University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Virginia Tech
|
Figure 7.20 EtherNet/IP Communication
Figure 7.21 Relay to Gateway Communications
In order to validate the system, packets were examined upon report MMS report header.
Taking the timestamp between consecutive reports on the IEC side determines the degradation of
gateway performance while loading. In addition, the time between EtherNet/IP messages to the
controller, based upon connection streams, determines how the gateway module can be loaded
on the controller side. Figure 7.22 shows an example of how analysis was performed. In this
example, the RPI, or requested packet interval, from the gateway to the controller is defined as
100 milliseconds. This means that information is first read from the gateway to the controller,
then written from the controller to the gateway. For this example, 10.203.37.57 is the gateway
module and 10.203.37.170 is the automation controller. The green boxes represent writing from
the PAC to the gateway, and the red boxes represent the reading from gateway to controller. In
this example, the first read from the controller occurs at 0.706740000, and the next occurs at
0.806760000. This is a difference of 1.020 milliseconds, an acceptable rate based upon a 100-
millisecond RPI time. The difference between the two green boxes for output comparison is 1.3
156 | P a ge
|
Virginia Tech
|
milliseconds, which is still acceptable. For the upcoming tests, the input stream from the
gateway to the controller is analyzed.
Figure 7.22 Data Transfer Example
In order to validate the device support of the gateway module solution, the following test
was conducted. Twenty identical feeder relays were gathered and assembled. A sample CID
file, including one data set of 40 fundamental parameters, was then created and downloaded to
each relay. One by one, a relay was added to the gateway module and controller.
The RPI, or requested packet interval, between the controller and the gateway module
was then reduced in increments of five milliseconds until the module did not respond or the CIP
connection was broken, i.e., there was no response by 2.5x the RPI time.
The dataset generated and downloaded to each relay can be seen in Figure 7.23. This
dataset consists of both fundamental electrical parameters, such as voltages, currents, and
powers, as well as thermal metering information from various RTD points. This dataset was
selected because it was a good representation of parameters that are of interest for mining
applications.
157 | P a ge
|
Virginia Tech
|
Figure 7.23 Test Dataset
A system was constructed, as shown in Figure 7.17, with a network of IEDs, managed
switch, PAC, and computer running the wire shark packet analyzer. To begin the testing
procedure, one feeder relay was added to the system. The RPI time between the controller and
gateway module was set to a value that resulted in a stable connection, in this case 40
milliseconds, as shown in Figure 7.24. Once the RPI change was made, the first three wire shark
time differences were taken between report publication to the controller and averaged, and the
resulting value was displayed in the chart as the actual RPI time. The RPI time from the
controller was then dropped by five milliseconds, and the procedure was repeated. The
procedure was repeated until the actual RPI time exceeded 2.5 times the requested RPI time,
which is the point defined by ODVA when the connection between the controller and remote I/O
device, i.e., the gateway, is to be terminated. The red line in Figure 7.24 shows this relationship
graphically. When the blue line intersects the red line in the plot, the connection timed out and
became faulted. At this point, the connection was terminated and the test trial was concluded.
An additional IED was then be added to the system and the process would be repeated. It can be
seen that the system failed at the 15 ms RPI time.
158 | P a ge
|
Virginia Tech
|
10 IED Module Performance Test
200
]
s
m 150
[
e
m
iT
100
Transfer Time
IP
R
la
u 50 Connection Time
t
c A Failure
0
45 50 55 60 65 70
Scheduled RPI [ms]
Figure 7.26 10 IED Test Results
Figure 8.27 shows test results for all 20 IEDs (in this case, feeder relays) for the system.
The curve can also be described as the capability curve of the system since it defines the
maximum operating points for the gateway module at each discrete transition in the system, i.e.,
the addition of an IED. It can be seen that the system can accommodate up to 20 IEDs as
specified by the conceptual design. Moving 40 parameters from 20 independent relays can be
achieved at an RPI rate of 100 milliseconds, leaving 300 milliseconds, or 75% of the time, for
other calculations.
The maximum number of IEDs supported by this module is 20 as that is the maximum
number of IEC 61850 driver threads that can run at once. The maximum scenario was tested
with a valid set of parameters to transfer to the control system and performance was set at 100
milliseconds for the data to transfer from gateway to controller for a stable operating point. It
was determined that the RPI time of 100 milliseconds was acceptable based upon current load
shedding algorithms. As a result, the device support portion of the conceptual design was met by
this testing.
160 | P a ge
|
Virginia Tech
|
8. Chapter 8 - Practical Implementation
Introduction
This chapter describes the practical implementation of the mine monitoring system
described in this dissertation. The practical implementation of the monitoring system was
installed at a bauxite mine to monitor the overall energy consumption and provide critical load
shedding functionality for the process. This chapter describes the overall system and how it is
utilized in real time production of the mine.
System Description
The system contains multiple sites of a mining operation that are separated over multiple
kilometers (km) of distance. In total there are approximately 350 IEDs distributed throughout
the system. This system is fed from two 22-kV feeders that monitor an intertie in the system, as
shown in Figure 8.1. Two important locations to be noted in Figure 8.1 are the 11-kV Generator
Buses A and B, located on the left of the diagram. An electrical intertie is located between them
that needs constant monitoring for load shedding possibilities.
Figure 8.1 Bauxite Mine Distribution System
163 | P a ge
|
Virginia Tech
|
Figure 8.2 Intertie Monitoring
The monitoring of the intertie between two generator buses is crucial to detect electrical
islanding and other unwanted power issues. Typical algorithms for islanding detection use
values such as rates of change of frequency, voltage, and current. The client uses the capabilities
of the IEC 61850 implementation to gather information, run calculation logic, and execute the
proper load shedding decision in an acceptable, long-time-constant scenario. A diagram of the
load shedding system can be seen in Figure 8.2.
With the ability to connect the electric distribution system to the process control system,
load shedding at the automation controller is realized. Measurands from various IEDs can be
configured to be sent on various time intervals to the controllers. Controllers can run simple
algorithms and make correct load shedding decisions, either over hard wired outputs or by
utilizing the remote bit‟s strategy for command and control over IEC 61850. Many load
shedding algorithms for magnitude protection are long-time constant algorithms. This means
that the time to propagate information from IED, through the system, to controller is on the order
of hundreds of milliseconds.
An example of data path propagation can be seen in Figure 8.3. By grouping parameters
of interest in a single report, the RPI of this Class 1 connection can be reduced to small RPI
times, from 10-20 milliseconds, in order to populate the controller data table.
164 | P a ge
|
Virginia Tech
|
The load shedding algorithm monitors frequency, voltage, and current magnitudes to
detect abnormal generator or motor behavior. These values do not themselves give good
indications of system performance, but the rates of change of these fundamental parameters are
used to detect abnormal and unwanted power system conditions. Since IEC 61850 provides the
timestamp with the associated measurand, a rolling window of data points (i.e. 5 frequencies and
times) is collected, and the first derivative is calculated and averaged over the window. Based on
this average calculated value, the automation controller can make a decision on whether to shed
load or take remedial action in order to provide a more stable operating point.
Controller
Network
Gateway Module Managed Switch
Managed Switch
Figure 8.3 Load Shedding Data Flow
Conclusions
In addition to monitoring various fundamental electrical measurands in a SCADA
application, the technology developed during this research was implemented in a bauxite mine in
order to provide critical load shedding. The system was interfaced to 350 IEDs distributed over
various sites that were separated by 25-30 kilometers. The load shedding algorithm collected
data from the IEC 61850 network and used the concept of rolling windows to calculate
fundamental changes in frequency, voltage, and current, in order to assess system health.
165 | P a ge
|
Virginia Tech
|
9. Chapter 9 - Conclusions and Future Work
Overall Conclusions
The IEC 61850 standard suite of protocols and methods provides a convenient method to
integrate power system IEDs into control system networks in order to provide greater electrical
system visibility for plant operators and engineers. Connecting these devices to the Supervisory
Control and Data Acquisition system through the gateway, as described in this dissertation, can
greatly enhance system command and control functionality.
This dissertation presented various SCADA standards and defined their benefits and
shortcomings in both generic comparisons and benefits to mining protection and automation
systems. The literature review of this document has shown that, with respect to interoperability
between vendors and a common naming convention, the IEC 61850 standard was the best
candidate to implement an automation and control solution that could interact with current
systems. It was also discussed that the IEC 61850 standard possessed the following
shortcomings that would need to be addressed by this solution: although the substation
configuration language (SCL) file defines how an IED communicates on a 61850 network, it
does not define configuration information for protection and control functions of each IED.
Each IED manufacturer has proprietary software and configuration tools used for
enabling and configuring various protection elements and control strategies. The IEC 61850
standard defines no methodology for designing communication-based assisted automation. In
order to develop a successful solution for the mining market, these two concerns needed to be
addressed within the conceptual design of the total solution, i.e., hardware, software, and
visualization.
The dissertation summarized the hardware gateway module developed to implement the
conceptual design. A key component to this design is a multi-threaded design to interface
between the IED IEC 61850 network and the process EtherNet/IP network. Each thread,
including the master control program, was defined and its basic functionality explained in
functional block diagram. The various types of data used in this project were then addressed,
followed by how data is packaged in packets to be sent to the process network via EtherNet/IP.
166 | P a ge
|
Virginia Tech
|
The concept of the tag database was then discussed, i.e., how it is utilized as a common
space of shared memory where both 61850 and EtherNet/IP drivers read and write various tags.
Also discussed was the use of semaphore tags in the database to avoid collisions in the tag
database.
This research presented software developments that provide a functional, user-friendly
solution to aid engineers in link the electric power system of a mine with its process control
system. The software developed to support this research had two components: software created
to interface with the gateway module and software created to move information to a process
historian. Software developed to interface with the gateway module reads CID files and allows
the user to map information that is required by the automation controller. After this information
is defined, these tags are mapped to corresponding EtherNet/IP tags according to the ODVA
standard. Once this mapping has occurred, the module configuration is downloaded to the
module and an Add-on instruction is generated so that the data stream can be interpreted by the
controller. Once the Add-on instruction is imported by the controller, information is correctly
parsed from the data stream and inserted in to the controller data table.
The historian software reads information from the Alarms and Events software database
and inserts the corresponding timestamp in post processing to the process historian within 15
microseconds. Once the data is in the automation controller from hardware and software
configurations, the remainder of this solution provides operators and engineers visual aids to help
enhance command and control of a mine process control system.
Visualization tools were also developed in this research. Items include faceplate
definition, human/machine interface discussion and definition, and data management solutions.
The goal of visualization was to represent an IED in logic and graphics to produce a virtual
faceplate that is familiar to users accustomed with a particular IED. This gives operators and
engineers the same look and feel experience they have had with the physical device. At the same
time, graphics were developed with various levels of security to allow only users with proper
credentials access to various command and control functions.
Additionally, the research solution adhered to various graphics standards for both power
and process control systems. The visualization trending functionality of the solution was then
discussed by providing an example of a trend from a process historian repository. Finally, an
167 | P a ge
|
Virginia Tech
|
example of a secondary selective main-tie-main scheme was presented for a processing facility.
The one-line diagram was converted to an HMI screen, and multiple faceplates were tied to
various instrumentation blocks in the diagram to enable remote monitoring and control of the
system.
Finally, testing and verification procedures conducted throughout this research were
presented and described. As this research‟s goal was to develop a mine monitoring solution
utilizing the IEC 61850 standard to link the electrical distribution system with the process control
system, it was determined that the best way to validate the work and functionality of the
proposed system was to benchmark performance test results against the conceptual design. For
milestones requiring more than just visual inspection, the experimental setup was defined, testing
procedure discussed, and experimental results discussed. It was determined that the proposed
solution meets all of the required specifications imposed by the conceptual design.
The proposed system developed in this research was then installed at a bauxite mine in
order to provide energy data to the process control system as well as provide data for critical load
shedding decisions.
Additionally, this dissertation presented how the structured IEC 61850 tag names were
maintained in the automation controller software, which provides greater transparency of the
data values and faster commissioning of a system. The protocols allow for the interconnection of
all IEDs and control systems at the industrial level, including mining, metals, pulp and paper,
semiconductor, oil and gas, and more. The faceplate solution described within this document
provides many advantages to process owners.
This solution is an all-encompassing plant-wide solution for not only the electrical
distribution system, but also for functions on the electrical process network. The solution
provides process owners with the ability to extend control down to the individual breaker or
contactor for functions such as load monitoring and load shedding. By incorporating a solution
that functions with a multitude of vendors, process owners can now easily manage their electrical
systems with a simple, standard network infrastructure.
168 | P a ge
|
Virginia Tech
|
Future Work Recommendations
The IEC 61850 standard provides a uniform method of communication among
multivendor intelligent electronic devices used for system monitoring, control, and protection.
Connecting these devices to the Supervisory Control and Data Acquisition system through the
gateway, as described in this dissertation, can greatly enhance system command and control
functionality. Benefits to the mining industry include improved system monitoring, improved
safety through remote control of power system components, reduction in downtime through the
reconstruction of the sequence of events leading to system failures, and the implementation of
demand management. Therefore, there are opportunities for future work in the development of
algorithms for achieving specific applications for monitoring and control of power systems.
These include
1. Power monitoring
2. Predictive maintenance
3. Demand side load management
4. Auto Transfer Switch (ATS) load shedding
5. Command and control of rotating machinery
6. Command and control of breakers
7. Interfacing energy measurements into speed regulator process models
8. Ventilation On Demand (VOD)
9. Fast motor bus transfer
10. Main-Tie-Main Distribution Schemes
11. Data Historian Applications
12. On Demand Topology Transitions
169 | P a ge
|
Virginia Tech
|
EPC- Engineering Procurement Construction
EtherNet/IP – Protocol owned by ODVA that specifies transfer of data for manufacturing control
applications over traditional 802.3 Ethernet.
