and finally the “Intake” image on the right.
3.2. Results and Discussion
To evaluate the effects of number of particles analyzed (n) on dust sample characterization results,
compositional distributions by particle number and mass were compared for a range of n values (Table 2).
For the Roof Bolter and Belt Drive samples, 200 particles in total were analyzed, and the resultant
compositional distributions were compared for the first 25, 50, 100, 150, and 200 particles (i.e., n = 25,
50, 100, 150 or 200); for the Intake sample, only 100 particles were analyzed in total, so n values of 25,
50, and 100 were compared. Somewhat surprisingly, when comparing compositional distribution of
particles by number, all samples showed relatively similar results across all n values—meaning that even
when n was increased 4- or 8-fold, little change was observed in the relative number of particles being
classified into each compositional category.
When comparing compositional distribution by mass, however, only the Belt Drive sample produced
similar results across all n values. For the other two samples, as n increased, the distributions changed
significantly. For example, in the Roof Bolter sample, the first 100 particles analyzed showed very little
carbonaceous material on a mass basis, but first 150 particles analyzed showed that over a quarter of the
mass was due to carbonaceous particles. This particular discrepancy was traced to a single very large
|
Virginia Tech
|
Resources 2015, 4 954
magnification to assess density of large particles, or elemental mapping at a relatively lower magnification
to assess compositional differences in larger particles).
Additionally, from Table 2, it appears that the six pre-determined compositional categories, and rules
outlined in Table 1 for particle classification, can account for most respirable particles in expected in
dust samples from underground coal mines. Indeed, of the 500 total particles analyzed across all three
samples, none required classification into the “other” category. It should of course be noted that dust
composition could vary with varying coal and rock geologies, and mining and operational practices—
and, thus, between mines. So, further verification of the developed method for dust characterization
should certainly be conducted using samples collected from multiple mines/regions of interest.
To demonstrate the robustness of the developed dust characterization method, size, and compositional
distributions (again by particle number and mass) for the three verification samples were generated by
the automated spreadsheet program. Figure 7 shows the results for the sample collected adjacent to a
roof bolter. With respect to composition, the sample largely consists of alumino-silicates, with
significant coal and mixed carbonaceous particles too. These results are consistent with expectations
based on the sampling location (i.e., the bolter was drilling into the roof, but the air being moved through
the mine also contains coal particles). These results additionally underscore the influence that large
particles can have on mass-based data. Figure 7 indicates that 1% of the particles in this sample, which
all happened to be carbonaceous, fell into the 4–8 μm size class—but these make up 19% of the total
mass. Figure 8 shows the relative angularity of particles in the “Intake” sample. This data indicates that
alumino-silicates and mixed carbonaceous particles tend to be primarily angular, while carbonaceous
particles can be more rounded.
(a) (b)
Figure 7. Particle size distribution by number (a) and by mass (b) for the Roof
Bolter sample; the relative number of particles in each compositional category is shown
within each bar.
|
Virginia Tech
|
Resources 2015, 4 955
Figure 8. Particle compositional distribution by number for the Intake sample; the relative
number of particles classified as having angular, transitional, or rounded shapes is shown
within each bar.
4. Conclusions
SEM-EDX is a powerful tool, which can be used for particle-level analysis of dust samples. This
paper describes a standard methodology developed for the purpose of achieving more comprehensive
characterization of respirable dusts in underground coal mines. Due to the large amounts of data that can
be generated by this method, a relatively simple spreadsheet program is recommended for automating
computational analyses to compare particles within and between dust samples. The recent availability
of automated particle analysis instrumentation to existing scanning electron microscopes could also
provide an even more robust analysis capability by increase the number of particles analyzed by at least
five to ten fold.
Future work should be geared toward further understanding particle uniformity, by both number and
size of particles, across the entire filter area and uniformity by particle size across the filter sub-section.
In cases of non-uniformity, such as agglomerated dust, characterization of >100 particles may be
necessary. The method is also user specific, and the steps outlined above are at the interpretation of the
user, such as in cases of exceptionally high dust density samples and increased numbers of large dust
particles. Although the method outlined in this paper was shown to classify particles properly from one
specific mine, to accommodate a mine of different mineralogy, the particle dust categories should be
altered prior to particle classification. The time required for this type of comprehensive analysis can be
a major drawback; however, the use of a standard methodology may increase analytical efficiency, as
well as consistency.
Acknowledgments
The authors would like to thank the Department of Mining and Minerals Engineering of Virginia
Tech and VT NCFL for assistance with this project.
|
Virginia Tech
|
15th North American Mine Ventilation Symposium, 2015 — Sarver, E., Schafrik, S., Jong, E., Luxbacher, K.
© 2015, Virginia Tech Department of Mining and Minerals Engineering
Considerations for an Automated SEM-EDX Routine for
Characterizing Respirable Coal Mine Dust
Victoria Anne Johanna, Emily Allyn Sarvera
aVirginia Tech, Blacksburg, Virginia, USA
Respirable dust in coal mining environments has long been a concern for occupational health. Over the past
several decades, much effort has been devoted to reducing dust exposures in these environments, and rates of coal
workers’ pneumoconiosis (CWP) have dropped significantly. However, in some regions, including parts of Central
Appalachia it appears that incidence of CWP has recently been on the rise. This trend is yet unexplained, but a
possible factor might be changes in specific dust characteristics, such as particle composition, size or shape.
Prior work in our research group has developed a standardized methodology for analyzing coal mine dust
particles on polycarbonate filter media using scanning electron microscopy with energy dispersive x-ray (SEM-
EDX). While the method allows individual particles to be characterized, it is very time-intensive because the
instrument user must interrogate each particle manually; this limits the number of particles that can practically be
characterized per sample. Moreover, results may be somewhat user-dependent since classification of particle
composition involves some interpretation of EDX spectra.
To overcome these problems, we aim to automate the current SEM-EDX method. The ability to analyze more
particles without user bias should increase reproducibility of results as well as statistical confidence (i.e., in
applying characteristics of the analyzed particles to the entire dust sample.) Some challenges do exist in creating an
automated routine, which are primarily related to ensuring that the available software is programmed to
differentiate individual particles from anomalies on the sample filter media, select and measure an appropriate
number of particles across a sufficient surface area of the filter, and classify particle compositions similarly to a
trained SEM-EDX user following a manual method. This paper discusses the benefits and challenges of an
automated routine for coal mine dust characterization, and progress to date toward this effort.
Keywords: Coal workers’ pneumoconiosis, Respirable dust, particle analysis, scanning electron microscopy,
Automated SEM
1. Introduction Automated SEM-EDX analysis has historically been
used for applications such as industrial process control
Coal mining operations generate dust which can be
and forensics [4]. However, SEM automated analysis
respired into the lungs of workers to cause occupational
hardware and software advancements have made it
health diseases such as coal workers’ pneumoconiosis
applicable to a variety of other applications, including
(CWP). The mining industry saw dramatic reductions in
mineral samples [4]. Automated SEM is able to analyze
CWP cases as dust standards and ventilation regulations
features such as inclusions in metals; porosity of
in underground coal have improved over the past few
geological samples, and samples containing wear debris
decades under the Federal Coal Mine Health and Safety
from combustion engines [4]. Another application for
Act of 1969 [1]. The Act also established the Coal
mineral samples is the detection of an anomalous
Workers’ Health Surveillance Program through which
particle within a grouping of thousands of particles of
NIOSH has witnessed first-hand the increase in CWP
other compositions [4]. This application might be
rates in the eastern United States, particularly Central
particularly useful to the occupational health field for the
Appalachia, since the mid-1990s [1-3]. This is of
analysis of dust samples containing atypical or
particular concern because the majority of cases have
hazardous particles.
been reported in young coal miners and many of the
cases are advanced [1-2]. Further research should be Some work has been conducted in the realm of
aimed toward determining the cause of increased automated dust particle analysis. Deboudt et al. [5]
incidence of CWP in Central Appalachia in order to performed automated SEM-EDX particle analysis on dust
improve miner health and safety [3]. Little is definitively samples collected on the Atlantic coast of Africa. As in
known regarding the effects of specific dust our project, this group collected airborne particulate
characteristics (such as size, shape, and chemical samples on polycarbonate filters and ran an SEM at an
composition) on lung disease occurrences in accelerating voltage of 15kV. Using the Link ISIS Series
underground miners. Analyzing these dust particle 300 Microanalysis system developed by Oxford
characteristics using scanning electron microscopy Instruments, this group was able to collect spectral data
(SEM) may be a good place to start. Automated SEM for individual particles with a 20 second acquisition time
analysis could be particularly advantageous in collecting [5]. Even faster rates of data collection can be achieved
data from more dust samples at a faster rate. though. Ritchie and Filip [6] recently undertook an effort
_______________________________________________________________________________
1
author’s email: [email protected]
|
Virginia Tech
|
to optimize the speed of automated particle analysis by range of 0.5-8.0μm in diameter. Two horizontal lines are
SEM-EDX and demonstrated data collection at drawn 2μm apart, centered on the screen and spanning
approximately three particles per second. They the width of the screen, using a line measurement tool,
employed a structured query language database that as depicted in Figure 1.
stores millions of particle records and is able to
simultaneously classify multiple particles to multiple
categories [6]. The authors’ research group does not
necessarily need to be analyzing particles at that rate,
though aiming to speed up analysis time by a few
seconds and perhaps creating a comprehensive particle
database could be beneficial toward research efforts.
Other researchers have also worked on multi-frame
particle analysis using the SEM. Fritz, Camus, and 2 μm
Rohde [7] have done work with automated microscope
stage analysis that can cover hundreds of frames in one
run to ensure total sample coverage. For applications in
the mining industry, this ability is particularly attractive
for particle sizing and the analysis of respirable dust
particles. Collecting data for particles over the entire
sample can be necessary for statistical significance [7].
The authors’ research group also plans to employ
automated multi-frame analysis.
Fig. 1. Example of the horizontal lines drawn 2 μm apart for
particle selection. Only particles touching this region are to be
selected for analysis [8].
2. Previously Developed Standard Dust
Characterization Method
Prior work in the authors’ research group has The stage is moved so that the first field to be
developed a standardized methodology for analyzing analyzed is three screen shifts from the outer, left edge
coal mine dust particles on a polycarbonate filter [8-10]. of the filter, 2.25mm (one quarter of the filter diameter)
This method was developed using an FEI Quanta 600 down from the top of the filter. Moving from left to right
FEG environmental scanning electron microscope and top to bottom each particle with a long dimension
(ESEM) (FEI, Hillsboro, OR) equipped with a Bruker greater than 0.5μm intersecting the space between the
Quantax 400 EDX spectroscope (Bruker, Ewing, NJ). two horizontal lines and falling completely within the
The ESEM is operated under high vacuum conditions at field of view is analyzed. Up to ten particles meeting the
a voltage of 15kV with a spot size of 5.0μm and at the specifications are characterized per field in order to
optimal working distance of 12-13mm. Bruker Esprit ensure that at least ten fields are analyzed and increase
software is used to collect spectra results for the the representativeness of the results. Figure 2 shows a
classification of individual particles. The “spot” analysis backscatter detector image of a typical field of ten
function of the ESEM software is used in conjunction particles that would be analyzed.
with the EDX software to generate elemental spectra.