EWS- Engineering Work Station; Graphical interface or terminal where information is provided
to engineers for calculations or operation
GOOSE- Generic Object Oriented Substation Event; Protocol defined by IEC 61850 for peer-to-
peer or one-to-many fast exchange of data
Global Object- images that will be repeated reused throughout an application; given global scope
so they can be referenced from any visualization screen
GUI-Graphical User Interface
I&C- Instrumentation and Control; usually applied to an engineering sector that defines,
specifies, and codes for process instrumentation inside and industrial process
ICD- IED Capability Description; A file containing all the information that can be published by
the IED
IEC – International Electrotechnic Commission; Governing body of standards based in Europe
IEC 61850- Standard developed by IEC for interfacing to Intelligent Electronic Devices
IED- Intelligent Electronic Device; e.g. relay, circuit breaker, meter
IEEE C37.1- A standard developed by IEEE that defines the performance of SCADA systems
177 | P a ge
|
Virginia Tech
|
Interface Node- A piece of software or hardware that qualifies data from the process control
system to be passed to the historian repository
ISA- Instrumentation Society of America
ISA 5.5- A standard developed by ISA which specifies schemes for graphics on process
instrumentation
IT- Information Technology
MAC- Media Access Control; Firmware address of a device defined for identification purposes
Main-Tie-Main- An electrical scheme where loads can be added or shed from the system in the
case of a fault so that loads can be prioritized, and the plant not completely shut down
MCP- Master Control Program; Program run on startup of the gateway module
MMS- Manufacturing Message Specification; Protocol defined by IEC 61850 for block transfer
of data from publisher to subscriber
ODVA- Open Device Vendor‟s Association; Organization that owns industrial protocols such as
EtherNet/IP
OSI-PI- A manufacturer of a popular process historian
OWS- Operator Work Station; Interface, usually graphical that will allow status feedback and
command and control of a process
PAC- Programmable Automation Controller
178 | P a ge
|
Virginia Tech
|
Persist- A database term used to “bind” data to a tag and populate it through the entire process
control system
PHY- Physical Interface; e.g. RJ45 Ethernet adapter
PLC- Programmable Logic Controller
PT-Potential Transformer; Analog sensor used to measure a presence of voltage on an electrical
system
PTP- Precision Time Protocol; Standard defined by IEEE 1588 v.2 which defines time
synchronization of devices over a standard 802.3 Ethernet media.
Rockwell Automation- A company dedicated to manufacturing PACs and other industrial
automation equipment
RPI-Requested Packet Interval; The rate at which information is shared across an EtherNet/IP
network, also used for timeout of certain devices
RS Logix 5000- PAC programming software developed by Rockwell Automation
SCADA- Supervisory Control and Data Acquisition; A scheme to gather data from distributed
sections of facilities
SCD- Substation Configuration Description; A file that defines all the IEDs on a network
SCL- Substation Configuration Language; XML scripted language defined by IEC 61850
SD- Secure Digital; A method for storing information on an external format
SEL- Schweitzer Engineering Labs; An IED manufacturer
179 | P a ge
|
Virginia Tech
|
Investigation of Coal Dust Remediation using a
Surfactant in an Aqueous Solution
Connor Brown
ABSTRACT
In addition to ventilation practices, the application of water via sprays is the
most economical and popular means of combating respirable dust in an
underground coal mine. Due to a noticeable increase in black lung among
coal miners and new dust regulations, surfactants or wetting agents have
been used to aid in dust suppression. The surfactant facilitates the wetting
process by lowering the surface tension and allowing the hydrophobic coal
dust to come into contact with the water.
One of the most straightforward and effective benchtop tests is a simple
wetting test. Although there are variations of this type of test, principle and
technique remain the same. A known amount of dust was placed on the
surface of a solution and the time it takes for all the dust to fall through the
interface would be the wetting rate. This investigation examined the specific
density of the bulk dust and concentration of a surfactant in solution and
their effects on the wetting rate. It was found that both factors were
significant in determining the wetting rate. It was seen that the surfactant
had a more significant effect on the dust which consisted mostly of coal
particle when compared to a dust with a higher non-coal mineral content.
Additionally, full-scale tests were conducted to determine the effect of the
surfactant at a constant concentration. During the field implementation, the
|
Virginia Tech
|
Investigation of Coal Dust Remediation using a
Surfactant in an Aqueous Solution
Connor Brown
GENERAL AUDIENCE ABSTRACT
People who work in mines are exposed to many dangers and illnesses. One
of the illnesses, which has in recent history resurged, is black lung. Black
lung is a disease caused by coal dust entering the lungs. The body’s reaction
to it is to build scar tissue around the piece of dust. If this happens enough
times over the miner’s career, then it becomes nearly impossible to breath.
Normally, to prevent this from happening, water is sprayed in the coal
before it is chipped off by the machine. Since this appears to no longer be
effective, soapy chemicals are added to the water, which helps to keep the
dust from lifting into the air in the first place.
One of the easiest ways to test whether the chemicals are working well or
not is to conduct a wetting test. When conducting a wetting test, a known
about of dust is placed on top of the water and chemical mixture, and the
time it takes for all of the dust to be wet is call the wetting rate. To get
better results in an actual mine, faster wetting rates were sought after. The
wetting test showed that the two main factors which determine the wetting
were how much coal is in the coal and rock dust mixture and how much
chemical is used. It was seen that the chemical had a more significant effect
on the dust which had mostly of coal particle when compared to dust with
more rock dust.
|
Virginia Tech
|
Another study was conducted at a mine with only one mixture of water and
chemical. During the study, the chemical was pumped through the mine and
to the cutter heads of the continuous miner. A continuous miner is the name
of the equipment used to mine coal and other soft material. The cutter head
is the piece of the equipment which actually makes contact with coal. Since
the conditions at the mine were not ideal and not enough data was taken, the
resulting effect of the chemical could not be certain. More long-term studies
need to be done in the future to help account for the less than ideal
conditions. There were, however, larger amounts of dust when using new
sampling equipment as opposed to the older equipment given the same
conditions. Also, smaller amounts of dust were seen when the miner
operators were allowed to activate the air cleaning attachments on the
continuous miner. These issues should be revisited in the future.
|
Virginia Tech
|
1 INTRODUCTION
There has been a recent surge of concern in the mining community on the
topic of occupational health. One of the areas of intrigue and interest
revolves around exposure of mine workers to respirable dust. Specifically,
the exposure of coal workers to respirable dust, coal, and silica-based dust.
Respirable dust in coal mines has been an issue for a number of years.
Efforts throughout the last half-century have greatly decreased the number
of cases of black lung to record lows in the early 2000’s. That being the
case, a recent uptick in the number of cases has been noticed. There are
many efforts under way to both identify the source or sources of the recent
increase and to regulate its influence on the health of coal workers. The
regulations which will be instated puts more pressure on the producers to
prevent and control hazards through better safe work practices or
administrative and engineering controls. A possible use of an engineering
control will be investigated in this paper. One of the more popular
engineering controls for abating respirable dust is the use of sprays during
the production and transport of ores. The continuous demand for energy,
driven by the rising populations, has forced mining companies to produce in
more difficult and thinner seams. The thinning of the seam has led to the
coal producers mining more rock than in previous operations. These
pressures on the industry have not only changed the mineralogy of the dust,
but the tonnages and raw production rates necessary have stretched the
effectiveness of sprays to their limits. There have been many attempts to
enlist the aid of surfactant in mitigating the exposure of underground coal
miners to respirable dust, and there have been mixed results both from lab
and field testing. The surfactant or wetting agent reduces the surface tension
1
|
Virginia Tech
|
2 LITERATURE REVIEW
Increasing population leads to increased demand for natural resources. The
demand leads to increased mining activity. Coupled with the decreasing
quality of ores and thinning of seams, increasing tonnages of excess rock
becomes an inescapable part of the mining process. For the employees
working in the mines, the dust becomes more difficult to control, resulting in
increased exposure. Specifically, this paper will discuss what Progressive
Massive Fibrosis (PMF), silicosis, and black lung are and how exposure to
coal and silica dust causes the diseases. This thesis will also examine
methods used to mitigate exposure to the dust, such as water sprays, as well
as test a surfactant in both a laboratory and a field environment.
This research was spurred by a noticeable uptick in the number of cases of
lung diseases among coal miners. Underground coal miners’ occupational
hazards, as they relate to mortality, have been extensively studied. Coal
mine dust has been found to be one of the most dangerous of these
occupational hazards due in part to its ability to cause “Miners’ Black
Lung”, or Coal Worker’s Pneumoconiosis (CWP) and silicosis. CWP and
silicosis are chronic occupational lung diseases caused by long-term
exposure to respirable dust (particles with 100% passing through a 10-
micron screen). After coming in contact with the alveoli, the dust triggers
inflammation, eventually resulting in fibrosis and irreversible lung damage.
(Laney et al., 2012) It has been found that the particles in the two to six-
micron range can be most damaging (Méndez-Vargas et al., 2013). CWP
and silicosis are both diagnosed by examining a chest X-ray. The
radiographic test will show opacities and will be classified by the size,
3
|
Virginia Tech
|
shape, and extent. The two diseases are only differentiated by the
individual's work history (Neimeier, 1993).
2.1 COAL WORKERS PNEUMOCONIOSIS
Some of the earliest references and investigations, including post-mortem
examinations, into what we know now as black lung took place in the early
1800’s. The term pneumoconiosis or “dusty lung” was not used until 1874
During this time, the tools used to study the disease were an examination of
the lung tissue and societal correlations (Meiklejohn, 1951). Around 1907,
an etiological investigation into the disease began in earnest with the use of
chest x-rays. The x-rays at this time, however, lacked the resolution for
early detection and could only be used for detection of very significant
pathological differences (Neimeier, 1993). It was not until the 1930’s-1950
that the investigators studying the disease had a wider range of tools and
innovative techniques including radiology, biochemistry pathology, and dust
sampling and analysis (Meiklejohn, 1952). In addition, for a number of
years, the causative agent was thought to be solely silica exposure. This
assumption, however, was brought into question after investigations found
pneumoconiosis among coal trimmers in the United Kingdom. The
occupational responsibility of the coal trimmer was to load and distribute
coal, which had been previously washed and separated from rock, into the
hold of ships (Collis & Gilchrist, 1928). Although the conclusions were
contested in the 1930’s, additional support was received in 1940 (Gough,
1940). Also, around this time, a technique of preemptive diagnosis, and
assessment of the disease were attempted, but the accuracy remained low
(Gough et al., 1949). Correlation between the total dust and the severity of
the pneumoconiosis was later found by analyzing radiographic and
4
|
Virginia Tech
|
pathologic data (King et al., 1956). There were some investigations
conducted in the US in the earlier part of the 20th century which also
mirrored the sentiments of British researchers, and brought forth evidence
that the lung disease among coal miners must be differentiated from the
broad label of silicosis (Dreessen & Jones, 1936). It was not until around
the late 1950’s, and 1960’s that more emphasis was placed on studying the
disease and finding an adequate method of reporting its incidence (Doyle et
al., 1958). In the US, a number of studies were conducted in Pennsylvania, a
region with a higher reported frequency of the disease (Baier & Diakun,
1961; Lieben & McBride, 1963; Mc Bride et al., 1963; McBride et al., 1966;
Tokuhata et al., 1970). Later studies looked at the larger Appalachian region
and other coal fields throughout the US (Lainhart, 1969; Morgan et al.,
1973; R. L. Naeye & Dellinger, 1970). Experts in the earlier part of the
1970’s began proposing that black lung was an agglomeration of complex
disorders differing in severity and frequency, all of which depended on
aggregate exposure, and individual predisposition based on personal habits
and exposure to pollutants in their respective communities (R. Naeye &
Dellinger, 1972).
The medical definition of CWP is a parenchymal lung disease resulting from
the body’s response to the deposited and retained coal dust in the lung
(Weeks & Wagner, 1986). The current legal definition of pneumoconiosis is
“a chronic dust disease of the lung and its sequelae, including respiratory
and pulmonary impairments, arising out of coal mine employment.” From
20 CFR 718.201. While the coal dust is predominately comprised of carbon,
the dust also has trace metals and inorganic minerals in its composition
which can be cytotoxic (Castranova & Vallyathan, 2000; Huang et al.,
2006). It has also been found and confirmed over the years that there is an
5
|
Virginia Tech
|
increased risk of contracting CWP with an increase in coal rank (M. D.
Attfield & Kuempel, 2008; M. Attfield & Morring, 1992; Morgan et al.,
1973). Additionally, the large surface area of the coal dust resulting from
the small size has the ability to absorb aromatic compounds such as benzene
and phenol present in the mine atmosphere. These compounds may have an
adverse effect on the lung tissue (Castranova & Vallyathan, 2000). In order
to start developing CWP an individual could normally have ten or more
years of exposure to respirable coal dust. The radiographic test will show
opacities <10mm (Neimeier, 1993) normally in the upper chest area
(Castranova & Vallyathan, 2000). An example of one of these x-rays can be
seen in Figure 1 below. Although the x-ray may show these opacities, the
individual may not be suffering from any symptoms (Neimeier, 1993;
Colinet et al., 2010). Once diagnosed, the individual is at a greater risk for
complicated CWP or PMF. When the opacities found on the x-ray combine
to cover an area greater than 1 cm then, the disease has progressed to the
point of complicated CWP. This can be seen in Figure 2 below. It is not
necessary for the individual to be diagnosed with simple CWP prior to this
diagnosis (Neimeier, 1993).
6
|
Virginia Tech
|
Figure 1: Simple CWP (Colinet, 2010) Figure 2: Complicated CWP (Colinet, 2010)
2.2 SILICOSIS
When respirable crystalline silica is deposited in the alveoli, it is possible for
silicosis to develop. This is due to the high reactivity of the cell membranes
and the surface of the crystalline silica. The silicon dioxides (SiO ) come in
2
contact with water and form silanol (-SiOH) which are hydrogen donors
(Castranova & Vallyathan, 2000). Hydrogen bonds are then formed due to
lone pair electrons on the oxygen and nitrogen that make up many biological
macromolecules. These bonds can lead to unfavorable interactions and cell
damage, ultimately developing into silicosis (Castranova & Vallyathan,
2000). This disease has four categories, chronic, complicated, accelerated,
or acute (Neimeier, 1993). Exposure to the respirable crystalline silica for
15 for more years can result in the contraction of chronic silicosis. The
silicotic nodule is the telltale feature. It is made up of an amorphous center
of a fibrous tissue surrounded by systematic hyalinized collagen fibers,
sometimes referred to as onion skinning (Castranova & Vallyathan, 2000;
Neimeier, 1993). An example of this can be seen in Figure 3. As with the
case of simple CWP, the individuals does not need to show any symptoms,
and the nodules will show up on the x-ray as opacities covering an area of
7
|
Virginia Tech
|
required each operator in a coal mine to continuously maintain the average
concentration of respirable dust in the mine atmosphere during each shift to
which each miner in the active workings of such mine is exposed at or below
2.0 milligrams of respirable dust per cubic meter of air (Congress, 1977).