Six compositional classification schemes were
developed for coal mine dust particles based on peak
elemental spectra heights of aluminum, calcium, carbon,
copper, iron, magnesium, oxygen, potassium, silicon,
sodium, sulfur, and titanium. The six classifications are
“alumino-silicate,” “carbonaceous,” “carbonate,” “heavy
mineral,” “mixed carbonaceous,” and “quartz.” Any
particles that do not fit into these categories are termed
“other.” Data on particle size and shape is also collected
in this dust characterization method. The long and
intermediate dimensions of each particle are measured
using the line measurement tool provided in the ESEM
imaging software. The shape is qualitatively classified
based on user interpretation as either “angular,”
“rounded,” or “transitional.”
The sample analysis routine begins by focusing the
SEM at 10,000x magnification to provide optimal Fig. 2. Example of a typical field of particles to be analyzed
resolution for analyzing particles in the desired size (at 10,000x magnification).
2
|
Virginia Tech
|
Once the first field has been analyzed, the next field backscatter detector image of a typical field of particles
of view to the right is analyzed. If fewer than 100 that would be analyzed using the automated routine.
particles have been analyzed upon reaching the edge of
the first row, the stage is shifted so that the field of view
is 4.5mm (one half of the filter diameter) from the top of
the filter and the same procedure for the previous row is
followed. If fewer than 100 particles have been analyzed
upon reaching the edge of the second row, the stage is
shifted once more so that the field of view is 6.75mm
(three quarters of the filter diameter) from the top of the
filter, following the previous procedure. Figure 3 depicts
this analysis routine in terms of filter navigation under
the SEM.
Fig. 4. Example of a typical field of particles (at 1,000x
magnification).
This image comes from the same sample as Figure 2;
however, by being able to analyze particles at 1,000x
versus 10,000x magnification, many more particles can
be analyzed per frame. Once the first frame is selected,
the imaging tool in the Esprit software is used to pull the
image of the frame from the SEM software and import it
for analysis. A special feature in Esprit allows for rules
Fig. 3. Illustration of a 9 mm diameter polycarbonate filter
and filters to be applied to the image so that the software
and navigation routing for SEM-EDX analysis. The box
is programmed to identify dust particles. Here, a binary
represents the first frame in which particles are selected for
image can be created and settings can be adjusted so that
characterization; the black arrows define the directions for
successive screen shifts between characterization frames. the software distinguishes particles as white and the
When one horizontal line of analysis is complete (black filter as black. Figure 5 depicts the binary imaging
arrows), the red arrows define shifting back to the left side of process in Esprit.
the filter to continue analysis on the next horizontal line [8].
The manual method has the capacity to analyze 100
particles on one sample in 75-90 minutes, depending on
user experience and sample characteristics (e.g., particle
density); despite the wealth of information that can be
obtained, the method is clearly too time-consuming to be
practical for a large number of samples.
3. Automation of the Standard Dust
Characterization Method
Considering the need to significantly speed up
particle characterization, efforts to automate the above
routine have recently been initiated. This work is being
developed using the same ESEM-EDX system
previously mentioned, and several special features
Fig. 5. Example of a binary image of a typical particle field (at
available for add-on to Bruker’s Esprit software. A
1,000x magnification). This is the same particle field as in
major benefit to automation is that the software can
Figure 4.
characterize particles at a magnification that is ten times
lower than the magnification required for the standard
dust characterization method. Figure 4 shows a
3
|
Virginia Tech
|
Once all settings have been adjusted for particle weight percentages of elements other than carbon and
identification, particle sizing can be conducted. Figure 6 oxygen. This does not allow them to be classified under
depicts the particle field once all the particles have been the original carbonaceous category that only sets
sized. The software outputs data such as length, width, parameters for carbon and oxygen weight percentages.
and shape factor to characterize the particles by size and Therefore, in this preliminary work, a second
shape, as shown in Figure 7. carbonaceous classification category was created with
additional elemental weight percentage rules (i.e.,
This single frame contains 171 particles that can be
“carbonaceous II” in Figure 9). By doing this,
analyzed, 71 more particles than would have been
carbonaceous particles that were previously not
analyzed in up to ten frames using the standard dust
classified are classified in the second carbonaceous
characterization method. This particle sizing routine
category. A particle chemistry analysis feature in the
minimizes user interpretation of the long and
Esprit software can classify each particle detected in the
intermediate particle dimensions which was required for
frame.
the standard dust characterization method.
Once sizing is completed, the particle classification
scheme can be implemented. The Esprit software allows
for particle classification based on the weight percent of
elements detected in spectral analysis. Therefore, rules
can be set for the maximum or minimum elemental
weight percentages required for various particle
classification categories. We have currently developed
rudimentary rules and particle classification categories to
demonstrate the utility of automated analysis and its
potential for respirable dust particles from coal mining
environments. The preliminary categories are based on
typical elemental weight percentages observed for
particles classified using the manual method. Figure 8
displays chemistry results for some particles identified in
Figure 6, showing the weight percentages of specific
elements considered by the classification rules. Figure 9
shows particle classification results in a bar chart as a
useful visual tool.
Fig. 6. Example of the particle sizing process in Esprit. This is
It should be noted that the chemistry classification the same particle field as in Figure 4. Each particle found that
is accepted for analysis is outlined in blue while undergoing
categories are not currently developed enough to
size classification.
accurately classify every particle (i.e., as it would be
classified manually). This is especially true for
carbonaceous particles because they can contain small
Fig. 7. Example of the particle sizing results for some particles shown in Figures 4-6. The accepted particles are all numbered and
listed in order in the first column. Other particle properties are provided for each particle in subsequent columns. At the bottom of the
results page, minimum, maximum, average, and standard deviation values are provided for each particle property.
4
|
Virginia Tech
|
The automated analysis at the particle density mineral identification and other key elements be used to
demonstrated in Figure 4 takes just over ten seconds per classify these particles in its place. We have determined
particle which allows for total analysis of this frame of that gold and palladium should definitely be
171 particles in approximately 30 minutes. At this rate, deconvoluted before spectral analysis because the
171 dust particles are being analyzed two or three times sample coating is comprised of these elements and can
faster than just 100 dust particles using the standard dust interfere with the chemical identification of the dust
characterization method. It is expected that with several particles.
modifications in basic operating parameters of the SEM-
EDX system, the efficiency can be greatly increased.
The majority of particles in this frame were classified as 5. Conclusion
carbonaceous while many particles were classified as
There is much work still to be done to refine the
alumino-silicates and fewer were classified as heavy
particle classification parameters in order to properly
minerals, carbonates, mixed carbonaceous, and quartz.
automate the particle analysis. Our goal is to determine
Another special feature in the Esprit software allows and implement the correct particle classification
for the automation of the microscope stage movement categories and their appropriate rules for our samples. To
from frame to frame. The user is able to designate the do this, we plan to:
starting frame position and the ending frame position
collect spectral data on many particles from our
along the sample filter. Once the entire area of the filter
samples
to be analyzed is determined, particle sizing and
chemistry can be run frame by frame and the software determine the appropriate elemental weight
will export the data after completion. This tool can even percentage thresholds for the classification
be used to run multiple samples consecutively so that up parameters
to 16 samples could be analyzed in one run. determine which elements should be deconvoluted
modify the currently developed particle classification
categories so that they are able to classify all particles
4. Discussion encountered
ensure that the software is properly classifying all
It seems as though developing an automated particle
particles
analysis routine is a step in the right direction for our
research. The standard dust characterization method is We aim to program the software to classify particles in
too time-intensive for the amount of particles that can the same manner that the user would classify them, but
practically be analyzed per sample because the user must without the human error.
interrogate each particle manually. A major advantage to
automated particle analysis is the amount of time saved
in the lab due to quick, electronic characterization of Acknowledgments
particles. This is also applied to data entry where
The authors acknowledge the Alpha Foundation for
automated analysis automatically exports data into
the Improvement of Mine Safety and Health for funding
Microsoft Excel while data obtained from the standard
this work. We would also like to thank Steve McCartney
method must be manually entered. Moreover, results
of Virginia Tech ICTAS-NCFL for operational
from the standard dust characterization method may be
assistance with SEM-EDX analysis, Ted Juzwak of
somewhat user-dependent since classification of particle
Bruker Corporation for assistance in learning the Esprit
composition and shape currently involves some
software capabilities, and Patrick Wynne for countless
interpretation of EDX spectra. Other benefits of an
hours of SEM work. We are also grateful to Meredith
automated particle analysis routine are a significant
Scaggs for her efforts to collect and prepare dust
increase in the number of particles than can be analyzed
samples.
per sample and minimization of user interpretation to
acquire results. The ability to analyze more particles
without user bias should increase reproducibility of
References
results as well as statistical confidence in obtaining
results from a representative portion of the sample.
However, some challenges do exist in creating an [1] Centers for Disease Control (CDC). Advanced
automated routine, including training the available Cases of Coal Workers’ Pneumoconiosis-Two
software to appropriately make multiple decisions such Counties, Virginia. MMWR, 55.33 (2006) 909-
as those involving differentiation of individual particles 913.
from anomalies on the filter media, selection of particles
[2] D. Blackley, C. Halldin, and A.S. Laney,
for analysis, and classification of particle composition.
Resurgence of a Debilitating and Entirely
Another challenge arises due to the filter media being
Preventable Respiratory Disease among Working
comprised of carbon and carbon being a key element in
Coal Miners, American Journal of Respiratory
the classification of carbonaceous, mixed carbonaceous,
and Critical Care Medicine 190.6 (2014) 708-709.
and carbonate minerals. We are working to determine
whether or not carbon should be deconvoluted for proper
|
Virginia Tech
|
Design of a Mine Roof Strata Analysis Device
Andrew James Reksten Russell
ABSTRACT
Because the roof lithology in an underground coal mine is typically variable and poorly
known, the safety and efficiency of these mines is reduced. To address this shortcoming, a device
for analyzing rock properties by way of scratching a mine roof borehole was designed and tested
in multiple different media with the goal of determining in situ mine roof properties with a
nondestructive technique. Tools were developed for measuring extraction force and position of
the scratching mechanism and those values were compared versus time for multiple tests to look
for changes in applied force over changing positions. Because of signal stability and
inconsistencies in scratch depths the data were found to contain too much variation to determine
any rock properties or changing rock conditions from the simulated roof material in the concrete
block. However, further scratch tests in a sandstone block indicated that increasing the diameter
of the wire scratchers (and therefore increasing their stiffness and accompanying normal force)
from 0.045 inches to 0.055 inches increased the average pull force from 6.24 to 9.96 lbs. Similar
to that test, a scratch test was performed in a PVC pipe where it was found that increasing the
scratcher diameter from 0.045 inches to 0.051 inches increased the pull force from a 2.81 lb
average to a 36.46 lb average, with considerably better gouging of the host material.
|
Virginia Tech
|
ACKNOWLEDGEMENTS
I would like to thank my advisor, Dr. Erik C. Westman, for the incredible opportunity to
be a graduate student in this program and for his perpetual support for me and my work. This
project was funded by the National Institute for Occupational Safety and Health (NIOSH) under
Contract 200-2011-40313 for “New Technologies for Identifying and Understanding Ground
Stability Hazards”. I would like to thank NIOSH for the funding to conduct this research
I thank Dr. Mario Karfakis for his consistent insight into my research and for helping me
develop my ideas and for access to his lab resources. I thank Dr. Kray Luxbacher for taking time
out of her schedule to be on my committee. I also would like to show gratitude to the entire
Mining and Minerals Engineering faculty and staff for their help and guidance.
I thank Ben Fahrman and Brent Slaker for their constant guidance and support. I would
like to thank Yuncong Teng and Ben Owsley for their assistance in the field. I would like to
thank Joseph Amante for his assistance with the lab equipment and Mike Kiser for his general
help. I would like to thank the other members of the ground control research group for their
input: Kyle Brashear, Billy Thomas, Xu Ma, Enji Sun and Will Conrad. I also thank every other
graduate student in our program for their help and support. Thanks to Amritpal Gill and Jacques
Delport for their help and resources for the electronic components of my research.
Thank you Jim Waddell for constructing the devices used in my project as well as Robert
Bratton for generously assisting with my research and nurturing my interest in mining and
mining technology. Thank you to J. H. Fletcher & Co. for allowing me to use their facility,
resources and employees to conduct my testing
And lastly, a huge thanks to my loving friends and family for everything they do for me
and for believing in me. Thank you to Scott Sr. and Jacqualine Russell for your love, support and
sacrifice on my behalf. Thank you to my parents, Scott Jr. and Donna Russell for their support
and investment in my success.
iii
|
Virginia Tech
|
Introduction
Mining has served the societies and economies of the world for hundreds of years by
providing jobs, resources and technology that have far reaching benefits. It is unlikely that
mining will ever not have a place in the world considering the growing rate with which humans
require valuable minerals and coal. Mining will continue to pose risks from use of vehicles,
electricity, powered haulage, as well as the hazards of fires, explosions, and gound falls.
Considering the nature of excavating rock masses and subjecting the rock material to physical
and chemical changes, the risk of ground related incidents is likely to exist as long as
underground structures are excavated.
According to the National Institute for Occupational Safety and Health “Mining Topic:
Ground Control Overview”, almost 40% of fatalities that occurred underground coal mining
between 1999 and 2008 and were the result of stability failure in the face, roof or rib (2012).
According to the same source, the fall of rocks between roof supports injures between 400 and
500 miners per year (“Mining Topic: Ground Control Overview”, 2012). It has been argued that
ground control will be more technically challenging as a result of the need to develop mines in
deeper areas or areas with more difficult conditions in order to combat dwindling resources.
The study of ground behavior in mining has brought to light a number of topics that
require further investigation. Some of these topics, such as large scale modeling, pillar bump
analysis, support optimization and subsidence are given considerable attention and research.
There seems, however, to be a lack of investigation in the ability to obtain strength
characteristics of mine roof strata at the face and along the panel. Possible prohibitive factors for
further investigation in this field may be the hurdles of intrinsically safe equipment certification
1
|
Virginia Tech
|
as well as a distrust in the ability to make good measurements in such a heterogeneous material.
This study covers the design and use of a tool that explores the task of analyzing and
categorizing mine roof.
The gathering of information about a rock mass requires the investigator to impart some
type of energy into the rock and then analyze the response of the rock. Several tests exist that do
this in a lab setting, such as Schmidt hammers which use rebound characteristics of a metal rod
on rock masses to determine rock traits. There also exist primary and shear wave tests that use
the transmission of waves through a rock mass to gather information about its properties. Most
notably, destructive test methods such as the Uniaxial Compression and point load tests or the
Brazilian test relay valuable information about rock behavior. Recently, several highly
instrumented and controlled testing methods have been developed that provide details about the
strength of rock masses in the lab setting without the use of destructive testing methods.
It is hopeful that one of these non-destructive, rock-surface implemented tests could
prove to be a means to gather strata properties in a mine setting. The current method of taking a
core sample at a mine and coupling it with laboratory analysis is simply too expensive to be able
to be considered a suitable roof strength analysis method. Primarily, rock core analysis lacks
enough resolution across a property to make widespread entry or panel scale judgments on
ground conditions because so few cores are logged on account of their cost. Core analysis also
lacks the ability to indicate the changes to the immediate roof strata that would be induced by
excavating a room and subjecting the strata to gravity loading. In light of these issues of
resolution and cost prohibitive factors, a cheaper, easier and more widespread testing method is
needed.
2
|
Virginia Tech
|
One of the current methods of analyzing the competence of mine roof mid-shift entails a
ground inspector running the tip of a tape measure (the off-the-shelf variety that can be found at
any hardware store) up the length of the legally required test holes. Test holes in this case refers
to the required empty drilled holes that extend 1 foot past the deepest bolt depth in the roof [30
C.F.R. §75.204 (f)(2)]. As the tape measure tip is run along the length of the hole, the bent metal
tip of the tape is expected to nest itself in any discontinuities that the roof material may have.
Additionally, the inability to extend the tape fully up the hole indicates an obstruction in the roof
hole such as shifted roof layers. This test is described in the interactive training lesson for roof
bolters (MSHA - Interactive Training - JTA Spiders – Roof Bolter Operator). This tape measure
test serves two functions, it indicates the presence of any such discontinuities and bed
separations as well as providing an indication of the distance up the hole that the discontinuities
exist. While being a good inexpensive test method, this fails to properly extract all the useful
information that may be contained in one of these test holes.
Another method of analyzing the composition of the roof is less desirable but more
informative. Ground control experts can learn about the roof by looking in areas where a large
scale collapse has already occurred. If they can inspect the roof cavity where the rock fell from,
they can often see fracture networks that are difficult to interpret by looking only at the skin of
the mine roof. Furthermore, this can be a good way to look at the stratigraphy of the immediate
roof and see if there are any obvious weaknesses in the roof layers. Depending on the conditions
of the roof fall, conclusions about the anchorage characteristics can be drawn. For instance, if the
grouting or anchor shell of the bolt is visible in the fallen roof material then it is likely that there
was some kind of slippage or loss of anchorage integrity. It also allows the damage inspector to
look more closely at the qualities of the strata that the anchorage mechanism is attached to and if
3
|
Virginia Tech
|
that rock mass has the strength characteristics that it was assumed to have when the bolt was first
installed.
In general, the practice of in-cycle roof analysis is often left up to the experience of the
bolters. They are the individuals that have the most information about how the drills behave in
the roof strata and if there are any conclusions that can be drawn from the drill behavior about
possible roof composition and competence. Stewart, et al. outline a roof control program in their
2006 paper that explains a coal mine’s use of “Lith-Graphs”, which are qualitative forms filled
out by bolters as each test hole is drilled. These forms have spaces where information about bit
wear, water presence and possible voids are recorded and then provided to the geologists and
engineers (Stewart, et al., 2006). Foremen and geologists are encouraged to introduce
supplementary support in areas where the roof is troublesome on the basis of information found
in these “Lith-Graphs” as well as from discussion with the bolters and observations of roof
conditions (Stewart, et al., 2006). The inclusion of information from a quantitative roof strength
evaluation test, i.e. the one outlined in this paper, would certainly help compliment and justify
some of the roof control decisions that ground control experts, foremen and mine workers will be
making about the presence and degree of installed auxiliary support.
The goal of this project is that attention can be given to the apparent lack of panel-scale
or entry-scale quantitative ground analysis tools. Prior literature is explored to see what kind of
tests exist that could be readapted to bridge this knowledge rift, or at the very least, inspire
additional investigation into that technique. The literature review should serve as a good
reference for anybody that hopes to find new ways to measure rock properties that go beyond the
traditional destruction methods such as, but not limited to Uniaxial Compressive Strength (UCS),
4
|
Virginia Tech
|
Point Load, and the Brazilian Tensile Test. The creation of a device that evaluates the selected
quantitative analysis method(s) for feasibility in a mining setting is also expected.
This device is to be used by people with little to no assistance from powered machinery
except for simple transportation of the unit and should therefore be portable. The device should
be useful in areas that are out of the mining cycle i.e. not modifying or adding to existing
equipment. Ease of use, portability and accuracy will encourage mines to use such a device
instead of avoiding it when the information gained from it is not worth the loss in productivity.
Above all else, this device should not diminish the stability of the current mine ground or
interfere with ventilation in any way and should put the operator at little risk of injury. Ideally,
this device will inspire others to continue research in the area of underground strata
characterization and hopefully provide enough technical information to advance the ability of
other researchers to develop their own devices or expand upon this one.
The principle aspect of the design is to control the motion of a device that interacts with
the walls of the borehole. The control of the motion allows each hole to be analyzed relative to
itself on account of the similar motion characteristics. Through the interactions with the borehole
wall, changes in forces applied to the device are expected in the presence of differing rock types.
By looking closely at these force changes and where their changes occur in the borehole, an
understanding of the strength and position of constituent rock layers can be gained. The device
created for this project addresses the position and removal force of a scratching mechanism in a
borehole and monitors those parameters in an effort to extract changes in rock type and
properties.
5
|
Virginia Tech
|
Literature Review
Geologic Background
Coal seams are notable for their deposition mechanisms and for the way that these
mechanisms ultimately define the nature of the adjacent strata. The coal deposits of the eastern
United States are the byproduct of swamps left undisturbed for many thousands of years such
that biomass accumulated and eventually began coalification (Molinda, 2003). Molinda
elaborates on the presence of these swamps and how their eventual coal thickness, middling
properties and roof content is subject to the dynamic behavior of ancient river deltas (2003). This
is due to the buildup of sediment within the delta network and how it forces a redirection of the
distributaries of the river as it drains into a larger body. What was once a swamp gets covered by
redirected river water and becomes a new depositional area for sands, silts, clays and other fine
rocks which form the roofs, riders and floors of future coal mines (Molinda, 2003). The variety
of mining conditions (both favorable and unfavorable) that these processes later induce prove to
be the core of the subject of ground control. This variability in ground conditions highlights the
concept that any good information on the geology of a coal seam and adjacent strata will be of
great help to engineers charged with mitigating their hazards.