Proclamation of the Mine Act occurred on April 8, 1980, and is currently
implemented under 30 CFR § 70.100, Respirable dust standards. In the case
where more than 5 percent silica is found in the respirable dust, then the
limit is determined using the following formula from 30 CFR § 71.101:
10
ɸ = (1.1)
% 𝑠𝑖𝑙𝑖𝑐𝑎
Where Φ = respirable dust limit (mg/m3)
% Silica = percent silica found in dust as a fraction
Since 1980, average coal dust exposures and incidence of CWP have
declined under the existing standards (National Mining Association, 2013),
but recently, “CWP has increased among experienced miners, and in some
cases, CWP has progressed rapidly to PMF” (Department of Labor - Mine
Safety and Health Administration, 2010). Additionally, it was found that
there is a geographic concentration of CWP cases in central Appalachia as
shown in Figure 5.
10
|
Virginia Tech
|
miners may be exposed to more toxic levels of respirable silica than in
previous years due to more difficult geology (Laney et al., 2010).
For this reason, in October of 2010, the Mine Safety and Health
Administration (MSHA) put forth a proposal to limit exposure further,
changing the limit from 2.0 mg/m3 1.0 mg/m3. Grounds for the proposal
were primarily based on a National Institute for Occupational Safety and
Health (NIOSH) report submitted to the Secretary of Labor Lynn Martin, on
November 7, 1995. The NIOSH report recommended that “exposures to
respirable coal mine dust be limited to 1.0 milligrams per cubic meter as a
time-weighted average” (NIOSH, 1995). The final ruling is a
comprehensive method which utilizes new continuous personal dust monitor
(CPDM) technology to display real-time dust concentrations, as well as
increased sampling and stricter standards. The final rule has gone into effect
as of August 1, 2014, with sections being introduced during the following
two years (Department of Labor - Mine Safty and Health Administration,
2014).
2.5 SPRAYS
Most of the coal dust production in a mine occurs at the face during cutting
and has the potential to produce up to 8600 grams of respirable dust per ton
of coal (Tien & Kim, 1997). In order to meet the new standards and be
proactive in the protection of miners from respirable dust, there are a number
of methods that can be implemented. The most predominate methods
include are the introduction of engineering controls. The engineering
controls used to limit the miners’ exposure to dust can vary from improved
ventilation controls to the introduction of scrubbers and collectors. One of
12
|
Virginia Tech
|
the most popular engineering controls used to curb the dust problem has
been the application of water via sprays in areas where there is significant
dust production.
There are three functions sprays play in limiting the exposure of a coal miner
to coal dust particulates. There are also a few factors regarding the
effectiveness of water spray systems depending in the function of the sprays;
including the nozzle type, pattern, flow, location, and pressure (Colinet et
al., 2010). The first function, redirection, can aid in separating the miner
from the primary source of the dust, by moving the direction of air flow
(Pollock & Organiscak, 2007). The key characteristics for these sprays are
high pressure and strategic location. The second function is capture of dust
particles, which has been demonstrated (Cheng, 1973; Goodman et al.,
2006). When capturing airborne dust with sprays, water droplet size and
velocity are the most influential parameters. Lastly is suppression and
wetting where low pressure high flowrates are the most advantageous
characteristics. The function of wetting coal is also thought to be the
primary mechanism for dust suppression, while capture and redirection are
secondary effects (Kissell, 2003).
There are a number of spray designs, each of which has a specific
application. Figure 6 shows the types of sprays using in mines to control
and mitigate dust.
13
|
Virginia Tech
|
Spraying Systems, 2013; Colinet et al., 2010). Hollow cone nozzles are
primarily used for combating the entrainment of dust and redirection of air
flow. These are also the most common type of spray placed behind the
cutting head on the boom of a continuous miner. In addition, the large
openings help to keep the nozzle from clogging and allow for higher flow
rates needed for wetting the host rock. The atomizing and fine spray nozzles
are used less in the industry due to their higher price and high maintenance
requirements (Colinet et al., 2010). The effectiveness of each type of spray
nozzle in capturing airborne dust can be seen in Figure 7.
Figure 7: Relative spray nozzle effectiveness (J Colinet et al., 2010)
2.6 SURFACE TENSION
Simple water sprays have now reached their useful limits in suppressing
respirable dust due to the physical parameters, and new effective suppression
aids are needed. The wettability of coal is limited due to its hydrophobic
nature and ultimately the surface tension of the water.
15
|
Virginia Tech
|
Surface tension is the phenomenon which results due to the cohesive force
between the water molecules. The cohesion arises from the hydrogen bonds
which are due to the polar nature of water molecules (Figure 8). The
cohesive forces within the bulk of the liquid are essentially uniform and
shared by all neighboring molecules. The exceptions are the molecules on
the surface. On the surface, since there are no molecules above, a film of
stronger attractive forces are formed.
Figure 8: Cohesion due to polarity (Tien & Kim, 1997)
These stronger attractive forces resist any distortion or increase of surface
area and act as an elastic membrane. This is the reason for the round shape
of water droplets or of air bubbles (Silberberg & Weberg, 2009). The round
shape allows for the least amount of surface area per volume and is the most
stable arrangement. Any distortion or increase in surface area will take
energy. The higher the surface tension, the more energy it takes to create a
new surface. The units of measurement are N/m, which is a force per unit
length. The factor with the largest effect on the amount of surface tension is
the difference in composition between the bulk and the surface. This
relationship is explained extensively in the work produced by J. W Gibbs. A
portion of that work is shown in the equation below (Gibbs, 1878).
−𝑑𝛾 = Γ 𝑑𝜇 + Γ 𝑑𝜇 (1.2)
1 1 2 2
16
|
Virginia Tech
|
Where 𝛾 is the surface tension in newton per meter, Γ is the surface excess of
a given component in mole per meter squared and 𝜇 is the chemical potential
of a given component in joules per mole. An example of this can be found
by examining the surface tensions of fresh water and salt water. The
freshwater will typically have a surface tension of around 73 dynes/cm while
the saltwater will be around 78 dynes/cm (Stewart, 2009). The increase in
surface tension occurs because on average the surface of the water is less
salty than the bulk, or in other words, the salt has a negative surface excess.
In this case, if a new surface is to be created, the molecules of salt within the
new surface layer must be forced back into the bulk (Chaplin, 2009). If
more salt is added to the solution and the potential is raised, it becomes
harder to push the molecules back into the bulk. As a result, the surface
tension is increased. In this case, the surface excess of the salt component is
negative, and any increase in the potential of the salt component leads to the
increase in the surface tension of that solution (Nalwa, 2001). The inverse is
also true. If a solution of soap and water were to be examined there would
be an excess of soap molecules on the surface of the solution and a lesser
concentration of soap in the bulk. In this case, the soap has a positive
surface excess, so when a new surface is formed soap molecules must be
taken from the bulk and placed in the new surface layer. If more soap is
added to the solution, increasing the potential, then the abundance of
molecules in the bulk are easily transferred to the surface. The fact that the
soap molecules are readily available makes the process easier and less
energy intensive, resulting in a lower surface tension. In this case, the
surface excess of the soap component is positive, and any increase in the
potential of the soap component leads to the decrease in the surface tension
of that solution (Nalwa, 2001).
17
|
Virginia Tech
|
with the hydrophobic sites, and the hydrophilic head group will be at the
interface with the water. This arrangement allows the coal surface to have
peso-hydrophilic properties.
2.7.1 Past Work
There have been many studies which have not reached a consensus on how
well surfactants improve dust capture or by what mechanism. In one study
conducted by Goldshmid and Calvert, the conclusion was that the interfacial
tension was the primary mechanism by which the collection efficiency of
water droplet could be greatly increased (Goldshmid & Calvert, 1963). On
the other hand, findings from Brauer and Varma suggested that the wetting
characteristics rather than interfacial tension, were the primary mechanism
(Brauer & Varma, 1981). There have also been observations where the
results were inconclusive, and the use of surfactants for the purpose of dust
control are not justified (Hargraves & McKinnon, 1961). Another study
found that surfactants are useful or effective in high dust concentration
areas, such as the face while cutting (Chander et al., 1991). Additionally, this
study by Chandler et al. found that the decrease in droplet diameter
increased collection efficacy. A study done by Tien and Kim found that it
was better wettability and not droplet parameters which lead to higher
collection efficiencies (Tien & Kim, 1997), and they also determined that
nonionic surfactants were found to have best collection efficiency (Tien &
Kim, 1997). Conversely, a study by Hu, Polat and Chandler observed that a
nonionic surfactant had a narrow maximum peak, and that cationic
surfactant had both a broad operating range and the highest collection
efficiency (Hu et al., 1992).
19
|
Virginia Tech
|
3 ASSESSMENT OF SURFACTANT VIABILITY FOR DUST
WETTING VIA SINK TEST
3.1 ABSTRACT
In addition to ventilation practices, the application of water via sprays is the
most economical and popular means of combating respirable dust in an
underground coal mine. Due to a noticeable increase in black lung among
coal miners and new dust regulations, the use of surfactants or wetting
agents have been to aid in dust suppression. The surfactant facilitates the
wetting process by lowering the surface tension and allowing the
hydrophobic coal dust to come into contact with the water. One of the most
straightforward and effective benchtop tests is a simple wetting test.
Although there are variations of this type of test, principle and technique
remain the same. In this study, a known amount of dust was placed on the
surface of a solution and the time it takes for all the dust to fall through the
interface was determined as the wetting rate. This investigation examined
the specific density of the bulk dust and concentration of a surfactant in
solution and their effects on the wetting rate. It was found that both factors
were significant in determining the wetting rate. It was seen that the
surfactant had a more significant effect on the dust, which consisted mostly
of coal particle when compared to a dust with a higher non-coal mineral
content.
3.2 INTRODUCTION
There has been a recent surge of concern in the mining community on the
topic of occupational health, specifically, the exposure of coal workers to the
21
|
Virginia Tech
|
respirable dust, including coal, and silica-based dust. Respirable dust in coal
mines has been seen as an issue for a number of years. Efforts throughout
the last half-century have greatly decreased the number of cases of black
lung to record lows in the early 2000, however, a recent uptick in the
number of cases has been noticed (Blackley et al. 2014). There are many
efforts under way to both identify the source or sources of the recent
increase and to regulate its influence on the health of coal workers. The
regulations instated put more pressure on coal producers to prevent and
control hazards, whether that be through better safe work practices or
administrative and engineering control. Advances in engineering controls
such as ventilation practices, remote control equipment, dust scrubbers, and
water spray systems have reduced the exposure of miners to respirable dust.
The use of water sprays as an engineering control for abating respirable dust
is common during the production and transport of ores. The mechanism for
abating dust is twofold; the airborne dust capture and the preemptive wetting
of the coal or rock, the latter of which is the primary means (Kissel, 2003).
The continuous demand for energy, driven by the rising populations, has
forced mining companies to produce in more difficult and thinner seams.
The thinning of the seam has led to the coal producers mining more of the
host strata than in previous operations. These pressures on the industry have
not only changed the mineralogy of the dust, but the sheer tonnages and rate
of raw material being moved have stretched the effectiveness of water sprays
to their limits. There been many attempts to enlist the aid of surfactants to
mitigate the exposure to respirable dust, with mixed results (Copeland, 2007;
Hu et al., 1992; Kost et al., 1980; Organiscak, 2014; Zeller, 1983). This
paper will consider the practicality of using a surfactant to aid in wetting
coal dust with differing mineral content using a variation of the Walker Sink
22
|
Virginia Tech
|
Test to gather preliminary data. Wetting rates have been used numerous
times in investigations to examine surfactant effectiveness to wet dust
(Chander et al., 2007; Copeland & Eisele, 2008; Copeland, 2007; Feldstein,
1981; Glanville & Haley, 1983; Glanville & Wightman, 1979; Kawatra,
2006; Kim & Tien, 1993; Tien & Kim, 1997; Zeller, 1983).
3.3 MATERIALS & METHOD
3.3.1 Description of study
The following sink test was conducted to determine the effectiveness of a
surfactant to improve the wettability of coal dust. Additionally, the study
seeks to determine how the mineral content affects the wettability of the
dust. Due to the occupational hazards of respirable dust when coupled with
the recent increase in the number of cases of lung disease among mine
employees, a practical surfactant which reduces the amount of respirable
dust which a mine employee is exposed to is needed.
The laboratory test is first needed in order to determine the viability for the
surfactant before running any experiment in the field. For this purpose, the
sink test was used. The following sink test experiment uses an analytical
balance to find the apparent weight of the wetted dust at predetermined time
intervals. By examining the changes in weight of the wetted dust over time,
a wetting rate can be ascertained. The subsequent section will describe in
detail the initial efforts to observe the effects of varying concentrations of a
surfactant, and the impact of mineral content on the wetting rates.
23
|
Virginia Tech
|
3.3.2 Experimental design
3.3.2.1 Equipment
For this experiment, the most important piece of equipment was the
analytical balance. The analytical balance used was an Ohaus Pioneer
PA214. This scale was chosen because it had the necessary capacity to hang
weights from the underside of the scale. It also had the added benefit of
being able to link to a laptop via a USB2.0/RS232 serial cable. With these
attributes, the scale could be programmed to send a measurement at
predetermined time intervals. A computer was used as the data storage and
data acquisition unit. The software used to control and manage the scale’s
output was WinWedge. A wire cable was used to connect a plastic platform
and counterweights to the underside of the analytical balance, and a small
glass basin was used to hold the solution.
3.3.2.2 Equipment setup
The analytical balance was set and suspended above a glass basin. Tied to
the underside of the analytical balance, via a wire, were the platform and the
counter weights. When the basin is filled with water or the solution, the
platform is submerged. The platform was cut to fit the basin with about one
to two millimeters of space along the edge. The RS232 end of the cable was
connected to the analytical balance, and the USB end was connected to a
computer. The computer had previously been loaded with WinWedge
software for data acquisition. The basic equipment setup can be seen in
Figure 10 below.
24
|
Virginia Tech
|
74.00 98.73 1.260 5.500 28.02 3.730 0.409 0.8000 0.4300
62.23 97.47 2.230 4.625 24.29 3.270 0.344 0.370 0.3700
52.33 95.24 3.730 3.889 21.02 2.760 0.289 0.0000 0.0000
44.00 91.51 5.490 3.270 18.26 2.340 0.243 0.0000 0.0000
37.00 86.02 6.910 2.750 15.92 2.060 0.204 0.0000 0.0000
31.11 79.11 7.280 2.312 13.86 1.890
After the screening, these samples were then oven-dried and collected in
plastic sample bags. Table 2 below shows the different gravity classes that
would be tested and the estimations for their respective mineral content.