A proper characterization of strata is widely believed to be one of the most crucial
elements of an effective ground control strategy. Iannacchione and Zelenko speak on the
importance of strata characterization in their 1995 work on coal mine pillar bumps by outlining
the relationship between thick sandstone layers above and below the coal seam and the
corresponding likelihood of violent pillar activity. It is probable that the presence of these
massive layers would be detected through traditional methods of geostatistics namely analysis of
6
|
Virginia Tech
|
Coal Mine Roof Characteristics
Generally, a robust sandstone layer in the roof is considered a favorable geologic feature
as it can provide a good anchorage region for a bolt. This is especially true in shallower mines
where having that massive sandstone layer wouldn’t subject the coal pillars to catastrophic burst
failure from the excessively high geostatic stresses seen in deeper mines. However, there are
accompanying negatives with having sandstone as a roof. Molinda and Mark describe in their
2010 work on ground failures in weak rock that an interface between sandstone and an
underlying shale layer can be riddled with discontinuities and can lead to a frail and unfavorable
roof. Although the strength properties of this interface may be difficult to quantify, the location
of the sandstone-shale transition is crucial knowledge whoever decides the bolt anchorage depth.
The benefits of knowing where sandstone layers are in the roof are obvious and a device that
could locate them would better inform choices about roof control design.
Mark and Molinda continue to describe negatives of sandstone by establishing that
sandstone layers can serve as a vector for groundwater contained within an aquifer (2010). This
sandstone can introduce water into the adjacent shales which are often sensitive to moisture and
lead to a crumbly, problematic roof layer (Molinda and Mark, 2010). It was later suggested in
that same work that the presence of a test hole may help to bleed the sandstone layer of the water
and help slow the time-dependent, moisture-induced degradation of the underlying shaly rock.
Again, the knowledge of the relative position and composition of these layers is important to
correctly mitigating their complications.
The ability to detect the presence of stackrock, thinly interbedded layers of shale and
other friable rock layers, is key to catering a ground control plan to local mine areas. Methods of
controlling these features are outlined in the 2008 Molinda, Mark, Pappas and Klemetti paper on
8
|
Virginia Tech
|
ground control issues in the Illinois basin. Generally, it is suggested that overlaying sandstone or
limestone beds are to be sought in the roof strata around these stacked layers to serve as a strong
anchorage horizon for the bolt (Molinda, Mark, Pappas, and Klemetti, 2008) (Molinda, Mark,
2010). In fact, Molinda, Mark, Pappas and Klemetti go on to suggest that in the presence of a
thick limestone layer, a solid one foot minimum of resin anchorage should be rooted in the
overlying limestone to give good suspension support to the weaker underlying layers (2008).
With changing thicknesses of limestone and stackrock formations, it is clear that the capacity to
make more detailed surveys on the position and dimension of these layers is advantageous. The
benefit that would come from having a tool to travel up a test hole and make analyses about the
roof composition is undeniable, especially considering the great number of test holes that can be
accessed.
The presence of rider seams in coal mines is another noteworthy geologic hazard.
According to Molinda in the 2003 Geologic Hazards and Roof Stability in Coal Mines, rider
seams are thin coal beds (6-48 inches) that overlay a thick, mineable seam. There is often a small
formation of shale between the main coal seam and the rider seam, this interlying shale layer has
a low formation strength of 28-40 for the CMRR index (Molinda, 2003). Rider seam thickness
and position in the roof can be difficult to categorize and there can be a number of them in the
immediate roof layer, further complicating any strategic plan for mapping them. If several
adjacent bolts anchor within a rider seam, failure can occur because this seam loses structural
integrity easier than other, more solid layers (Molinda, 2003). The wide array of dimensions for
rider seams requires a systematic, consistent approach for monitoring. The most effective method
for detecting and categorizing rider seams is regular test holes that cross the rider layers and
allow the ground control expert to make conclusions about their location (Molinda, 2003).
9
|
Virginia Tech
|
The presence of clay in the roof matter is another of the principal hazards that changes
ground integrity. Clays typically form in veins that intersect the seam and can have a dramatic
impact on the solidarity of the roof, so much so, that they were the cause of 90% of ground falls
at some mines in Pennsylvania and Illinois (Molinda, 2003). As is typically the case with these
roofs, the suspension of the weak clay and shale from a sturdy limestone or sandstone beam of
suitable thickness is essential to keeping the entry open (Molinda, 2003). There is a great need to
collect and systematically process the thicknesses of the roof material in these areas, as having
conclusive strata thicknesses and positions are necessary for the proper anchoring of the bolts
and implementation of sufficient supplementary support (Mark, Molinda, and Burke, 2004).
Due to the wide variety of roof conditions outlined in this ground control review, the
need to determine their presence in a mine setting is deemed pressing. The fact that the modes of
sedimentary rock deposition tend to manifest themselves in horizontal formations, is important
because a vertical test hole would likely cross several different rock types between the collar and
its deepest point. This provides a valuable opportunity to use a single analysis technique in a
single hole that would establish interaction with several different rock layers and expose possible
risks to miners that are not immediately visible to them. There are several different analysis
techniques to explore that can indicate strength characteristics of these roof rocks.
10
|
Virginia Tech
|
Rock Analysis Methods
The use of roof drilling parameters to back analyze the characteristics of rock is a
promising technology that has the support from industry experts and academic researchers alike.
This technology, while providing relevant information on roof strata composition, is beyond the
nature of this study because it is purely within the operation cycle and bypasses the desired
portability and free implementation of the roof strata analysis device to be constructed in this
project. Using a roof bolter to drill new holes for collecting strata strength data would be
ineffective at analyzing the roof strength characteristics of mine areas that may not have had any
new roof bolts installed in a few years and a roof drill is unlikely to venture again. A portable
roof analysis device would function well in a place where it would be economically unfeasible to
bring a well-instrumented rock drill into the area to drill a handful of exploratory holes that
would determine roof strength parameters.
The process of categorizing roof conditions from drilling data has been attempted on
several different occasions. The bulk of attempts have capitalized on Teale’s original
calculations of drilling parameters and how they relate to the intrinsic specific energy (ε ) of the
10
rock (1964). The equation for specific energy obtained from drilling parameters can be seen
below:
𝐹 2∗𝜋 𝑁𝑇
ε (𝑝𝑠𝑖) = ( ) + ( )( ) Equation (2.1)
10
𝐴 𝐴 𝑢
Where T (in.*lb) is the torque applied to the drill string, u is the penetration rate (in/sec), N is the
rotational velocity (rev/sec), F is the penetration thrust (lbs), and A is the area of the hole (in.2)
(Teale 1964). This equation is the sum of the constituent elements of a drill’s cutting mechanism,
a rotational scraping and a thrust gouging. Taking these two elements into account, the amount
11
|
Virginia Tech
|
of energy require to remove rock material can be calculated, this same principle will be later
discussed to describe energy used to excavate rock via scratch mechanisms. In recent years,
several additional authors have used this information to try and incorporate it into the analysis of
strata properties. Most recently, the application of systematic evaluation of drilling parameters
(and therefore strata properties) as well as a suitable background on the evolution of research in
the field can be found in the 2013 work by Bahrampour, et al..
Adapting the technique of using drilling parameters to analyze rock would be useful if it
could be made portable and effective. This method was explored in 1996 by Reddish and Yasar
wherein an ammeter was run in line with a hand drill that was attached to a drill mount to
determine the electrical current applied to a motor for torque and rpm values. Useful parameters
about drilled rock samples, namely intrinsic specific energy and therefore UCS, were obtained
by standardizing the bit properties and the torque/rpm applied to the bit and by using a mount to
keep a standard penetration pattern (Reddish and Yasar, 1996). This rock analysis process could
be made relevant to ground control experts because they would be able to take hand size samples
from the roof strata and extract the strength characteristics of the material.
A further application of this test method would be to note the location of collected
underground rocks and take them to a lab on the surface to have them analyzed and categorized,
thus giving a location specific database of strength properties of certain roof layers. This
systemic approach would also bypass the issues of rendering this hand drill safe for methane air
mixtures as the drilling and analysis would take place outside of the mine environment. This test
would be biased towards the shallowest roof layers (the skin layer) as it is the one most likely to
be falling at any given time. Nonetheless, this idea still may prove useful for providing inputs to
determine ground control techniques for controlling the behavior of the skin and immediate roof.
12
|
Virginia Tech
|
where L is the applied force in (in kN), and D is the tip penetration depth (in mm) (1998). These
values are recorded throughout the experiment and when plotted on the same graph, stop at the
point of first chipping in the sample. If no chipping occurred, then the reference becomes
penetration depth at a 20 kN load or the associated load at the predetermined depth of 1 mm,
using whichever condition that is reached first (Szwedzicki, 1998). The Indentation Hardness
Index was correlated with uniaxial compressive strength (IHI) to yield an obvious trend, the
trend can be expressed by the equation:
𝑈𝐶𝑆 = 3.1∗𝐼𝐻𝐼1.09 Equation (2.3)
where UCS is in MPa (Szwedzicki, 1998). The simplicity of this test is one of its advantages,
requiring few inputs and no extreme testing procedures or tools. Additionally the variability in its
results are comparable to other rock strength test such as the UCS test and the Brazilian Tensile
Test (Szwedzicki, 1998). It also reinforces the notion that UCS indices can be obtained from
simple tests looking at forces and displacement within the rock.
The final rock strength analysis method to be analyzed in this paper is the scratch test
methodology developed by G. Schei and E. Fjᴁr SINTEF Petroleum Research based on work
conducted by University of Minnesota professor Dr. Emmanuel Detournay. This sedimentary
rock testing technique is predicated on continuously logging information about certain cutting
parameters of a scratch bit that is equipped with precise kinematic and force controls and is
dragged along the length of a core sample that is saddled in a housing (G. Schei et al., 2000).
This type of testing, commonly referred to as “scratch testing”, shows great promise for
providing inexpensive, quick and useful information about strength characteristics of
sedimentary rock without the need to destroy the sample in the process (G. Schei et al., 2000).
14
|
Virginia Tech
|
The test works by compiling force and cutting depth values and assessing how they correspond
to rock strength properties such as compressive strength and elastic modulus.
Primarily, the test controls the depth of cut and the velocity at which the cutting head
moves along the core sample (on the order of several mm/s) while monitoring the force applied
to a cutting head (G. Schei et al., 2000). The depth of cut and velocity of the cutter are controlled
electronically by a computer that sends user inputs about preferred depth and velocity to stepper
motors that adjust these parameters. Schei et al. explain that the reason that cutting depth is
controlled is that for shallow depths of cut, between 0.5 and 2 mm, the rock behaves in a ductile
fashion along the leading edge of the cutting surface (2000). When the cutting depth is increased
past this range, macro-scale rock failure behavior takes over and the rock begins to fail with
larger, more sporadic failures in the form of chipping (Suarez-Rivera et al., 2002). It is in the
ductile region that the scratch test is performed.
The value of horizontal force used in the scratch test is essential for computing the value
that actually correlates with uniaxial compressive strength, intrinsic specific energy (ε ). This
10
unit is the same as what was mentioned in R. Teale’s research and equals the amount of energy
needed to remove a given rock volume, in this case by scratching it (Suarez-Rivera et al., 2002).
The formula for this value as obtained from Suarez-Rivera’s 2002 paper is seen below:
𝐹
𝜀 (𝑝𝑠𝑖) = ℎ Equation (2.4)
10
𝑤∗𝑑
where F is the horizontal force average from the tested zone, the w is the cutter width and d is
h
the depth of cut (width of cut times depth of cut equals extracted rock area). Note that the units
for Equation 2.4 are equal to that of pressure, which is the unit for intrinsic specific energy
(Energy per unit volume) where one of the length dimensions from the energy and volume term
15
|
Virginia Tech
|
Overview of Current Rock Analysis Devices
This scratching technology was brought to patent status by inventors Bertrand Peltier,
Emmanuel Detournay (mentioned earlier in scratch test research), and Anthony Booer. In this
patent, US 5,323,648 A, a tool for being lowered into a gas well borehole with scratching
capability was described, featuring transducers for measuring forces and scratch depth. The
scratchers suggested in the patent are made from Polycrystalline Diamond, and are imparted into
the rock by an unspecified force generating element (US Patent No. 5, 323,648 A, 1994). No
information regarding the successful development or implementation of this device in a field
setting was found.
The Formation Evaluation Tool patent was later referenced for the development of a
laboratory core log analysis device for which the patent was awarded to Terratek inc. out of Salt
Lake City, Utah. This device, patented under the title “Apparatus for Continuous Measurement
of Heterogeneity of Geomaterials”, was invented by a team of individuals of which Roberto
Suarez-Rivera (author of a technical paper on scratching referenced earlier in this paper) was a
member. This device functions by traversing a scratching head under precise kinematic
conditions and closely monitoring the resulting force and depth of cut being applied to the
scratcher (US 8,234,912 B2, 2012). This device entered development and is used by
Schlumberger to do scratch evaluations on cores that they logged for the development of oil and
gas wells.
An additional scratching device has already been patented for evaluating borehole wells
in situ. The patent is held by Chee Phuat Tan of Kuala Lumpur, Malaysia under the assignment
of Schlumberger Technology Corporation. This patent descriptions explain that the device was
designed to function by making scratches in the rock mass with powered arms and measuring
17
|
Virginia Tech
|
Product Design
Design Introduction
The objectives of this design are for the device to identify changing layers of rock by
interacting with the wall of a borehole. To take full advantage of all the information that may be
contained in the borehole, it is important to try and extract information such as rock strength and
changes in rock type that the previously explained tape measure/inspection hole test overlooks.
Further, the use of outputs from this device to calculate strength values of the rock layers that it
is interacting with is desired.
The design of this device must abide by several constraints in order to be useful in the
location in which it is expected to work. The first element is the size constraint, the device must
fit in the legally required test holes which are commonly one inch in diameter. This prevents the
complication of drilling another hole in the roof and opens up many old areas to roof analysis. It
additionally is constrained to be safe to operate, portable and quick to assemble. It must function
in a self-contained manner, based on forces that a user generates, and with no extra power
systems being run to it and it would ideally be permissible for methane air mixtures.
Given the size constraints on any device expected to fit in a test hole, a certain hierarchy
was given to each possible rock analysis method so that the best process would be used. Issues of
portability remained in the forefront of the design choice, but seeing as it is accomplished on
account of the scale of the device’s operating conditions, namely the one inch hole, and the
widespread availability of transport equipment in most mining settings, maintaining mobility was
easy. The acknowledgement of the size of the borehole led to the conclusion that the more of the
19
|
Virginia Tech
|
device that could exist outside its very limiting size constraints, the better a design choice it
would be.
Immediately, issues arose when considering the use of the indentation analysis system
namely because of the restrictions on size. When the experiments were conducted by
Szewedzicki, the large hydraulic cylinder allowed precise displacement control normal to the
rock face (1998). Furthermore, they had the potential to generate tens of kilonewtons of force
and often had forces of that magnitude for Szewedzicki’s test (1998). This is prohibitive in the
field setting because in order to increase the force imparted on the rock surface by hydraulic
pressure, the hydraulic cylinder must either increase in area, or the indentation tip must be
reduced in diameter.
Noting that the hydraulic piston could never get larger than the borehole diameter of one
inch (and even in the best conditions would still need to be considerably smaller than that), an
enormous amount of hydraulic pressure, on the order of 20,000 psi, would need to be generated
to get comparable forces (4000-8000 lbs) seen in Szewedzicki’s 1998 experiment. This would
also force the user to have to maintain the hydraulic fluid levels in the system, not impossible,
but adding an undesired level of complexity. Additional pitfalls with indentation testing include
the safety aspect of working with high pressure fluids as well as the lack of precise pressure
control of the hand pumps that would have to be used to generate the necessary pressures.
The borehole size also restricts the ability to determine the amount that an indentation tip
displaces into a rock surface, which is necessary to derive UCS by Szwedzicki’s methodology
(2002). Precise linear displacement transducers that determine such movement are very
expensive and one could not be found that could conceivably be expected to fit in a borehole in
such a configuration that it would be able to measure relevant displacement. Especially
20
|
Virginia Tech
|
Scratch Head and Scratchers
The size constraints indicated that a scratching mechanism may be the most useful rock
analysis method for such a small area. It provides a useful, quick, nondestructive method for
determining position-specific (and therefore strata-specific) mine strata characteristics. This
categorization would be useful for analyzing the competency of the anchorage layer of roof bolts
and cable bolts as the anchorage depth for these devices is regulated in a mine. Moreover, by
having a continuous scratch log of the wall of the borehole, any serious discontinuities may
manifest in the data, which would provide useful information about the jointing network in
addition to the roof formation’s strength characteristics.
The acknowledgement of the merits of the scratch analysis method require one to
consider the way that a scratcher would be inserted into the borehole. Firstly, the scratcher has to
constantly be applying force to the rock face that lodge the scratcher tips deep enough in the rock
surface to enter the ductile rock failure phase of scratching. It is in this depth of rock scratching
(0.5 to 2 mm) that Schei et al. explained that scratch tests are valid and that their equations
explain rock failure (2000). Insertion of this mechanism into the rock mass would ideally be
done in a manner where the scratch tips would be allowed to expand into the surrounding strata
after insertion. In other words, this device works by an unobstructed insertion followed by
expansion of the scratch heads and then a well instrumented, controlled removal of the resistive
scratch head wherein the relevant scratching parameters would be monitored. It is under this
basic design principle that the Mine Roof Strata Analysis Device (MRSAD) was created (in
other literature about this project, the device is referred to as the In-Situ Technical Compression
and Hardness Evaluation System, or ITCHES).
22
|
Virginia Tech
|
The scratch head is the pivotal element of the system and its proper design is essential for
the device to function as the theory would require. The scratch head serves as the housing for the
scratching mechanism as well as the element that translates forces generated by the pull of the
user to the scratcher mechanism and then ultimately to the rock surface. The scratch head is a
modified one inch diameter steel rod with a hole drilled the full diameter of the rod perpendicular
to its long axis and a 5/8” diameter rod welded to the top to act as a wraparound for the tension
cable. The material chosen for the head is stainless steel because of its hardness and resistance to
oxidation, which was expected upon use in a moist environment. The following image, Figure
3.1, was taken of the scratch head, with a number three on it, displayed next to a one inch
borehole:
Side
Scratcher
Opening
Scratcher
Cavity
3
Figure 3.1: Scratch Head Immediately Prior to Insertion with Red Box around Side Scratcher
Opening and Blue Box around Scratcher Cavity
The hole that the scratcher tips comes out of is viewable just to the left of the installers thumb,
highlighted by the red box. This hole, called the side scratcher opening, continues through the
housing to the other side of the head. There is a cavity on the bottom of the head, outlined in the
blue rectangle, which provides an area to aid in the installation of the scratcher and scratch head.
23
|
Virginia Tech
|
In the following image, Figure 3.4, this process is shown in a similar fashion to the one
seen in previously in Figure 3.3:
Force
Figure 3.4: Retraction of Blue Scratcher Arms into Scratch Head Cavity from Downward Force
Take note of the fact that the loop moves downward from an induced force between Figure 3.3
and Figure 3.4, indicating that tension is being put on the loop of the scratcher causing the tips to
retract into the openings of the head. The scratcher loop seen in Figures 3.3 and 3.4 would be
pulled down by hand by a loop of wire that runs out the length of the borehole, this added tension
pulls the tips of the scratcher into the head housing and the head can then be inserted into the
borehole without the resistance of the scratch heads against the strata.
The scratchers are made from ASTM A228 stainless steel music wire of diameters
0.045”, 0.051” and 0.055”, as all were explored as possible sizes. Music wire was selected
because of its hardness and resistance to fatigue, while still having the flexibility to undergo the
27
|
Virginia Tech
|
necessary deformation to be installed. The wire scratcher does scratch the sides of the borehole
upon insertion as is visible in the following figure, Figure 3.5:
Figure 3.5: Scratch Demonstration on Sandstone Sample
This scratching demonstration was performed on a sample of sandstone in the lab. The lines
coming down the side of the hole are from the scratchers housed in the head. No data was
collected from this particular scratching test, it served simply to verify that there was the capacity
to install the scratcher according to the method of pulling the scratcher loop into the cavity and
inserting it in the borehole. The image above indicates this was done a number of times as the
areas of rock removal are clearly outlined on the borehole wall. The scratcher is installed by way
of a 0.029” pull wire that wraps around the loop at the base of the scratcher in the cavity, and
applies tension to the scratcher, causing it to deform into the cavity at the base of the head.