Table 2: Tabulation of gravity class and estimated mineral content
Gravity class Mineral content
1.3 Float 0%
1.3 - 1.35 0 - 4%
1.35 - 1.4 4 - 7%
1.4 - 1.45 7 - 11%
1.45 - 1.5 11 - 15%
1.5 - 1.6 15 - 22%
1.6 - 1.7 22 - 30%
1.7 - 1.8 30 - 37%
1.8 Sink > 37 %
The mineral content was estimated by assuming the specific gravity of coal
and non-coal or mineral were 1.3 and 2.65 respectively.
The surfactant being used was initially found to be too viscous to be drawn
with a micropipette. For this reason, the concentration was diluted mixing
equal parts by volume of deionized water and surfactant. The surfactant was
measured via a 100 ml syringe. A large syringe was used because of its
large ratio of volume to surface area. There was concern that the viscosity
of the surfactant may lead to an excess of material adhering to the inside of
the nozzle of the measuring device. The surfactant was added to the
container first. Next, the then the deionized water was added with a separate
100 ml syringe. Using the solution, the syringe that was used to measure the
surfactant was then rinsed to free any surfactant that was still adhering to the
27
|
Virginia Tech
|
inner surface of the syringe. The tested ratios are tabulated in Table 3 below
along with their respective volumetric percent of the solution.
Table 3: Tabulation of tested surfactant to DI water ratios
Ratio of Surfactant as a
surfactant to DI percent of solution
water
0 0.000%
1:7000 0.014%
1:6000 0.017%
1:5000 0.020%
1:4000 0.025%
1:3000 0.033%
1:2000 0.050%
1:1000 0.100%
1:500 0.200%
3.3.3 Procedure
Initially, the suspended analytical balance was leveled. After randomizing
the order of sample combinations to be tested, the concentration of the
solution to be tested was prepared. This was accomplished by mixing 300
ml of deionized water with the appropriate amount of diluted surfactant.
The deionized water was measured 50 ml at a time with a 60 ml syringe and
place in a 500 ml beaker. The surfactant was measured with a micropipette.
The prepared solution or unaltered deionized water was then poured into the
basin. Next, the platform and weights were placed in the solution and the
basin centered beneath the analytical balance. Due to the movement
involved, the system was allowed to settle before taring the analytical
balance. In this time, the dust sample was prepared. The dust sample to be
tested was weighed in a separate analytical balance. A dust sample of at
least 500 mg, but no greater than 510 mg was taken from the sample bags.
Due to the drying process, some of the dust formed cakes which were
broken up. After breaking up the cakes, the dust was placed on a spatula,
which would be used to place the dust on the surface of the liquid. In
28
|
Virginia Tech
|
combination with the spatula, a sieve was used in order to disperse the dust
evenly on the surface of the solution. After the analytical balance had
settled, it was zeroed, and the software was set to record the incoming data.
After the first entry in excel appeared, the dust was distributed on the surface
of the liquid. The data were recorded as long as the dust continued to settle
on the submerged platform. The test was halted only when the recorded data
appeared to reach steady state, and it was observed that the dust would no
longer wet. After the test was halted the platform and the counter weights
were removed from the wire. Lastly, the platform, basin, and beaker were
all washed and rinsed thoroughly with deionized water a total of three times
and then allowed to dry.
3.4 RESULTS AND ANALYSIS
3.4.1 Analysis
The results that were recorded yielded some formatting errors. Before the
results could be analyzed the errors had to be removed. In some cases, a
negative sign would be incorporated into the data, not representative of any
physical phenomena, and some entries had a misplaced decimal point or had
multiple decimal points with the entry. Finally, the data that were used had
to be corrected due to displacement to depict the actual weight of the dust.
In its recorded form, the data were showing the apparent weight of the dust
in the solution. The first steps began with correcting the apparent weight to
reflect the actual weight of the dust. This was achieved for each gravity
class by using Archimedes’ principle to find a correction factor. The
correction factor was found, assuming the gravity of the dust in each gravity
class was the arithmetic mean of the limits for each class. Next, the problem
29
|
Virginia Tech
|
of negative entries and misplaced decimals, which led to entries having
values an order of magnitude larger than they should have been, were
corrected. This was done by setting limits of zero and 500 mg around the
recorded data set. It is important to note that errors remained in the data in
some cases. (An entry of 320 milligrams when it should have been 32 mg or
vice versa would remain unaffected.) When the corrections were made the
data entries were replaced with blanks as placeholders in order to not skew
the data and to ensure the data points remained at the right time tag.
After the data were refined, a model was selected. The initial form of the
data has two parts. The first is the settling of the majority of the dust,
approximately 0 to 4000 seconds in Figure 13. The shallower incline of the
second section of the plot describes the slower settling times of the very
small particle and may be due to the mixing effect caused by the settling of
the larger particles of dust, approximately 4000 to 30000 seconds in Figure
13. In order to find the wetting rate, the model used a piecewise function
with the first part being a linear function and the second part being a power
function. Although there were models that fit the data better, this one was
chosen because the wetting rates could be determined directly from the
model. The slope of the first part of the piecewise function is considered the
wetting rate of the sample. The model was initially rough fitted by manually
altering the constants in the model. On a point to point basis, the model was
compared to the recorded data. The difference from each point was squared,
and these values were then summed. It was this sum which was minimized
by fine tuning the constants of the model. An example of the result is shown
in Figure 13.
30
|
Virginia Tech
|
Figure 14: Recorded data and fitted model for gravity class 1.3 - 1.35 with a surfactant ratio of 1:5000 after alteration
The most common problem in all models was accurately depicting the
beginning of the experiment, possibly due to the manual deposition of dust
onto the surface of the solution. The wetting rates of some dust might
appear higher in the beginning if the particles were agglomerated or released
from a higher position than other samples. Additionally, the higher wetting
rates may be a result of higher than normal deposition of dust directly on the
connecting wire or hook of the experimental setup. In other cases, the initial
measurements are lower than expected or negative, shifting the y-intercept
of the model below zero. This may be a result of the data transfer or
interpretation between the analytical balance and the computer. These
depiction problems can be seen very well in the Observation versus
Predicted (OvP) plots for each model. Examples of these OvP plots for
both the higher and lower than expected wetting rates can be seen in Figure
15 and Figure 16 respectively.
32
|
Virginia Tech
|
significant difference in the wetting rates which correlate with both the
surfactant dosage and the gravity class. There is also a less considerable, but
still significant, interaction between the two variables.
After conducting the initial experiments, additional data points were sought.
In order to get a better picture of the correlation between the wetting rates to
the gravity class and surfactant, additional parameters were tested. The
same procedures were followed. The Observed vs. Predicted plots were also
used to aid in the decision process of removing anomalous data. In addition,
the coefficients of determination had to be greater than 0.95. The results are
shown in Figure 21.
Figure 21: Wetting rates resulting from dosage ratio
Figure 21 shows a general trend of increased wetting rates with increased
silica content and dosage ratio. Wetting rates are also seen to increase at a
slower rate at the higher concentrations of 1:500 and 1:1000. At the high
dosage ratios and high gravity, the dust applied to the surface of the water
38
|
Virginia Tech
|
immediately fell through, leading to the wetting rates which are highly
dependent on application time. When examining dosage ratios, the largest
increases are seen when a small amount of surfactant was added and
compared to the runs with no surfactant at all. Also, the effect of the
surfactant on the different gravity classes is evident. Figure 22 below
compares the wetting rates of two high gravity dusts and two low gravity
dusts.
Figure 22: Wetting rates of high and low gravity dust
It can be seen in Figure 22 that the wetting rates from the high gravity dust
increase about an order of magnitude from no surfactant to a 0.03%
surfactant solution. At the same time, the lower gravity dust increases
almost three orders of magnitude. This noticeable difference may be due to
the high gravity coal already being close to a maximum wetting rate as a
result of being more hydrophilic. The hydrophilic characteristic of the dust
does not allow for a significant improvements to be made.
39
|
Virginia Tech
|
deposition via compressed air or the like. Such a device should also bring
about more consistency and even surface coverage with less agglomeration
of the dust particles. This burst mechanism along with tests run in triplicate
would also aid to make sure the regular oscillations were not skewing the
results. On the same note, additional improvements could be made to the
test apparatus with regards to isolation. Due to the longer period of the
regular oscillations, the source is most likely mechanical or environmental.
Additional support under the analytical balance would help in removing
some of the higher frequency oscillations. The very low-frequency
oscillations may be due to some environmental interference. A small
enclosure around the apparatus would aid to isolate the tests from these
interferences.
Additionally, scanning electron microscopy (SEM) analysis of the dust
would also be beneficial in determining the true mineral content. The use of
SEM in determining the mineral content of the wetted dust would also be
very informative. Taking samples at various points during the wetting
process would also help to determine whether the non-coal fraction is
wetting before the coal fraction due to the high degree of liberation.
Under laboratory conditions, it has been shown that the surfactant is an
effective additive for wetting dust. Until this point, no studies have been
conducted in order to test the effectiveness of the surfactant at reducing the
amount of airborne dust. As a result, the potential to take the surfactant to a
more dynamic environment to examine the effects of a sprayer type
application, either in a lab on in the field, does exist.
41
|
Virginia Tech
|
4 EVALUATION OF SPRAY WATER WITH THE ADDITION
OF SURFACTANT FOR THE PURPOSE OF RESPIRABLE
COAL DUST SUPPRESSION
4.1 ABSTRACT
The application of water sprays is one of the most popular means of
suppressing respirable dust in an underground coal mine. Due to a recent
spike in the number of cases of black lung among coal miners, the use of
surfactants has been investigated to aid in dust wetting. The surfactant
lowers the surface tension thereby allowing the hydrophobic coal dust to
come into contact with the water more easily. Full-scale field testing is
needed to assess the effectiveness of surfactant in an underground coal
environment. During the field implementation, the surfactant was pumped
into the section mine water to the cutter heads of the continuous miner. The
study implemented continuous personal dust monitors and personal dust
samplers, which were placed in the mine environment to monitor the dust
concentrations. Measurements were taken when the surfactant was off and
on. When comparing the level of dust concentrations, no significant
differences were found. There was, however, a significant difference as a
result of location or distance from the face. Additionally, the volumetric air
flow showed significant impact reinforcing the use of ventilation as a
primary engineering control. Interestingly, the dust measurement between
the two devices and activation of scrubbers were significantly impactful.
44
|
Virginia Tech
|
4.2 INTRODUCTION
The continuous demand for energy, driven by a rising population and
subsequent increased demand for energy has forced many coal mines to
produce in thinner and more difficult seams. These difficult conditions have
led to an increase of host rock being mined in the process. In addition,
several stakeholders have pointed out the increase of occupational lung
disease among coal miners in recent years (Blackley et al., 2014). Some
effort seeks to lessen the impact of these diseases through regulation. These
new regulations, along with market pressure have placed more onus on the
coal producers to implement better engineering controls. The most popular
method is the better implementation of ventilation which dilutes and
removes the dust from the mine atmosphere. The next most cost effective
and popular method is the use of water sprays for preemptive wetting of the
coal before cutting and removal of dust from air near the source. The
pressures felt by the mining community have stretched the practical
application of water sprays to their limit. The method which is investigated
in this paper is the addition of a surfactant into the spray water. The
surfactant lowers the surface tension of water allowing for easier wetting of
coal at the interface. The aid of surfactants to mitigating the exposure to
respirable dust has been investigated in the past, but with mixed results for
both laboratory and field testing (Copeland, 2007; Hargraves & McKinnon,
1961; Hu et al., 1992; Kost et al., 1980; Organiscak, 2014; Zeller, 1983).
This paper will consider the practicality of using a surfactant to aid in
suppressing respirable coal dust in a mine environment.
45
|
Virginia Tech
|
4.3 MATERIALS & METHOD
4.3.1 Description of study
The field test described in this chapter was conducted to determine the
effectiveness of a surfactant to improve the wettability of coal dust, and
subsequently, decrease the dust concentrations to which miners would be
exposed. Additionally, the study seeks to determine how well the surfactant
mitigates the silica concentration of the dust. Due to the known
occupational hazards of respirable dust and the recent increase in the number
of cases of lung disease among mine employees, new and practical
engineering controls which reduce the amount of respirable dust to which a
mine employee is exposed to are urgently needed. This is even more
important when it comes to suppressing the more toxic silica portion of the
airborne dust in the mine atmosphere, since crystalline silica is a classified
carcinogen. The wettability rates resulting from previous laboratory sink
tests suggest the surfactant plays a vital role in the wettability of dust
particulate. In addition, it was observed that dust with higher concentrations
of silica was more easily wetted. The following field testing used a number
of CPDM’s and personal dust monitors to determine the dust concentrations
experienced at locations near the face as well as locations further outby and
inby active continuous mining operations. The measurements were taken
during active cutting, with and without the surfactant addition. These
measurements collected during the test will be subjected to a number of
analyses to determine the total dust concentrations and the silica content. By
examining the differences between the measurements, this study hopes to
determine the effectiveness of the surfactant. The subsequent sections will
46
|
Virginia Tech
|
describe in detail how the testing was conducted as well as the significance
of the resulting measurements.
4.3.2 Experiment Location
The experiment was conducted in West Virginia, in an underground
bituminous coal mine, utilizing room and pillar mining technique. Seam
thickness at this mine is about nine feet. The mine had two supersections in
operation. A split-type ventilation on each supersection was used with an
exhausting fan. For this study, a single supersection with two continuous
miner production units was monitored. The two continuous mining units
alternated production and included three shuttle cars for haulage. Budgeted
production was 300 ft of advance per shift using 40 ft deep cuts.
4.3.3 Equipment
For this testing, two pieces of equipment were used to gather the data. The
first of which was Thermo Scientific PDM3600, which acted as the
continuous personal dust monitor (CPDM). The CPDM is a respirable
personal dust monitor designed for US-based mining applications and
provides real-time measurements. Battery powered pumps draw a
continuous sample, the respirable portion of which is collected and measured
on an exchangeable filter. The PC-based software allows for the recorded
data to be downloaded and reviewed. This piece of equipment can be seen
in Figure 23.