A user generated force that pulls on the scratch head is the process by which the scratch
head is retracted from the hole and the primary means that energy gets imparted into the rock
mass. The scratch head has a loop of steel cable that runs through the body over the top of the
28
|
Virginia Tech
|
threaded bolt, called a connecting bolt, which is welded to the cable. This bolt can be unscrewed
to isolate the scratching head and corresponding tension cable as one unit. When the unit is fully
installed, with the scratch tips extending into the borehole wall, a coupling at the connecting bolt
will provide the linkage for the rest of the tension cable to extend the remainder of the way out of
the hole. It is outside the hole where tension is applied by the user to move the scratch head
through the rock mass.
The interdependence of the borehole, the scratcher and the head that houses it is the
principal design feature of the MRSAD testing unit. This means that the instrumentation choices
were to be made after the analysis method was chosen as it was the core of the system design.
When it was determined that scratch testing was going to be the method of rock analysis, the
relevant parameters needed to be fully instrumented. It was known from the 2002 Suarez-Rivera
et al. paper that the formula for intrinsic specific energy as a result of rock scratching had three
inputs, horizontal scratching force in the numerator with scratch tip width (0.045”, 0.051” and
0.055” diameter scratchers are used in the design) and scratching depth being multiplied in the
denominator (Equation 2.4 in this paper). This requires the proper instrumentation of the force in
the direction of scratching. Based off the use of pull forces, the movement of the scratch head
would directly correspond to the movement of the cable that it attaches to and therefore any pull
force in the cable is also applied to the scratch head. It is also important to note that any force
imparted on the head detected by instrumentation will be divided over its two points of contact
with the borehole wall.
30
|
Virginia Tech
|
Instrumentation
A load cell was constructed to relate the amount of force that is applied to the scratching
head. This strain sensing element needs to be durable, accurate, and small enough to fit in the
borehole with enough space to allow the presence of necessary signal wires and any additional
installation devices. It was decided that strain gauges mounted to a specially designed element
would serve well as a force transducer and would provide enough extra space to work with the
size constraints.
The design of a strain element is predicated around determining the expected values of
strain beforehand based on predicted load on the element and the elastic properties (Young’s
Modulus of Aluminum = 69 GPa) of the strained material. If the geometric dimensions of the
strain element are known, then forces distributed over the area lead to calculable stresses. These
stresses correspond to strain values by way of Hooke’s law and elastic moduli. When strain
gauges are applied to strain element, their bond to the material can be rendered ineffective if they
are overstrained. It is important to ensure that applied force values should be within a range that
these damaging levels of strain are not reached. Strain gauges are analog devices which only
change their resistance in from deformation due to forces, meaning that the smallest useful strain
value is limited by the accuracy of the data acquisition system and the environmental noise.
An I-shaped tension rod was designed to translate the tensile forces to the strain gauges.
The rod is wider at the ends so that there is enough room for connecting bolts to screw into to the
tension rod, the rod then thins in the middle to provide a flat surface to which the strain gauges
are affixed.
31
|
Virginia Tech
|
the thickness of this tension element was designed to be wide enough to ensure easy application
of the strain gauges but not so thin that it interferes with the durability of the tension member
The configuration of the strain gauges was done at the recommendation of strain gauge
design guides such as “The Strain Gauge” and “Strain Gauge Configuration Types”, both sources
are web documents from leading instrumentation manufacturers that indicate the merits of a full
Wheatstone bridge (their source information is available in the references section). This positions
two of the four strain gauges parallel to the primary deformation direction and places two that
run perpendicular to this deformation, but still in the same plane as the first set of gauges. This
full-bridge configuration has the added benefit of automatic temperature compensation (“Strain
Gauge Configuration Types”). The strain gauges and installation kit were purchased from Micro
Measurements in Raleigh, North Carolina. The strain gauges were the 250BF-EA-13 model at
350Ω resistance each. The wires were connected with 134-AWP Solid Copper Wire included in
the GAK-2-AE-10 installation kit, while 22 gauge AWG wire was used on the more exposed
parts. The voltage signal coming off the gauges was sent via a 25 foot 4 channel, shielded,
braided and jacketed cable (Model 426-BSV) also purchased through Micro Measurements. This
cable was then wired to the I/O module on the data acquisition system through the five volt
excitation port, ground port and +/- inputs.
The tension force transducer had a number of features that protect it from some of the
damaging circumstances that it was likely to experience in a borehole setting. Firstly, the profile
of the wiring was kept as low as possible, this reduced the outer diameter of the tension member
to reduce the likelihood that parts of it would snag on things in the hole or during installation.
Secondly, the entire tension device was insulated with 5/8” flexible plastic tubing that was cut
34
|
Virginia Tech
|
and then taped to the exterior of the device. The protected strain gauge can be seen in the in
Figure 3.9 below:
Figure 3.9: Force Transducer in Protective Plastic Jacket
the series of interwoven wires coming off the right of the protected load cell are the signal wires
from the strain gauges. This covering protected the transducer from shock, abrasion and puncture
and when that was wrapped in electrical tape, became very robust, while still being small enough
to fit in the hole. The last design requirement was the use of connecting terminals between strain
gauges. This provided a buffer between the different gauges that if there was any strain put on
the signal wire that wasn’t absorbed by other preventative features such as the taped tubing, the
stress didn’t manifest itself on the gauges themselves, which could damage them beyond what
could be repaired in a field setting.
The method for determining the position in the borehole, as well as observing the
velocity of the head through the hole is to use a string displacement transducer, otherwise known
35
|
Virginia Tech
|
as an extensometer. The extensometer, otherwise known as a linear position transducer, chosen
for this system is the Unimeasure HX-PA-300-L3M. Extensometers function by placing a
variable resistor in connection with a rotating shaft, with the shaft affixed to a string whose
position is changing. A voltage change across the resistor, corresponding to a change in position,
can be measured by a data acquisition system with a voltage readout. The end of the string on the
extensometer was to be attached to the scratching device somewhere just below the tension
sensing member, that way any force applied by the winding spool on the extensometer would
automatically be accounted for in the force on the scratchers. According to the Unimeasure, inc.
datasheet for this device, the tension on the spool is 2.25 lbs.
The range for the device is to be at least 20’, allowing for a seven foot mining roof height
as well as a 13’ journey up the borehole. A steel wire extensometer made by Unimeasure was
chosen that had a range of 25’, allowing for a buffer to prevent overdrawing the spool which
would damage the device. The potentiometer in the transducer is a one kΩ, ten-turn resistor that
has a linear taper and was attached to the winding shaft by way of a precision gear. The signal
wire coming of the extensometer has a ten foot length, which was suitable to attach it to our data
acquisition system (an image of the signal connection between an extensometer and our DAQ
can be seen in Figure A.1 in the Appendix), while keeping more fragile electronics out of the
way of the operating area.
36
|
Virginia Tech
|
Data Acquisition and Management
The data acquisition unit used in this experiment is the National Instruments USB-6211
module. It features a USB bus port that permits configuration with National Instruments
LabVIEW software to allow for easy data acquisition and management. The data acquisition
system (DAQ) features an I/O module, on board five volt excitation for powering laboratory
instruments, and has ports for analog as well as digital signals. There is no earth ground for this
device, only a chassis ground, so all instrumentation cable shielding is grounded separately from
the module via a cable that is hooked to a metal anchor. The ground circuit on the chassis simply
provides a reference for the excitation voltage, without it, no current flows through the
instrument. The device is not MSHA permissible, but it was assumed that the data acquisition
system would function similar to one that was safe for methane air mixtures and that a future
suitable data acquisition system could be substituted for this one. The USB-6211 runs off the
USB device it is plugged into, making this project’s entire instrumentation system (the strain
gauge transducer and extensometer) fully portable when combined with a charged laptop, giving
hours of portable use.
The LabVIEW software features a unique feature for assisting in making the DAQ
communicate with the program. A module within the LabVIEW software, called DAQMX (or
DAQ Assistant) contains all the elements to collect information and process it electronically. For
this project, two channels were configured in DAQMX that allowed separate inputs both
transducers while allowing them to run off the same five volt excitation source.
37
|
Virginia Tech
|
Mounting and Installation System
A few methods of applying tensile force to the cable were explored. It was first thought
that a constant-velocity electric motor would be used, but it was determined that this would be
prohibitively expensive and difficult to implement/power in coal mines. Then the thought of a
person simply pulling the device out of the hole by hand was considered, but the risk of injury
and lack of ability to adequately control pull velocity rendered that idea unusable. This led to the
adoption of using a winch, this would simultaneously handle the issues of organizing the tension
cable as it came out of the borehole as well as providing a safer, more precise, tension generator
that doesn’t rely on electricity.
The winch selected for this device was a 1500 lb hand-cranked winch made by Torin Big
Red Jacks. The spool had a selector that could do smooth, uninterrupted coiling or extending and
it could do ratcheting coiling or extending so that no matter which direction was under load,
hazardous and undesired slipping would not occur. A plastic drum was added to the original
winding spool to increase the winding diameter to 3.45”, this would make the winding process
faster which is important considering how long it would take to hand wind the 30’ attached to the
winch. The increased wind up rate is due to the fact that each revolution had a larger
circumference around which the cable wound, but this also put the steel cable under less stress to
wind around the shaft which is safer, more organized and prolongs the life of the cable. It was
also determined that the same 1/8” cable used for the scratch head would also be used for the rest
of the tension cable.
In order to ensure a controlled insertion of the signal wires, tension cables, scratch head,
scratchers and force transducer, an installation rod was designed. This rod is simply a piece of 4’
long conduit pipe of around 3/4” with a slot cut the entire length. The top one foot of the device
39
|
Virginia Tech
|
features the majority of the pipe cut away to make a small pushing element. The idea behind this
design is that all the wires will go through the slot into the center part of the pipe, protecting
them from pinch points in the hole that could damage them. The part at the top that is cut away
serves as a platform for the tension transducer (the largest diameter item in the borehole) and
gives a tip with which the scratcher can be pushed up the hole. This device can be seen in
Figures A.7 and A.8 of the Appendix. Upon full installation of the system in the borehole, the
pipe could be removed without pulling any of the elements, all the while the cable would be fed
through the slot cut through the side. This would leave all the cable in the hole without cutting it
or damaging it while also removing the installation rod.
It was determined that for reasons of stability and consistency, that the entire MRSAD
system would be affixed to a post that would brace itself against the roof and floor of the mine.
This would reduce the problem of any tension (applied to the cable to remove the scratchers)
manifesting itself in ways that would lead to motion of whatever was applying the tension either
a person, winch or motor. Considering mines may require testing in low, and higher coal
situations, this element was required to have a good deal of versatility in what variety of roof
heights in which it could operate.