47
|
Virginia Tech
|
In addition, a fifth CPDM and cage sampling apparatus was constructed.
This would have been used as a reserve if one of the other CPDM’s failed.
The reserve CPMD started at the same time as all of the others and was hung
in the intake measuring incoming air as a secondary objective.
4.3.6 Experimental schedule
The experiment was run within a two week time period. The mine operated
with three shifts. The first two, morning and evening shifts, were normally
for production and the third, night shift, is for maintenance. Prior to the
study beginning, pipes to the continuous miner were flushed with regular
mine water. This was done since the mine was, at the time, using the
surfactant in their water. The intent of running normal water before the
study was to make sure the pipes were swept of any surfactant before the
study began. The study began on Monday, July 14th, 2014 with the
surfactant turned off. Sampling was conducted during the evening shift.
Sampling with the surfactant turned off was repeated during the evening
shift on the 15th. At the end of the second run, the surfactant was turned on,
and the surfactant was allowed to flow through the system. This was done to
ensure that the sampling conducted on the 16th was measuring a thoroughly
mixed solution. Sampling was conducted on the evening shifts of the 16th
and 17th with the surfactant turned on. After sampling on the 17th, the
surfactant was turned off. This was done to make sure any surfactant in the
pipe would be washed out by the beginning of the following week. The
entire process was repeated the following week beginning on the 21st of July
and ending on the 24th. Table 4 below shows the experimental schedule in
detail with each week, the three shifts and dates. The red cells denotes when
the surfactant was off, and blue denotes when the surfactant was on.
52
|
Virginia Tech
|
Table 4: Testing schedule
Week Shift Mon. Tue. Wed. Thur. Fri. Sat. Sun.
7/14/2014 7/15/2014 7/16/2014 7/17/2014 7/18/2014 7/19/2014 7/20/2014
1 morning
evening sampled sampled sampled sampled
night
7/21/2014 7/22/2014 7/23/2014 7/24/2014 7/25/2014 7/26/2014 7/27/2014
2 morning
evening sampled sampled sampled sampled
night
Due to changes in the availability of personnel and environmental conditions
the procedure was altered. The new procedures only had one team for both
continuous miners. This change may have led to less detailed observations
during the cut because the researcher had to leave before the cut of the first
continuous miner was finished. It was necessary, due to the need to remove
the other sampling unit from behind the curtain, to allow for tramming of the
second continuous miner as well as placement behind the line curtain at the
new location of the second continuous miner. In addition, the total number
of feet of advance had to be reduced due to the presence of poor roof
conditions. The instability of the roof restricted the amount of advance per
cut (normally 40 ft) to 10-20 ft per cut, depending on location. The
increased frequency and less productive characteristics of the cuts increased
the relative amount of time used for tramming of equipment. Since it
became clear that 100 ft would not be achievable for each CM, for every
shift, the target feet of linear advance was decreased to 60 ft. It is important
to note that the initial plan was followed for the first shift, as it pertains to
personnel. Due to the roof conditions the targeted 100 ft of advance was
never attained.
53
|
Virginia Tech
|
4.3.7 Pre-test
A day prior to the testing the CPDMs were cleaned and programmed. The
mass transducer and the grit pot were removed, and compressed air was used
to clear the sample line and the grit pot. In addition, the surfaces were wiped
down. While the mass transducers were disconnected, the filters were
carefully replaced using the designated filter replacement tool. Then, the
mass transducers were placed back into the CPDM units and secured. The
software WinPDM v7.20, provided by Thermo Fisher Scientific, was used to
program the PDMs to startup and stop times in accordance with shift’s
scheduled start and stop times. After the units had been programmed, they
were charged overnight.
The Zefon gravimetric samplers were also broken down and cleaned. All of
the components of the cyclone including the fittings and holding apparatus
were washed with water and a detergent. Extra attention and care were
taken not to scratch the cyclones surfaces. After washing, the components
were allowed to dry overnight, and the pumps were charged.
Before testing began, the dried components for the Zefon gravimetric
samplers were inspected and assembled. The CPDM units were be placed in
the cages along with the two Zefon gravimetric samplers. When the CPDMs
began to warm up, the cassettes containing the filters were be attached to the
Zefon gravimetric samplers.
4.3.8 Test
Before the production shifts began, the flow data from the main water pump
which fed the mine were gathered. In addition, the tote which contained the
54
|
Virginia Tech
|
surfactant was examined, and the level marked with a date and time
examined. After arrival to the mining section, the team would hang the
cages and record the location during which the mining crew conducted the
pre-shift checklist. In the pre-shift checklist the mining crew would measure
the water pressure on the continuous miner. This data was relayed by the
mine foreman and recorded. Prior to each cut, a water sample was taken
from each CM’s wash-down hose; in addition to the volumetric flow of air
behind the curtain with an anemometer by the mine personnel. After the
cages had been hung, the personal sampling pumps were turned on as the
CM made its first cut, and time was recorded. After each cut was
completed, the feet of advance were recorded in addition to the number of
bits changed. Cuts were directly observed and any changes in the
environment were recorded. The instability of the roof restricted the amount
of advance per cut (normally 40 ft) to 10-20 ft per cut, depending on
location. The increased frequency and less productive characteristics of the
cuts increased the relative amount of time used for tramming of equipment.
During each shift, a target of 60 ft of liner advance was for each CM was set.
Recording continued until a total of approximately 60 ft of advance were
logged from each CM. After the predetermined amount of material was
mined, the Zefon gravimetric samplers were turned off and the time was
recorded. The sampling units were then collected, and the flow was
recorded at the main pump. Finally, the chemical level was inspected and
marked along with date and time. Depending on the day, the surfactant
pump may also be turned on or off.
55
|
Virginia Tech
|
Figure 30: Summarized surface tensions measurement Figure 31: Summarized surface tensions measurement
for week one for week two
It can be seen from the plots in Figure 30 and Figure 31, that when the
surfactant was added to the spray water, the surface tension was consistently
reduced to approximately 36 N/m2. The erratic nature seen in at the
beginning of the second week, with the surface tension ranging from 46 to
74 N/m2 may be due to a lack of flow running through the wash-down hose.
If the wash down hose was not utilized enough between the first and second
week of testing, then the insufficient flow through the hose may have led to
some of the surfactant remaining in the hose at the beginning of the second
week. The flow running through to the sprayer heads of the continuous
miners were as assumed to be ample to clear out the surfactant from the
week before.
4.4.2 Silica Analysis
After the field testing had been conducted, the sealed gravimetric samples
were sent to NIOSH for silica analysis. The silica analysis conducted by
NIOSH used an experimental non-destructive method. The results of the
silica analysis at tabulated in Table 5.
.
57
|
Virginia Tech
|
The limit of detection (LOD) of the nondestructive method is five
micrograms. Using the silica mass of each sample and the total time that the
samples were taken, the concentrations of each sample were calculated.
When compared to the total dust concentration, a silica content could be
derived. Multivariate regression was conducted to determine the variables
which significantly affected the silica content. The analysis looked at the
presence of the surfactant, which CM operator, left or right, and the location,
behind the curtain or in the returns. None of the variables were deemed
significant.
When comparing the average levels of silica, it could be seen that the silica
percentage dropped about 2% when the surfactant was turned on. The drop
in average levels of silica content can be seen below in Figure 32.
Figure 32: Comparison of silica content before and after the addition of surfactant
It can be seen in Figure 32 that the data also had a slightly more narrow
distribution when the surfactant was turn on. Although a drop in silica
content can be observed, it is not statistically significant.
59
|
Virginia Tech
|
Figure 35: Example of a typical total dust plot collected from a continuous miner
In order to extract the average concentrations during testing, the start and
stop time of the CM ripper motors were used. The total dust for the entire
test was found using the total dust levels at each start and stop time. The
sampling rates of 2.2 l/min and the total time of sampling allows the one to
calculate the average concentrations. After the field testing was completed,
the sealed gravimetric samples from the PDS were sent to NIOSH for
analysis. The sampling rate for the PDS pumps was 2.0 l/min. The resulting
dust concentrations from the two devices are shown below in Table 6.
Table 6: Resulting dust concentration from PDS and CPDM
PDS CPDM
Location Respirable Sampling Respirable dust Respirable Sampling Respirable dust
dust (mg) time (min) concentration dust (mg) time (min) concentration
(mg/m3) (mg/m3)
Day 1 LC 0.641 161 1.99 0.62 161 1.75
RC 0.767 179 2.14 1.12 179 2.84
LR 0.295 180 0.82 0.32 180 0.81
RR 0.381 184 1.04 0.28 184 0.69
Day 2 LC 1.112 340 1.64 1.79 340 2.39
RC 1.317 301 2.19 2.38 301 3.59
LR 0.848 409 1.04 1.74 409 1.93
RR 0.217 52 2.09 0.3 51 2.67
Day 3 LC 0.954 270 1.77 2.35 270 3.96
RC 0.755 221 1.71 1.3 221 2.67
LR 0.712 271 1.31 0.86 271 1.44
RR 0.26 227 0.57 0.4 227 0.80
Day 4 LC 1.314 197 3.34 2.06 197 4.75
RC 0.937 175 2.68 1.2 175 3.12
62
|
Virginia Tech
|
LR 0.556 194 1.43 0.91 194 2.13
RR 0.206 180 0.57 0.33 180 0.83
Day 5 LC 1.39 204 3.41 2.18 204 4.86
RC 1.309 165 3.97 2.61 165 7.19
LR 0.324 205 0.79 1.36 205 3.02
RR 0.39 165 1.18 0.72 165 1.98
Day 6 LC 1.211 198 3.06 1.84 198 4.22
RC 0.419 161 1.3 0.69 161 1.95
LR 0.725 214 1.69 0.99 214 2.10
RR 0.213 157 0.68 0.31 157 0.90
Day 7 LC 1.154 222 2.6 2.52 222 5.16
RC 1.491 198 3.77 2.94 198 6.75
LR 0.738 222 1.66 1.14 222 2.33
RR 0.239 193 0.62 0.45 193 1.06
Day 8 LC 1.768 309 2.86 2.5 309 3.68
RC 1.178 228 2.58 1.92 228 3.83
LR 0.414 311 0.67 0.68 311 0.99
RR 0.416 230 0.9 1.07 230 2.11
With the pre and post test weights, the total dust was determined for each
sample. Using the net gains of each sample and the total time that the
samples were taken, the concentrations of each sample were calculated.
Multivariate regression was conducted to determine the variables which
significantly affected the dust concentration. The analysis examined the
impacts of the measuring device themselves, the location in the airway, the
CM operator, and the addition of surfactant. The variables which were
found to have a significant impact were the sample location in the airway
and the measuring devices.
4.4.3.1 PDS vs. CPDM
When examining the influence of the two devices, the average level of dust
concentration was higher for the CPDM when compared to the PDS. The
difference between the two devices can be seen in Figure 36 below.
63
|
Virginia Tech
|
Table 7: Tabulation of cut by cut dust concentrations
Cut depth Starting Surfactant Left/right Curtain/ Number of bits Air flow Dust conc. for
(ft.) depth (ft.) CM return changed (#) (CFM) cut (mg/m3)
20 20 w/o L L 12 14,904 2.88
20 20 w/o L R 12 14,904 0.44
20 28 w/o R L 11 12,690 1.60
20 28 w/o R R 11 12,690 1.07
20 32 w/o L L 10 10,458 3.23
20 32 w/o L R 10 10,458 0.31
20 10 w/o R L 10 11,088 3.63
20 10 w/o R R 10 11,088 1.02
20 40 w/o L L 10 9,261 4.51
20 40 w/o L R 10 9,261 0.28
25 0 w/o R L 9 11,016 5.29
25 0 w/o R R 9 11,016 0.85
20 0 w/o L L 12 11,795 5.04
20 0 w/o L R 12 11,795 1.89
25 0 w/o R L 19 12,587 8.77
25 0 w/o L R 10 13,680 8.44
25 0 w/o L L 10 13,680 6.76
25 33 w/o R R 0 11,875 2.25
25 33 w/o R L 0 11,875 1.07
15 0 w/o R R 18 12,300 8.56
15 0 w/o R L 18 12,300 2.76
15 20 w/o L R 12 9,980 1.50
15 20 w/o L L 12 9,980 5.90
20 20 w/ R R 15 9,643 2.72
20 20 w/ R L 15 9,643 0.61
20 0 w/ R R 23 12,690 5.67
20 0 w/ R L 23 12,690 1.40
25 55 w/ R R 14 13,797 4.82
25 55 w/ R L 14 13,797 2.63
25 0 w/ L R 21 14,960 13.63
25 0 w/ L L 21 14,960 5.62
20 46 w/ R R 11 10,260 4.03
20 46 w/ R R 11 10,260 1.62
20 25 w/ L L 13 9,010 3.52
20 25 w/ L R 13 9,010 1.41
30 54 w/ R L 10 9,324 7.19
30 54 w/ R R 10 9,324 0.14
10 25 w/ L L 0 11,178 4.89
10 25 w/ L R 0 11,178 1.92
20 0 w/o L L 24 18,080 10.24
20 0 w/o L R 24 18,080 7.14
15 20 w/o R L 11 9,207 5.82
15 20 w/o R R 11 9,207 4.33
20 30 w/o L L 4 11,560 3.82
20 30 w/o L R 4 11,560 1.61
15 50 w/o R L 6 10,044 6.64
15 50 w/o R R 6 10,044 0.69
15 20 w/o L L 9 11,760 1.60
15 20 w/o L R 9 11,760 7.26
30 20 w/o R L 12 9,820 7.88
30 20 w/o R R 12 9,820 2.85
10 0 w/o L L 9 15,066 6.72
10 0 w/o L R 9 15,066 1.71
30 65 w/o R L 20 12,690 3.18
30 65 w/o R R 20 12,690 1.66
35 0 w/o L L 18 14,784 13.14
67
|
Virginia Tech
|
Cut depth Starting Surfactant Left/right Curtain/ Number of bits Air flow Dust conc. for
(ft.) depth (ft.) CM return changed (#) (CFM) cut (mg/m3)
20 20 w/o L L 12 14,904 2.88
35 0 w/o L R 18 14,784 4.70
30 87 w/o R L 7 11,088 2.29
30 87 w/o R R 7 11,088 1.31
15 36 w/o L L 0 15,620 10.38
15 36 w/o L R 0 15,620 2.09
30 0 w/ R L 6 17,388 7.21
30 0 w/ R R 6 17,388 1.64
30 0 w/ L L 19 15,302 9.35
30 0 w/ L R 19 15,302 1.90
10 0 w/ L L 6 16,855 21.38
10 0 w/ L R 6 16,855 6.39
35 0 w/ R L 10 12,852 10.14
35 0 w/ R R 10 12,852 0.56
35 0 w/ R L 17 10,792 5.47
35 0 w/ R R 17 10,792 3.39
10 30 w/ L L 5 13,008 4.71
10 30 w/ L R 5 13,008 1.58
20 37 w/ L L 25 14,769 11.97
20 37 w/ L R 25 14,769 1.97
10 0 w/ L L 8 12,600 7.19
10 0 w/ L R 8 12,600 3.16
25 20 w/ R L 7 16,870 9.03
25 20 w/ R R 7 16,870 3.86
10 40 w/ L L 9 11,576 3.10
10 40 w/ L R 9 11,576 1.19
10 47 w/ L L 8 10,240 3.86
10 47 w/ L R 8 10,240 1.94
During the testing, unfavorable roof conditions restricted the CM operators
from taking normal deep cuts of 40 ft. In a few cases, roof falls were
experienced. A number or variables were recorded for each cut and
examined to determine their significance. The first and most obvious
variables were the locations of the monitors: ventilation curtain versus in the
returns. The results are shown in Figure 40.
68
|
Virginia Tech
|
the unfavorable roof conditions. The lack of productivity on the left caused
the right CM to move further away from the right returns in some cases.
This atypical cut schedule and the inability to keep the measurement
apparatuses at a constant distance from the active face may be the cause of
some of the variability seen in the data. Next, the influence of the
surfactants on the levels of dust concentration on a percent basis were
examined. The resulting average dust concentrations are shown in Figure
42.
Figure 42: Comparison of dust concentration with and without surfactant on a cut by cut basis
A slight increase in the average dust concentration can be seen when the
surfactant was added, but this difference was found to be insignificant. In
addition, other variables that could impact dustiness were examined. The
first of these variables were the number of bits changes after the cut. Figure
43 below shows a plot of the average dust concentration versus the number
of bits changed in three groups. The groups chosen for the study were 0-9,
10-19, and 20+ bits changed.
70
|
Virginia Tech
|
The results show that there was a significant difference in the observed dust
concentrations correlating with changes in the volumetric airflow behind the
curtain. The general trend shows that an increase in the volumetric flow
behind lead to higher dust concentrations behind the curtain. The differences
were statistically significant which shows the effectiveness of ventilation as
a primary engineering control which, when used correctly, can limit the
miners exposure by sweeping more dust away from their location at the face.
Lastly, an analysis was conducted with regards to the starting cut depth. The
starting cut depth describes how many feet of coal were previously cut
relative to the beginning of the pillar. The starting cut depth data was split
into five groups; ≤10 ft, 11-20 ft., 21-30 ft., 31-40 ft. and greater than 40 ft.
The analysis shows that the differences between the dust concentrations of
the different groups are significant, but the only group which stood out was
the starting depth of less than or equal to 10 ft. This can be seen in Figure
45.
Figure 45: Dust concentration per cut against the starting cut depth
In an effort to understand the possible cause of the significant difference,
total dust plots were inspected. The primary difference noted was the use of
scrubbers.
72
|
Virginia Tech
|
4.4.5 Influence of Scrubbers
4.4.5.1 Scrubbers
While examining the total dust, discernable spikes were noticed. A
correlation was evident between the large spikes and the starting cut depth.
When the starting cut depth was flush with the pillars, the dust
concentrations were much higher. According to the dust control plan of the
mine, the scrubbers were only to be activated after the CM had reached a
cutting depth of 20 ft. An example of the spike in total dust can be seen in
Figure 46 below.
Figure 46: Example of a typical total dust plot collected from a continuous miner including a highlighted spike
Due to the irregularity of the spikes from test to test, it was decided that each
of the spikes would be isolated, and their characteristics examined. The
criteria for determining whether the cut was identified as “scrubber off”
were as followed. The cut had to start below 20 ft or flush with the pillars,
and the ending depth could not be more than 20 ft. Cuts which were
identified as “scrubber on” cuts had an initial starting depth greater than 20
ft. This ensured that the scrubbers would always be on, as detailed in the
mine ventilation plan. In some cases, the starting depth was at 10 or 15 ft,
73
|
Virginia Tech
|
and then 20 ft of advance would be made. In this case, the scrubber would
be off at the beginning and on at the end. These “hybrid cuts” were not
included in the analysis. This is clarified with the examples below in Figure
47, Figure 48 and Figure 49.
Figure 47. Environmental setup of Figure 48. Environmental setup of Figure 49. Environmental setup of
when scrubber was off during the when scrubber was both on and off when scrubber was on during cut.
cut. during cut.
The following figures show the isolated cuts as total dust plots. In addition,
the series of figures are split into the subgroups with and without the
surfactant addition. Figure 50 shows the isolated cuts with the surfactant off
and scrubbers off. Figure 51 shows the cuts with the surfactant on and
scrubbers off. Figure 52 show the cuts with the surfactant off and scrubbers
on. Lastly, Figure 53 shows the cuts with surfactant added and the scrubbers
on.
74
|
Virginia Tech
|
Figure 53: Collection of isolated cuts which used the surfactant and operated scrubber
From the series of figures (Figure 50, Figure 51, Figure 52 and Figure 53) it
can be seen that the plots depicting the lack of scrubber activation are much
more varied and have a step-like feature.
From these cuts, specific cuts were examined which had similarly recorded
characteristics. The recorded characteristics used to determine the specific
cuts were the number of new bits after the cut, cut depth, and starting depth.
The cuts have been stacked to avoid clutter at the origin. The cuts are also
stacked chronologically, with the first cut starting at the origin. The second
cut starts at the same time and total dust level as the end of the first cut. The
same procedure was followed for the third cut. The basic statistics for the
cuts characteristics are shown in Table 8 below, followed by the cuts plotted
in Figure 54.
Table 8: Statistics of cut characteristics with inactive scrubbers
Water w/o scrubbers Water and surfactant w/o scrubbers
Total 𝑋̅ σ Total 𝑋̅ σ
New bits 54 18 6 53 18 8
Cut depth (ft) 55 18 3 65 22 3
Starting depth (ft) 0 0 - 0 0 -
76
|
Virginia Tech
|
Figure 57: Plot of average total respirable dust from isolated cuts with and without operating scrubber
In Figure 57 a very clear and consistent lowering of the average total dust
can be seen when the scrubbers are activated.
4.5 DISCUSSION
The data gathered during the field test which was presented in Figure 34 and
Figure 35 demonstrates the effectiveness of the surfactant in successfully
and consistently lowering the surface tension of the mine spray water. There
was a perceived issue of the surfactant being left in the washed down hose
water after the first week of testing. Since the hose was not in significant
use and the hose not being left to run for a sufficient amount of time, the
surface tension was not able to return to its normal level of about 72 N/m2 at
the beginning of the second week. Regardless of the still substantial
difference in the surface tension, no significant difference in the dust
concentration was observed with the monitoring instrumentation.
A significant difference was observed, however, between the two pieces of
instrumentation, the CPDM, and the PDS. The CPDM measured a higher
concentration when compared to the PDS. Although this may be an
79
|
Virginia Tech
|
important distinction for the required implementation the CPDM, it has no
important impacts with regards to the aims of the study via relative dust
levels. Additional work, however, should be done to confirm the deviation
between the two devices. This does have much larger implication with
regards to enforcement of the new dust standards, but was not explored with
vigor in this study.
Additionally, a significant difference was observed between the testing
locations, in the returns and directly behind the curtain. This difference was
revealed in the analysis shown in Figure 38. This would be expected, since
given enough time, some amount of the dust will fall to a lower position in
the entry or settle out completely by agglomerating together and acting as a
large sized particle. Additionally, the respirable particle could be attached to
the larger particle and then fall out of suspension.
Some additional factors examined were the number of bits changed and the
quantity of air flow. The number of bits changed was initially thought to be
an influencing factor, but the resulting data collection and analysis showed
otherwise. Airflow, however, was a considerable factor. The data showed
that the higher volumetric flows were sweeping larger amounts of dust from
the working face. This reinforces the use of ventilation as a primary
engineering control for diluting and transporting dust out of the immediate
mine atmosphere.
The changing distances from the face and other environmental conditions
lead to an almost impossible comparison of data (need multiple samples with
the same start and stop depths as well as air velocities). Ideally, a study on
the order of months is required to look at these parameters in depth.
80
|
Virginia Tech
|
However, the nature of mining assures us that more variables, particularly
geologic, would likely be introduced.
Although not initially taken into account, the pronounced effect of scrubbers
on concentrations was clearly observed. An argument for the continuous use
of the scrubbers could be made, but not without first studying the effects of
dust concentration around the CM operator, as well as potential methane
concentrations. To get a complete picture of the effectiveness of the
scrubber, additional work should be done which places a dust measuring a
device by the CM operator. In the past, there were concerns that the
scrubbers would cause recirculation at the face resulting in a pocket of
methane developing (Kissell & Bielicki, 1975). Past studies have shown
that the recirculation of methane at the face will occur when the scrubber is
moving more air than the ventilation (Divers et al., 1981). No deterioration
of ventilation performances was observed in one study, but these were at
curtain setbacks of 25 ft or more with a blowing system (Halfinger, 1984).
Recently, a study in three mines using exhausting ventilation found that
there was an improvement in the respirable content of dust outby the CM,
but not by the CM operator for cuts within 20 ft of the face. Interestingly,
there was also no significant difference in the dust level at the CM operator
location when the scrubber was off or on (Colinet et al., 2013). A strong
case could be made to investigate the role of scrubbers further with
conditions that elevate concerns of methane buildup and negative alterations
to the vent controls. In addition, this preliminary study indicates that dust
exposure may be high for operators when starting a new cut and that
studying the use of specialized engineering controls for this scenario could
prove meaningful in reducing exposure.
81
|
Virginia Tech
|
5 CONCLUSION
This study evaluated the potential of using a surfactant to aid in coal dust
remediation to lower the risk of coal miner developing coal workers
pneumoconiosis. This study specifically examined the effects of adding a
surfactant to water at varying concentrations and varying the specific gravity
of the exposed dust in a static environment. The testing method used in this
instance was a particle wetting test. More detail on the testing procedure can
be found in Section 3.3.3. Additionally, full-scale testing of the surfactant
described in Section 4.4, was as also implemented. A number other
investigations have been conducted in the past with other surfactants. Due
to the lack of a consensus, these investigations were undertaken to shed
more light on the issue.
To evaluate the additive, the concentration of the surfactant in the aqueous
solution was varied in conjunction with a varying specific gravity of the coal
dust. The current mineral content of the coal dust, to which coal workers are
exposed, is of concern since it is known that silica dust is more harmful to
human health than coal dust alone. For this reason, the assumed mineral
content of the coal dust was studied using the specific gravity of the coal
dust as a proxy. It should be noted that the mineral content of the coal dust
tested was not however confirmed. The first battery of tests conducted
looked at three specific gravities and three surfactant concentrations with
each cell of the 3x3 test matrix tested in triplicate. Both of the variables
were found to be significant. The wetting rates of the dust increased with
increasing specific gravity and surfactant concentration. There was also
significant, albeit to a lesser extent, interactions between the two variables.
84
|
Virginia Tech
|
The second battery of wetting test extended the boundaries of the initial
study by including more surfactant concentrations and more classes of
specific gravity. The test matrix included nine surfactant concentrations and
nine gravity classes. The results of the tests showed that both variables
again were significant, and the increase in either one was correlated with an
increase in the wetting rate. A stronger effect from the surfactant was
observed on dust with a smaller specific gravity or a larger coal fraction in
the dust. This may be due to the non-coal fraction, which is already
hydrophilic, being close to its fastest wetting rate with little room for
improvement.
The second effort in this investigation was geared towards a full-scale study.
The surfactant which was being examined was kept at a constant
concentration of 1:5000. This solution was pumped through the mine to the
continuous miner. The solution was then used in the CM’s sprays during
production. A control of no surfactant was also used. Dust concentration
measurements were collected outby the working face both behind the curtain
and further in the returns. Two different types of samples were collected, the
continuous personal dust monitor (CPDM), and the personal dust sampler
(PDS). Both devices were placed in the same location, and the data was
aggregated and examined to see if there was a significant change in the dust
concertation when the surfactant was being used. No significant changes in
the dust levels were observed. However, in the CPDM data there were
distinctive spikes in some places of the data which correlated to the depth of
the cut. The mine ventilation plan did not allow for the scrubbers to be
turned on until a cut depth of 20 ft was reached. The significant decreases in
the dust concentrations from the activation of scrubbers are grounds for
continued work on the topic. This area of future study should be focused on
85
|
Virginia Tech
|
6 FUTURE WORK
There are noticeable critiques and further assessment which should be added
as addenda to these two main studies. With regards to the first study of the
wetting rates, analysis of the dust contained in the material should be
conducted. The analysis should be conducted on the dust before and after
the wetting test. This will both confirm the actual silica content and allow
the investigator to determine whether the higher gravity class material is
selectively wetting the most hydrophilic components of the dust. The
additional knowledge will also shift the correction factor used to derive the
wetting rate.
It has also been noted that the manual method for applying the dust on the
surface of the solution became inadequate when testing high gravity class
material or greater concentrations of the surfactant in solution. If these are
areas of future interest, the concern may be alleviated by introducing an
automated mechanism for the application of dust. The automation should
allow for the rapid and even application of the dust onto the solution surface.
Some populations of the apparatus may include a very tight enclosure with
compressed air as a dispersing mechanism.
One of the more justifiable critiques for future studies was the lack of testing
of the dynamic capability of the surfactant to strip the dust from the air. All
of the benchtop work executed dealt with the wetting rates in a static
environment. Although wetting rates have been successfully used to some
extent in the past for the preselection of surfactants, the abilities of the
additive should still be scrutinized within a controllable mock setup. A
number of tests could be run including particle sizing of the water droplet
from sprays and airborne dust alike using laser diffraction. Relative
87
|
Virginia Tech
|
concentrations of dust could be assessed immediately downwind of the area
of application with an obscuration meter, which actual concentration could
be measured further down a mock entry. This would also allow investigators
to quickly change a variable like pressure, orientation, and sprayer type in
addition to the variable examined in this first study.