There are several advantages of using the stand that made it well suited for its purpose. It
is simple in construction and quite robust meaning it can be assembled and dissassembled
quickly and easily (on the order of one minute) and it won’t get damaged from being stored in a
container. The stand is comprised of a four-post base that provides stability from it tipping over,
a series of fitted middle sections and a screw top for fine adjustments. The stand provides a
platform to affix other elements of the MRSAD device. The winch and displacement transducer
were all attached to the stand, keeping them from moving under the tension put on them, as well
40
|
Virginia Tech
|
Upon conclusion of the construction of the MRSAD, a testing facility was identified that
possessed resources that would be suitable to evaluate some of the features of the device and to
see if the theory behind the remote field scratch test were valid. The people at the research
facility indicated that they were not able to put a rock sample in the roof mount as was planned
originally and that modifications would have to be made to the MRSAD unit in order that it
retain its function in the horizontal direction. This required slight modification of a few of the
elements as well as relying on some useful coincidences with previous design choices. The
primary concern was the need to change the winch and extensometer cable direction from
vertical to horizontal. Due to the way that the winch was designed, with several crossbars across
the body, the pulley and extensometer wire were simply snaked around one of these crossbars
and allowed them to function sideways. In order to compensate for friction on the cables as well
as any damage that may be induced from wrapping cables around small diameter cylinders, a
larger pipe was fitted to one of the crossbar elements and seemed to serve very suitably for stress
relief.
Additionally, the MRSAD tower has the ability to brace against a rock sample
horizontally. For this change in orientation, a series of three inch by three inch by two inch
pieces of wood were arranged to brace against the base and keep it from sliding toward the
direction of the tensile force. In a prior iteration of the design cycle of the MRSAD brace, a 1/4"
threaded bolt was put through the frame, an additional piece of wood was drilled with a hole the
diameter of that bolt and this served as a mount for this element.
44
|
Virginia Tech
|
Calibration
Expected values for tension to remove the scratch unit from its installation were on the
order of 20 to 100 lbs. The tension values were estimated by considering the amount of force that
a person could generate with their own strength, as the device was constrained to not be
externally powered. The first calibration was done with 0, 20.325, 40.325, and 47.110 lbs, which
were the order of magnitude of the forces that a person could generate to pull on the device.
Calibration of the force device was predicated on taking 1,000,000 samples from the strain
transducer as weights of increasing mass were applied to induce tension on the load cell. This
sampling was done for four different weights (including zero) three times each. The gauge
output can be seen in Figure 3.14 below:
Voltage vs. Applied Tension y = 5E-07x + 7E-0
R² = 0.9881
0.00003
0.000025
s
t lo 0.00002
V
,t
u0.000015
p
t
u
O 0.00001
e
g
a t0.000005
lo
V
0
-0.000005
0 5 10 15 20 25 30 35 40 45 50
Applied Tension, lbs
Figure 3.14: Voltage vs. Applied Tension with Trend Line, Equation and Coefficient of
Determination
The trend line and coefficient of determination are also visible and indicate a linear voltage
change with changing force. There was an error with the data reader that was not recognized
46
|
Virginia Tech
|
until after calibration that cause there to be only one data collection test instead of three for the
zero force point. This means the variance for this force value is not as well-known as it is for the
other force values.
The results for the first calibration, while indicative of the linear deformation of the strain
element, did little to indicate the variability of force readings at smaller force changes. This led
to further calibration efforts that focused on smaller scale forces (0, 0.506, 1.016, 3.016, and 5.25
lbs). When the device output was analyzed under the smaller scale loads later seen in the tests,
the following output was obtained, visible in Figure 3.15:
Voltage vs. Applied Tension y = 8E-07x + 1E-06
R² = 0.5996
0.000006
0.000005
0.000004
s
t
lo
V 0.000003
,t
u 0.000002
p
t
u
O 0.000001
e
g
a 0
t
lo
V -0.000001
-0.000002
-0.000003
0 1 2 3 4 5 6
Applied Tension, lbs
Figure 3.15: Voltage vs. Applied Tension with Trend Line, Equation and Coefficient of
Determination
Note the trend line and coefficient of determination, these values are not the same as in the
original calibration. Additionally, the variation of the voltage at the zero force level has a
dramatically different variance than the rest of the values. If this zero-force voltage output is
neglected and the graph properties recalculated, the coefficient of determination returns to its
47
|
Virginia Tech
|
previous level of precision. This omission of the zero-force could be justified on the grounds that
the tension sensing member would never actually be measuring forces at the zero-load level
because at the very least, it will have the tension of the extensometer applied to it. Additionally,
when the full device comes out of the hole, the scratch head will still be attached to the tension
transducer and will continue to apply tension as it dangles from the MRSAD unit. The modified
graph can be seen in the next image, Figure 3.16:
Voltage vs. Applied Tension y = 4E-07x -9E-08
R² = 0.9893
0.0000025
0.000002
s
t
lo 0.0000015
v
,e
g
a
t 0.000001
lo
V
0.0000005
0
0 1 2 3 4 5 6
Force, lbs
Figure 3.16: Voltage vs. Applied Tension - Omitting Zero-Force Values - with Trend Line,
Equation and Coefficient of Determination
Take note of the graph returning to an almost 99% linear fit among the data. There are
differences between the first graph, which calibrates weight between zero and about 45 lbs and
the last one which does zero to five lbs. The intercepts can be ignored because they serve only to
shift the outputs up and down. If the zero force value for the graph is set based off a low point on
the graph, then the intercept is not important because the force values change linearly. Since the
plots from the data acquisition system reference the zero force values computed from the graphs
themselves, the intercepts are ignored. The equation from Figure 3.14 is the most justifiable for
48
|
Virginia Tech
|
determining the force exerted on the tension member and is rearranged to relate force applied to
the load cell in the following equation, Equation 3.1:
𝑇𝑒𝑛𝑠𝑖𝑜𝑛 𝐹𝑜𝑟𝑐𝑒 (𝑙𝑏𝑠) = 𝑉𝑜𝑙𝑡𝑎𝑔𝑒∗2,500,000 Equation (3.1)
note that the voltage input is the voltage from the graph. Equation 3.1 converts the force voltage
values from the data to force values. The data were recorded with a scaling factor included, so if
further processing of the results from the original experiments is desired, the above value of
2,500,000 should be reduced to 50,000.
The same calibration principles were applied to the displacement transducer. The string
on the extensometer was drawn out to predetermined lengths and held for 1,000,000 samples,
done three times for each length. These samples were averaged and then all three tests plotted on
a graph which can be seen in Figure 3.17:
Voltage vs. Displacement y = 0.1802x -0.0419
R² = 0.9894
0.6
s
lt
o
V
,t
0.5
u
p
t
u 0.4
O
e
g
a t 0.3
lo
V
t
n 0.2
e
m
e
c a 0.1
lp
s
iD
0
0 0.5 1 1.5 2 2.5 3 3.5
Displacement, ft
Figure 3.17: Voltage vs. Displacement with Trend Line, Equation and Coefficient of
Determination
49
|
Virginia Tech
|
Design Summary
The components for the device consist of a stand, a hand cranked pulling winch, a
position transducer, a load cell, a scratch head, an installation rod and scratchers. These parts,
and the parts that they are made from are categorized in the bill of materials. The bill of materials
for the device is included in the following table, Table 3-I:
Table 3-I: Bill of Materials for Formation Evaluation Tool
Tier Component Unit Number Make/Buy
Mine Roof Strata
Analysis Device
(MRSAD or
1 ITCHES) Tool 1 Make
2 Mount Stand 1 Make
Interchangable
3 Metal Assembly Stand 1 Buy
Adjustable Roof
3 Threaded Bolt Bolt 1 Buy
4 Brace Plate Plate 1 Make
4 Bolt Nuts Nut 3 Buy
1" thick Plastic
Mount for
Extensometer and
3 Winch Sheet 2 Buy
4 Bolt Bolt 2 Buy
4 Nut Nut 2 Buy
2 Hand Winch Winch 1 Buy
3 Mounting Bolts Bolt 3 Buy
3 Spool Spool 1 Make
4 3" Plastic Cylinder Plastic 1 Buy
3 3/16" Steel Cable Feet 35 Buy
4 Connecting Bolts Bolt 4 Buy
4 Crimps Device 3
Unimeasure HX-
PA-300-L3M
Position
2 Transducer Device 1 Buy
3 Signal Wire Feet 10 Buy
2 Load Cell Device 1 Make
52
|
Virginia Tech
|
350 Ohm 250BF-
EA-13 Strain
3 Gauge Gauge 4 Buy
134-AWP Signal
3 Wire Feet 10 Buy
426-BSV Braided
Shielded Signal
3 Wire Feet 30 Buy
5/8" Diameter
3 Aluminum Rod Feet 1 Buy
GAK-2-AE-10
Strain Gauge
3 Installation Kit Kit 1 Buy
Plastic Tube for
3 Protection Feet 1 Buy
2 Installation Rod Device 1 Make
3/4" Conduit Pipe
3 5' long Pipe 1
2 Scratch Head Device 1 Make
Stainless Steel
3 Rod 1" Feet 1 Make
Stainless Steel
3 Rod 5/8" Feet 1 Make
2 Scratchers Make
0.045" Music
3 Wire Spool 1 Buy
0.051" Music
3 Wire Spool 1 Buy
0.055" Music
3 Wire Spool 1 Buy
USB 6211 Data
Acquisition
2 System Unit 2 Buy
The table works on a tier system, the most encompassing aspect of the project is labeled tier one
and in this case is the Mine Roof Strata Analysis Device. Components in tier two assemble to
form tier one items and similarly, tier three assembles into tier two and finally tier four
assembles into tier three. This allows for easy categorization about what is needed for the device
and how the parts come together. The unit section dictates how the product is sold with number
53
|
Virginia Tech
|
Experiment
Concrete Block Scratch Test
The testing of the device took place at a facility that researches the design of mine roof
drilling equipment. This was selected because of its close proximity to Virginia Tech as well as
the availability of pre-drilled 10,000 psi concrete blocks with river gravel that had one inch blind
holes that are drilled four feet into the rock. A picture of the full scale concrete block can be seen
in Figure A.2 in the Appendix. These blocks came pre-installed with two different rock types
that were an artifact of prior research. The presence of these differing rocks in the block matrix
was desired because it means that the MRSAD can look for differing rock types in the block
material in a manner that is similar to analyzing the varying strata in a coal mine roof.
This goal of this concrete block test was to control the velocity of the scratch head in the
borehole and try to see if the force varied with any regularity as a result of scratching and
motion. It was thought that an increase in force over sections of the block would indicate a
stronger or more resistive rock type and that the presence of decreases in force could indicate
weaker rock types. Controlling the velocity of the scratch head for each test with steady winding
of the winch allows the results of one hole’s similar velocities to be compared. In essence, if the
velocity of each scratcher pull test is held constant, the forces seen in each individual attempt can
be compared to other forces in the same attempt because it moved at a constant velocity through
the hole.