The second study is also not without its own areas of improvement. The
principal amendment to the full-scale study is simply further investigation
for the purpose of controlling for a large amount of environmental variation
and atypical mining condition within which the study took place. Due to
unforeseen circumstance, the study was adapted to gather as much data as
possible, even if the conditions and production cycles were not the norm.
Through the data analysis, however, intriguing trends were observed with
the operation of scrubbers. Another engineering control used mine to
diminish the dust exposure. Further investigations should also place
measurement devices on or near the CM operator at the working face.
Measuring the CM operators immediate environment would aid
investigators in confirming the observed improvement in the mine
environment. This is not just the case for examining surfactant or scrubbers,
but also substantiating the assertions made with regards to the high airflow
rates.
The deviation of the CPDM and PDS from each other is also of interest.
The deviations seen should be reexamined by exposing both instruments to
similar conditions with the aim of repeating the findings. The implications
of this difference in measurement are far reaching with regards to the new
dust regulations.
88
|
Virginia Tech
|
Analysis: ANOVA Two-Factor With Replication (α=0.05)
H : There is no significant difference in the wetting rates correlated to the surfactant dosage ratio
0a
H : There is no significant difference in the wetting rates correlated to the gravity class
0b
H : There is no interaction between the variables mention in H and H
0c 0a 0b
SUMMARY 1.3- 1.4-1.45 1.5- Total
1.35 1.6
1:5000 Count 3 3 3 9
Sum 0.48 0.55 1.01 2.04
Average 0.16 0.18 0.34 0.23
Variance 0.00 0.01 0.01 0.01
1:3000 Count 3 3 3 9
Sum 0.58 0.76 1.71 3.05
Average 0.19 0.25 0.57 0.34
Variance 0.00 0.00 0.00 0.03
1:1000 Count 3 3 3 9
Sum 1.81 2.45 5.63 9.89
Average 0.60 0.82 1.88 1.10
Variance 0.00 0.01 0.32 0.43
AVOVA Source of SS df MS F P- F crit
Variation value
Surfactant 4.05 2.00 2.02 51.41 0.00 3.55
Dosage
Gravity Class 1.93 2.00 0.97 24.55 0.00 3.55
Interaction 1.17 4.00 0.29 7.43 0.00 2.93
Within 0.71 18 0.04
Total 7.86 26
Results: H is rejected, H is rejected, and H is rejected
0a 0b 0c
Conclusion: There is a significant difference in the wetting rates which correlate with both the
Surfactant dosage and the Gravity class. There is also, a less considerable, but still significant
interaction between the two variables.
116
|
Virginia Tech
|
Table 55: Multivariate regression of variables during the collection of dust concentration on a cut by cut basis
Data:
Cut Starting Surf Left/ Curtain Number Air Dust
Depth Depth Right /Return od Bits Flow Conc. for
(ft) (ft) CM Changed (CFM) Cut
(#) (mg/m^3
)
20 20 w/o L L 12 14904 2.88
20 20 w/o L R 12 14904 0.44
20 28 w/o R L 11 12690 1.60
20 28 w/o R R 11 12690 1.07
20 32 w/o L L 10 10458 3.23
20 32 w/o L R 10 10458 0.31
20 10 w/o R L 10 11088 3.63
20 10 w/o R R 10 11088 1.02
20 40 w/o L L 10 9261 4.51
20 40 w/o L R 10 9261 0.28
25 0 w/o R L 9 11016 5.29
25 0 w/o R R 9 11016 0.85
20 0 w/o L L 12 11795 5.04
20 0 w/o L R 12 11795 1.89
25 0 w/o R L 19 12587 8.77
25 0 w/o L R 10 13680 8.44
25 0 w/o L L 10 13680 6.76
25 33 w/o R R 0 11875 2.25
25 33 w/o R L 0 11875 1.07
15 0 w/o R R 18 12300 8.56
15 0 w/o R L 18 12300 2.76
15 20 w/o L R 12 9980 1.50
15 20 w/o L L 12 9980 5.90
20 20 w/ R R 15 9643 2.72
20 20 w/ R L 15 9643 0.61
20 0 w/ R R 23 12690 5.67
20 0 w/ R L 23 12690 1.40
25 55 w/ R R 14 13797 4.82
25 55 w/ R L 14 13797 2.63
25 0 w/ L R 21 14960 13.63
25 0 w/ L L 21 14960 5.62
20 46 w/ R R 11 10260 4.03
20 46 w/ R R 11 10260 1.62
20 25 w/ L L 13 9010 3.52
20 25 w/ L R 13 9010 1.41
30 54 w/ R L 10 9324 7.19
30 54 w/ R R 10 9324 0.14
10 25 w/ L L 0 11178 4.89
10 25 w/ L R 0 11178 1.92
20 0 w/o L L 24 18080 10.24
20 0 w/o L R 24 18080 7.14
15 20 w/o R L 11 9207 5.82
15 20 w/o R R 11 9207 4.33
20 30 w/o L L 4 11560 3.82
20 30 w/o L R 4 11560 1.61
15 50 w/o R L 6 10044 6.64
15 50 w/o R R 6 10044 0.69
15 20 w/o L L 9 11760 1.60
15 20 w/o L R 9 11760 7.26
30 20 w/o R L 12 9820 7.88
177
|
Virginia Tech
|
30 20 w/o R R 12 9820 2.85
10 0 w/o L L 9 15066 6.72
10 0 w/o L R 9 15066 1.71
30 65 w/o R L 20 12690 3.18
30 65 w/o R R 20 12690 1.66
35 0 w/o L L 18 14784 13.14
35 0 w/o L R 18 14784 4.70
30 87 w/o R L 7 11088 2.29
30 87 w/o R R 7 11088 1.31
15 36 w/o L L 0 15620 10.38
15 36 w/o L R 0 15620 2.09
30 0 w/ R L 6 17388 7.21
30 0 w/ R R 6 17388 1.64
30 0 w/ L L 19 15302 9.35
30 0 w/ L R 19 15302 1.90
10 0 w/ L L 6 16855 21.38
10 0 w/ L R 6 16855 6.39
35 0 w/ R L 10 12852 10.14
35 0 w/ R R 10 12852 0.56
35 0 w/ R L 17 10792 5.47
35 0 w/ R R 17 10792 3.39
10 30 w/ L L 5 13008 4.71
10 30 w/ L R 5 13008 1.58
20 37 w/ L L 25 14769 11.97
20 37 w/ L R 25 14769 1.97
10 0 w/ L L 8 12600 7.19
10 0 w/ L R 8 12600 3.16
25 20 w/ R L 7 16870 9.03
25 20 w/ R R 7 16870 3.86
10 40 w/ L L 9 11576 3.10
10 40 w/ L R 9 11576 1.19
10 47 w/ L L 8 10240 3.86
10 47 w/ L R 8 10240 1.94
178
|
Virginia Tech
|
Simulation and Analysis of Particle Flow
Through an Aggregate Stockpile
Brian Mark Parker
Abstract
For many aggregate mining facilities, the stockpile is the preferred method of
storing rock. In many aggregate mines, as well as other mines using stockpiling
techniques, understanding the timing and flow of particles through a stockpile is
important for correctly timing samples, making proper process adjustments and overall
stockpile safety. Because much of the research of today lacks important information
regarding actual interior particle movement within a stockpile, a series of Real Time
Distribution (RTD) analyses and stockpile flow models have been prepared and analyzed
for this study in order to better understand the flow characteristics of a stockpile.
A series of three RTD analyses performed on three separate stockpiles provides
information leading to the assumption that stockpiles tend to operate similar to a plug
flow system. While conveyor loading techniques may lead to separation of rocks prior to
traveling through the stockpile, the majority of the rock particles entering the pile remain
near the point of entry, or within the “action” area, and will travel through the pile in a
plug flow, rather than a mixed flow, manner. High Peclet number results for each
analysis prove this assumption to be accurate.
A series of models on three separate stockpiles have been created using PFC3d.
Mainly, the simulations prove PFC3d is capable of showing how stockpile particles move
in three dimensions while monitoring specific particles within the pile. In addition, these
models provided simulation results similar to the results obtained within the RTD
analyses. Results show that particles located directly above the discharge point, or
“action” area, travel through the pile at a faster rate than particles surrounding this area.
Velocity results obtained from the simulations show particles accelerating as they get
closer to the discharge points while also providing evidence of “arching” during the
simulation process. These findings provide a better understanding of internal flow within
the stockpile and ways to possibly predict future stockpile flow issues that may be
encountered.
|
Virginia Tech
|
Chapter 1: Introduction
1. General
Stockpiles are used in many applications within the mining industry. A simple design,
usually associated with little moving machinery, it is an essential part of most mines. In the coal
industry, stockpiles are primarily usually used as short or long term storage. The metals industry
sometimes uses stockpiles as leach beds, with the design and proper layout of the stockpile being
very important. The fine sands and minerals industry uses stockpiles as both a storage and
discharge location with a feeder usually located beneath the pile. The aggregate industry is
similar in that it also uses stockpiles for storage and discharge purposes, but with larger material.
The geometry of a stockpile varies according to the type of material being stored and the
environment the pile is in. The angle of repose, or the angle at which the material makes when
allowed to settle, usually remains consistent around the base of the pile. Various environments
can cause many geometric changes to the stockpile. Such changes include slumping and sliding
along the base due to heavy rain and stockpile deformation due to heavy wind. The following
Figure shows the general shape and design of a stockpile (LuckStone 2005).
1
|
Virginia Tech
|
Figure 1.1 - General aggregate rock stockpile
Stockpiles are used throughout mining over other storage methods because of the various
loading/unloading techniques and the ability to actually store the rock using this method. Silos,
and other storage devices, are an unnecessary added cost which is needed only if the material to
be stored is toxic, flows freely, or requires a special discharge technique. Stockpiling also
eliminates additional maintenance brought forth by corrosion and general wear and tear.
In the case of aggregate mining, stockpiling is the number one method of storing rock.
Stockpiles are used within the plant process primarily as short term storage. Discharge tunnels
beneath the stockpile constantly remove rock from the pile to be taken to other locations within
the plant. This is usually happening while the stockpile is also being loaded by a conveyor.
Stockpiles are also used as short to long term storage of final product. Each product size usually
has its own pile where a front-end loader can easily load drive-up trucks at any point around the
base of the pile.
2
|
Virginia Tech
|
2. Purpose of Research
While stockpile storage of aggregate material is preferred within the aggregate industry,
little research has been done on stockpile flows. In the case of aggregates, sampling the
discharge of stockpiles is important for quality control and when specific tests are needed to rate
performance. Sometimes stockpile feeding or discharge adjustments can be very difficult to
accurately perform and can skew data or cause excessive wait time in order to assure a sample of
a specific rock. There are also instances where various changes in rock hardness and size can
manipulate a plant or cause damage to various parts of a plant. An overall better understanding
of rock movement and general flow through a stockpile would provide knowledge to perform
proper tests and improve overall mine efficiency.
In cases outside of aggregate mining, stockpile flow research would also provide a better
understanding in quality control and safety. In terms of control, understanding when various ore
types are within a pile would allow for proper adjustments at the processing plant. For instance,
knowing when random changes in oxidation levels would be entering a processing plant would
allow for proper adjustments. Stockpile flow research could also assist in understanding
bridging and ratholing. From 2000-2007 there were 49 reportable injuries related to the climbing
of piled material alone (NIOSH, 2009). Considering there were additional accidents related to
stockpile flow and feeder issues, it is evident that many dangers exist while working on or
around stockpiles. Understanding these issues would allow for preventive procedures to be
developed to reduce stockpile accidents.
3. Structure of Research
This research discusses the attempts to prepare a Residence Time Distribution (RTD)
analysis on a discharging stockpile based on various introductory methods and the use of distinct
element method to simulate particle movement through a stockpile. Both of these research
topics will include a summary of program and field work preparation and a discussion of the
results determined after the simulation and analysis have been completed. Final simulation
results hope to provide a better understanding of how rocks react within a stockpile during
discharge from beneath the pile. A completed RTD analysis hopes to provide a better
understanding on how introducing material to a pile can affect the quality of material discharging
from the pile. While the results may not provide answers to all stockpile related issues and
questions, the results should provide a foundation for future research.
3
|
Virginia Tech
|
Chapter 2: Literature Review
1. Literature Introduction
The scope of stockpile research is very limited in terms of particle and rock flow. The
following review will discuss the various simulations performed on stockpiles and research
necessary in the completion of this paper. Past research was limited to quality control, safety and
particle interaction. Stockpile flow research is either non-existent or extremely limited. For this
reason, some hopper flow research and research lacking flow content was reviewed in order to
understand the importance of particle flow. The stockpile flow research which will be covered in
this thesis will relate to all of the research included in this literature review. The understanding
of how particles flow in a pile will hopefully assist any future research of the stockpile.
2. Discrete Element Method
Numerical methods for computing motion of objects are used in many research
applications today. Some methods focus on the problem as a holistic, or continuum. Other
methods focus on many tiny pieces within the system in order to represent the system in a model.
Discrete element method (DEM) is a numerical method for computing motion of a large number
of particles that represent a system.
DEM was first applied by Cundall in 1971 in order to address issues in rock mechanics
(Cundall and Strack 1979). Cundall’s other early studies more focused on microscopic and
macroscopic characteristics of many tiny discs. In one paper, the behavior of soil was modeled
(Cundall and Strack 1983). This was a 2D model that mainly determined material behavior after
applying many different exterior boundary conditions. The program Cundall used during this
time was titled BALL. This program appears to be much like the simpler 2D DEM programs
used today.
Willliams, Hocking and Mustoe in 1985 explained that DEM was more of a generalized
finite element method. This is true in that DEM looks at the individual particles within the
system and provides data according to each individual sphere or circle. Finite element method
breaks down objects or flows into many small spheres or circles, but only does this to relate to
the system as a whole. Finite element method is more of a continuum model while DEM treats
each particle within the model as its own, or as many tiny finite elements (Williams, Hocking et
al. 1985).
4
|
Virginia Tech
|
While DEM uses spherical or round particles in its models, the method is actually
capable of modeling particles with non-spherical shape. This is usually done so by grouping or
bonding many particles together to represent a random mass. This is extremely useful in many
flow applications where non-spherical particles are present.
Discrete element models are very detailed models that usually require a lot of computer
power. Because each particle is treated as its own and has its own set of data, the model usually
is limited to a specific number of particles. While advances in software are beginning to take
place, some programs with many of thousands of particles still need a quick processor (Yasar,
Kuraoka et al. 1997; Wikipedia 2006).