The blocks were taken into the research warehouse and the pre-drilled holes were then
analyzed to see if they would be able to serve as good hosts for the scratch head by inserting the
scratch head into each of the holes on the face. Immediately, the geometry of the holes became
problematic because it was difficult to insert the device more than a few inches into the block.
55
|
Virginia Tech
|
There were aspects of the initial drilling of the holes that made them have an undulating profile
along their length which was difficult for the scratch head to move past. However, observed
holes where at least a foot of depth could be reached were considered promising, and were
marked to be used for testing of the MRSAD.
The following image, Figure 4.1, shows the scratch head being installed with the
horizontal installation rod pushing the scratcher into the borehole:
Figure 4.1: Installation of the Scratcher into the Borehole
The MRSAD unit is in the bottom right of the image, it is moved away for the installation to
allow more room to work, and then placed back up against the block for the test. The test holes
and their markings can be seen in Figure A.3 and Figure A.5 in the Appendix.
During the first two attempts at installation, the device did not install correctly and it was
decided that modifications to the scratch head would help with the process. The facility assistant
56
|
Virginia Tech
|
This graph has a distinct increase in force during the motion phase which is indicative of
resistance from the rock. The velocity for the first motion region between 28 and 51 seconds is
0.045 ft/sec (0.54 in./sec) with an R2 of 0.98 for that velocity for that interval. The velocity of the
second motion phase, between 51 and 82 seconds is 0.041 ft./sec. (0.49 in./sec.) with an R2 of
0.999 for the velocity over that time window. The region between 40 and 80 seconds shows an
applied force value that although spiking, hovers around the same value while the displacement
changes constantly over the course of the hole. This motion corresponds to two velocity regions
the first one between about 40 seconds and 50 seconds and then the second, slower one between
50 and 80 seconds. The velocity of the scratch head is dictated by the user carefully cranking the
winch handle and when it is under operation, keeping a consistent, steady pace is needed to look
for changes in the rock type along constant velocity sections.
58
|
Virginia Tech
|
It was decided at this point to resume testing in a different hole, hole two, adjacent to
hole one. The data readout for this test can be seen in the subsequent image, Figure 4.4:
Figure 4.4: Force and Displacement vs. Time for Third Test in Concrete Block
This test begins with a very distinct spike in force at 15 seconds, this is showing the amount of
force it took to initiate motion of the scratch head in the borehole because at that point, the
displacement profile is flat, indicating no motion is occurring. The velocity of the first phase of
motion between 15 and 20 seconds is 0.11 ft./sec. (1.36 in./sec.) with and R2 of 0.989 for that
velocity over that interval. The velocity seen in the second part of the motion phase between 21
and 41 seconds is 0.055 ft./sec. (0.66 in./sec.) with an R2 of 0.996 over that interval for that
average velocity value. It is often the case that the head will lodge itself in the rock with the
60
|
Virginia Tech
|
The pull forces for the successful tests (tests one through three) were averaged over the
motion phases and arranged in the following table, Table 4-I:
Table 4-I: Concrete Block Test Pull Force Averages and Statistics
Scratcher Concrete Block
Diameter Pull Force Averages (lbs) Test Statistics Overall Statistics
(Inches) Test 1 Test 2 Test 3 Test 1 Stdev Test 2 Stdev Test 3 StDev Average Std. Dev.
0.045 4.4795 3.85 10.23 3.49 3.87 10.818 6.19 2.87
The averages were taken in areas where displacement was occurring by highlighting the dataset
with the data selection tool in the MATLAB plot function, making it a variable, and finding its
average and standard deviation. The test statistics in the middle are the amount that the pull
forces varied as it traveled the length of the hole for each test. The overall statistics on the right
are the averages and standard deviations for the three test’s pull force averages.
The results of this test were indicative of the successful application of resistive forces to
the scratch head and borehole as well as a demonstration of the instrumentation and installation
tools. However, there was a variability in the test circumstances especially in controlling
variables such as rock composition, scratcher dimensions and anchorage characteristics which
meant a lack of definitive conclusions. The failure of the mechanism and installation was
important because it illuminated weak points in the design. The data readouts from the failed
tests can be seen in Figure A.25 and Figure A.26 in the Appendix. The decision was made to
reapply the MRSAD device to better regulated circumstances in the lab which entailed testing it
a one inch inner diameter PVC pipe and a small sandstone block to serve as more homogenous
controls.
63
|
Virginia Tech
|
PVC Pipe Scratch Test
The PVC Test serves as a platform for comparing differing wire scratcher diameters in a
more controlled material, in this case PVC plastic. Two wire scratcher diameters, 0.045 inches
and 0.051 inches were used in this experiment. The test paired MRSAD unit with a five foot long
PVC pipe. The pipe was clamped down to a table and then the stand for MRSAD was placed
next to the table with horizontal braces installed. The pipe setup can be seen in Figure 4.6 below:
Figure 4.6: Clamped PVC Pipe with Scratch Head Visible Emerging from Pipe
This image was taken after the scratcher had been set and shows the direction that the head
moves in the pipe as it is drawn toward the far hole. The plastic pipe was longer than the
installation rod used for the concrete block test and this meant that that installation method
would no longer be feasible. Instead, the scratch head was dropped down the pipe with the other
64
|
Virginia Tech
|
cables attached to it having gravity pull it down. The scratcher was installed when the head
emerged from the other side as seen in the following image, Figure 4.7:
Figure 4.7: Scratch Head Embedded in PVC Pipe Walls Prior to Extraction
There is a small gap between the head and the PVC pipe, it is thought that this gap is one of the
most critical elements of the test. If this annulus is too small, the head is more likely to get stuck
in the hole. If it is too wide, the thin scratchers will get bent into the gap and not engage in the
type of scratching behavior desired. Being able to see the scratcher from the other side of the
hole removed a lot of the uncertainty about anchorage behavior because it allowed for a close
look from the other open end of the PVC pipe where it could be easily determined that the
scratch tips had contact with the plastic.
65
|
Virginia Tech
|
the hanging wire. This ensured that the cranking could commence as soon as the data acquisition
program was initiated, instead of cutting in to the data logging time as the spool was wound and
reorganized. When the initial force is the same as the traveling force, it is a byproduct of this
testing procedure.
It was decided that for the third test, the scratch tips would be changed from the 0.045
inch diameter tips to the 0.051 inch diameter tips. With the homogenous nature of the plastic
pipe, more control could be expected for the behavior of the surrounding matrix and the
consequences of changing the diameter of the scratcher wire could be evaluated. A picture
showing prepared scratch tips laying on a 0.196 inch x 0.196 inch grid is visible in Figure 4.11:
0.051”
0.051”
1”
0.045”
0.045”
Figure 4.11: Prepared 0.051” and 0.045” Scratchers with One Inch Parallel Lines Simulating
Borehole Dimensions
All the scratchers are the same width and all extend 1.1811 inches from tip to tip, which is 0.091
inches wider on each side than a one inch borehole. The tips of the wire scratchers were filed
69
|
Virginia Tech
|
after being cut to attempt to control the profile of the scratcher surface. The fact that they are the
same dimensions was important to consider for control as having them at the same cut width,
1.1811 inches, kept ruled out the width of scratchers as a variable.
Once installed, the pull test was performed on the PVC pipe with the new 0.051”
scratchers. Scratching commenced and immediately more resistance was observed, which at this
point felt like gouging of the plastic. The clamps were losing their hold and the pipe had to be
held by two people as the head was drawn out. Again, the time was cut short, with the output for
the first part of this test seen in Figure 4.12 below:
Figure 4.12: Force and Displacement vs. Time for First Part of First PVC Test with 0.051”
Scratcher
Although the force applied to it is highly erratic, the velocity of the scratcher stays largely the
same (R2 for velocity of 0.99) throughout the first part of this test at 0.037 ft./sec. (0.44 in./sec.).
70
|
Virginia Tech
|
Take note that the large magnitude tensile forces correspond strongly with the motion through
the PVC between 10 and 60 seconds. The velocity of the motion over the interval between 12
and 60 seconds is 0.043 ft./sec. (0.516 in./sec.) with an R2 of 0.998.
Table 4-II is shows the averages of the forces for the differing scratch tip diameters in the
PVC tests and can be seen below:
Table 4-II: PVC Test Pull Force Averages and Statistics for Differing Scratcher Sizes
Scratcher PVC Pipe
Diameter Pull Force Averages (lbs) Test Statistics Overall Statistics
(Inches) Test 1.1 Test 1.2 Test 2 Test 1.1 Std. Dev. Test 1.2 Std. Dev. Average St. Dev.
0.045 2.48 3.23 2.71 0.48 0.69 2.81 0.38
0.051 39.27 33.65 n/a 7.35 7.69 36.46 3.97
This table highlights the difference in force between the PVC tests. The test showed a large
change in magnitude of pull force from an increase in scratch tip diameter, 2.81 lbs as opposed to
a 36.46 lbs average. The test statistics section in the middle refers to how much the forces were
varying during the testing phase for test one for each scratcher type. The test with the 0.051”
scratch diameter had more force variability over its motion (7.35 & 7.69 lbs) than the 0.045”
scratch diameter (0.4771 & 0.69 lbs). The overall statistics to the right are the averages and
standard deviations of the pull force averages in the left section of the table.
When the scratch head was removed from the hole, it left interesting marks on the inside
of the pipe. A pairing of deep cuts corresponding to the diametrically opposed scratch tips
traversed the entire length of the pipe’s inner surface. The marks from the first two tests (0.045”
scratcher diameter) were almost indistinguishable compared to the profile of the pipe, but the
third (0.051” scratcher diameter) left cuts on the pipe.
72
|
Virginia Tech
|
Sandstone Scratch Test
0.045”, 0.051”, and 0.055” wire scratchers were then used on a new test, one looking at
their interaction with a small sandstone sample. This was to mimic the control of the PVC test
but while still testing the device on rock. The holes after being tested can be seen in the
following image, Figure 4.15:
0.045” 0.051” 0.055”
Figure 4.15: Sandstone Holes after Testing the 0.045”, 0.051” and 0.055” Wire Scratchers
Markings are visible on the rock surface resulting from contact with the scratchers. The sample
was held to the table by a series of clamps and was carefully monitored to ensure that it was not
undergoing slippage during the duration of the test. This rock was a suitable length to obtain a
velocity profile and to measure the associated pull force as can be seen in the next several
images. The plots of time, displacement and force were recorded and graphed for each run of the
tests. The results for the first test from every scratch diameter are displayed below with the
remainder of the graphs being available in the Appendix in Figures A.9 through A.24.
74
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.