Considering DEM models can take many separate, discrete particles and shape them for
use in many different experiments, it is very useful in many fields. DEM is sometimes used in
liquid applications where the particles are given parameters that allow the application to flow
freely. Bulk materials are also modeled with DEM, such as with silo storage. DEM is also used
in a more straight forward application such as with granular material and powder flow where
each particle acts as a single grain of sand or powder. Each of these examples may represent an
application used within many types of jobs. Though, DEM is more commonly used in mining,
pharmaceuticals, oil and gas, agriculture and food handling fields.
There are many commercial and non-commercial DEM software packages that are
available today. The original Cundall designed software, BALL, is non-commercial software
and is one of the older programs available. Particle Flow Code in 3D, or PFC3D, is one of the
popular programs used today. Considering it is in three dimensions, it is the most realistic of the
programs and was used to develop the models in this thesis.
2.1. Particle Flow Code in Three Dimensions (PFC3D)
The numerical “Particle Flow Code,” or PFC, models sphere particle movement and
interaction using the discrete element method. PFC can operate in either 2D or 3D. Calculations
in PFC are taken over a series of time steps. This makes it easy on the computer memory in that
dynamic equations are performed on every time step, rather than saving and reapplying matrices
as in an implicit scheme. However, many thousand time steps are run, and calculations are done
on each particle for each step. Considering each step is usually a fraction of a second, the
program run time is rather lengthy. This can be an issue considering many applications require a
large amount of particles to properly simulate a system (Lorig, Gibson et al. 1995; Itasca
Consulting Group 2004).
5
|
Virginia Tech
|
A PFC3D simulation is started by developing a model outline. When first developing a
model the schematics must be inserted. This includes any walls or particles that are necessary in
representing the system to be modeled. Information can be inserted into PFC either through a
command line or through an input file coded for PFC. Once all particles and walls have been
properly inserted the program will run a series of cycles. Each cycle performs the specific
calculations stated by the programmer and allows the particles to interact accordingly. After
each cycle is complete, the computer takes the new particle data and starts a new cycle. The
cycling is then repeated until the model is complete. Information requested can then be collected
for analysis (Itasca Consulting Group 2004).
3. Associated Flow Research
3.1. Quality Control Models
While there is limited research on stockpiles and their internal flow, there is much
research involving quality control. Quality control is important in controlling product and
throughput for a plant. Understanding particle flow through a stockpile has a large influence on
quality control. Most of the quality control research of today focuses on quality of what enters or
exits a pile and ignores what goes on inside of the pile. The addition of internal flow research
will assist in perfecting quality control.
Past quality control research, which more focuses on final product, lacks important
information about what goes on within the pile. For example, P. Keleher’s paper discusses how
QMASTOR uses stockpile performance numbers to make proper adjustments to shipments
(Keleher, Cameron et al. 1998). J.E. Everett discussed how simple adjustments to discharge
rates and pickup timing can improve mine life and quality of product (Everett and Kamperman
2000). Similar to Everett’s paper, M.L. Smith showed how production scheduling and inventory
control can be controlled by proper discharge techniques (Smith 1999). Another example is seen
in M.G. Nelson’s paper where simulations were performed on random mine data in order to
determine how to properly blend product using a real-time analyzer (Nelson and Riddle 2004).
In each of these papers, final product was solely focused on while internal flow was ignored.
The use of internal flow data within the stockpiles would have provided a more detailed
understanding of how to both load, and unload, the stockpiles efficiently.
Perfecting blending techniques to control product has also been heavily researched over
the years. Resent developments in simulation techniques have lead to a number of new papers
relating to the topic. Along with Nelson’s paper on simulating a real time analyzer are papers
where simple visual basic programs are used to properly layout blending schemes. Francis
6
|
Virginia Tech
|
Pavloudakis used this technique in his paper to layout stockpile yards in ways that would
properly blend material based on conveyor introduction to the pile. While this blending method
did prove efficient, it lacked information on how the product blended while being discharged
(Pavloudakis and Agioutantis 2003). The following figure presents how specified areas of flow
react during discharge as depicted in an image from a paper by Prescott and Barnum titled “On
Powder Flowability.” Having similar information could have led to an even better program for
Pavloudakis capable of providing higher blending efficiencies.
Figure 2.1 - Flow profile blending during discharge
Aggregate studies have proven that proper mine design and operation timing is important
to maintain an efficient mine. Minestar, by caterpillar, is used in some mines today as an
operational management system. The precise data capture of Minestar enables for better
blasting, quarry control, exploratory drilling, hauling, etc. While this system mostly looks at the
mine efficiency as a whole, it does control stockpile deliveries and blends. Automatic controls
adjust flow and blending of stockpiles in order to produce the desired product (Snow and
Cousland 2002). Once again, this quality control program only adjusts according to trial and
error. There is no flow data that is captured or understood by the program.
G.K. Robinson analyzes blending results based on predictions and proper material
introduction to the stockpile. While this paper similarly discusses blending based on various
stacking methods, this paper focuses on blending material as much as possible before removing
product from a pile. The stockpile model called “CHASM” was used to determine the results of
7
|
Virginia Tech
|
various stockpile stacking procedures. This program along with his research was able to predict
variability of output based on predicted geometry and loading technique. In the research no flow
data was actually needed to get an accurate stockpile output, however, Robinson did mention that
output sampling theory was important to understanding the best possible blending that could be
achieved (Robinson 2004). So while internal flow was negated, stockpile output was deemed
important in determining the most efficient blending technique. Combining these parameters
along with internal flow would have allowed a more accurate model.
Michael Binkowski approached the quality control aspect of stockpiling mathematically.
The goal of his research was to develop a mathematical framework that would be able to model,
analyze and optimize stockyard design. In other words, his research focused on finding the
number and size of stockpiles needed to maximize throughput and cost at the mine. Using a
series of equations, Michael was able to determine a method to optimize stockpile configuration
so that the mine would operate more efficiently (Binkowski and Mccarragher 1999). Once
again, ignoring blending during flow may have made this model slightly inefficient.
Understanding how the material flows while leaving the stockpiles may have led to a different
stockyard layout.
3.2. Granular Flow Studies
Another paper by a consultant at Jenike and Johanson discusses solid flow problems in
bins, hoppers, and feeders. This paper states how walls in a hopper or bin that are too steep or
rough may result in particle hang up, thus resulting in quality control issues (Marinelli and
Carson 2001). In the case of aggregate stockpiles, this is evident, but is not as much of a
concern. For instance, the stagnant rock within a stockpile acts as the walls of pile and
resembles that of a silo. While stagnant rock in a silo may lead to excessive particle segregation
and spoilage, stagnant aggregate rock in a stockpile is absent of these problems mostly because
of the unimportance of aggregate rock remaining on the outside of the pile. The following figure
compares a silo to an aggregate stockpile.
8
|
Virginia Tech
|
Figure 2.2 - Comparison of moving and stagnant material of silo and stockpile.
While issues of rock remaining stagnant around the lower boundaries of a pile have no
effect on the quality of the pile, it does lead to arching within the pile. When the pile’s discharge
location gets smaller, the more capable it is of creating an interlocking arch. Maynard
researched ways to prevent this from happening for cohesive fine material, but did not describe
its effect on larger material (such as aggregate rock) (Maynard 2004). While arching has been
witnessed to occur on occasion in aggregate mining, its effects on quality are mostly due to
downtime. The bigger issue with arching is sudden collapses leading to cave-in and blow-out at
the discharge point. This is of course a major safety issue that will be better explained later.
In order to obtain a consistent, even flow, proper dischargers must be used. Dischargers
encourage the flow of material from a pile and cannot control the rate at which the material
flows. Volumetric or gravimetric feeders are required to actually control rate and exist usually in
fine material applications (Carson and Petro). In the case of an aggregate stockpile, only a
discharger is used to assist in creating flow from a stockpile. This is mostly due to the size of the
material.
Another stockpile flow issue is segregation. Segregation of particles within the stockpile
has been a quality control issue in mining for many years. Overall, segregation occurs over a
wide range of mining applications, though in coal it is especially evident. This is because coal
tends to have very fine material blended with larger particles in order to create a required
product. This usually occurs on a stockpile, where one material is blended with another while
being presented to the pile. This is where sifting occurs. Sifting is simply the movement of finer
particles through a mixture where larger particles are present. Sifting usually leads to a gathering
9
|
Virginia Tech
|
of smaller particles closer to the center of the pile (Carson, Royal et al.). While there are many
methods used today to help prevent this, it tends to be ignored in the aggregate industry. This is
mostly because the stockpile discharge goes through a crusher or series of screens before being
declared product.
3.3. Flow Safety Research
While “ratholing” and internal voids are more of a common quality control problem in
the fine materials industry, they can be very dangerous in the mining industry. P.C. Arnold
performed research on the flow of fine coal stockpiles (Arnold, Fielder et al. 1983). His research
examined the need for a better feeder, or discharging system beneath the pile. Earlier reclaim
tunnels had a feeder design that had low recovery and produced “ratholes” and voids. These
voids were dangerous in that they were capable of collapsing while dozers were mixing on the
pile. In aggregate mines, there is also the danger of a sudden release of energy at the arch
creating somewhat of a “rockburst.” The research eventually led to a feeder design that
produced a 90% increase in recovery and the elimination of stockpile voids and “ratholes.”
Future flow research would be able to determine well ahead of time when this problem was
occurring and possibly locate the problem within the pile.
Other safety research is usually associated with displacement of piles and the results
associated with displacement. B. A. Quinn (1995) had an article which more focused on the
displacement of a stockpile due to fines. He performed lab tests to determine whether excessive
fines and lack of medium sized particles can cause a decrease in stockpile slope stability. His
results showed the changes in shear strength of a stockpile at different layers as a result of fine
coal percolation and wet conditions. As a result of this research, stockpile instability was better
understood and better methods of calculating factors of safety for stockpiles were determined.
Some of the tertiary or secondary stockpiles within the aggregate industry may take on similar
risks during wet conditions.
3.4. Flow Simulations and Models
Models simulating flow of granular material are developed to determine how particles
interact while flowing within a system. Most of the simulations are done using distinct element
method. The following papers use various techniques and methods to simulate flow and particle
interaction within a system. While no stockpile flow research was found, many papers
simulating silo flow were found. These papers were used to gain a general understanding of
flow in a system similar to that of a stockpile.
10
|
Virginia Tech
|
A study which led to successful results on flow within a hopper was reviewed by P.A.
Langston (Langston, Tuzun et al. 1995). This study described the results of a distinct element
method simulation of hopper flow. Using a simple hopper design, similar to the simple stockpile
design used to simulate in this paper, Langston was able to develop accurate velocity, contact
forces and mass flow. The real purpose of his paper was to determine whether this exercise was
more of a research or design tool. At the time of this paper, he concluded that the results were
better used as a research tool until a more user friendly design was available in the future. At
this point in time, there still exists no user-friendly design, and all flow research using distinct
element method mostly goes into research and development. However, some companies do use
very detailed models to assist in equipment design (Metso 2006). Due to the limited stockpile
flow studies, this paper should be deemed an introduction to the subject and thus used as a tool
for future research.
A similar successful hopper flow study was completed by H.P. Zue in 2002 (Zue and Yu
2002). In this paper, Zue stated that granular flow can be classified into two categories:
macroscopic and microscopic. The macroscopic approach is more of a continuum approach
which relies on a series of boundary conditions that balance out a model. The microscopic
approach is more discrete and is based on the analysis of individual particles. This method is a
more accurate way to describe flow because of its focus on individual particles. However, it
usually requires many particles and can lead to an extensive program. Zue’s research has
determined a way to link microscopic variables in the discrete approach to macroscopic
variables. By averaging microscopic results at a probe point a macroscopic solution is
determined. This method was validated on a hopper where averaged stress components were
compared to actual stresses within a hopper (Zue and Yu 2005). This method best represents
how PFC determines macroscopic variables like velocity and stress within its model. Zue’s
method averages after DEM has been completed while PFC actually averages within the
program. With this being understood, flow results calculated similarly through PFC can be
deemed accurate when estimating velocities and flows of granular material.
It is believed that the used of conglomerated particles within PFC best represents rocks
and other randomly sized particles. For example, L. Li used PFC2D to represent angular grains
to realistically simulate the behavior of cemented granular material (Li and Holt 2001). While
monitoring rock breakage was a factor in Li’s experiment, it still proved how the use of
conglomerated particles best represented an actual particle shape. Similarly, Hamid Nazeri used
PFC3D to simulate rock movement through an ore pass (Nazeri, Mustoe et al. 2000). His paper
chose to compare clustered particles to circular particles and compared these results to that of the
mine. It was determined in this paper that stresses were inconsistent with that of circular ore and
that material flow tended to “hang up” in ore passes when using clustered particles. Both of
11
|
Virginia Tech
|
these papers support the use of clustered, conglomerated particles to represent rocks within a
system. However, both of these papers used clustered rocks to represent forces and not rock
flows. Even Nazeri’s paper only found that “hang ups” were more evident when trying to
compare flow.
While some research did recommend the use of clustered, agglomerated particles within
PFC to represent actual rocks, other research mentioned its unimportance. A paper similar to
Nazeri’s decided to just use large circles combined with fine, much smaller, circles to show
actual hang up and ore flow (Iverson 2003). This paper showed acceptable results using circular
particles in PFC2D. S. Masson used PFC2D to compare particle mechanical properties on silo
flow, and confirmed the efficiency of PFC for leading numerical analysis of granular flow
(Masson and Martinez 1999). Fine particle research also involved single spherical particles.
This analysis discovered the flowability of smaller particles representing small glass beads
(Moreno-Atanasio, Antony et al. 2005). In each of these cases, circular particles represented
flow through a silo successfully.
Other simulations include a large number of particles in attempt to represent an actual
silo. D.R. Parisi along with S. Masson (2004) used discrete element method to develop a hopper
that housed 20,000 – 170,000 particles. This method actually layered the particles in order to
validate the model to actual silo flow. Stresses and velocities were also calculated by the model.
By separating the large system of particles into subdomains, the program was able to run a large
number of particles. The results of this paper showed a very small discrepancy in the
displacement fields. The overall results, however, were very reliable, and once again prove that
flow can be measured using spherical particles.
4. Residence Time Distributions
Residence time distributions (RTD’s) are the results determined by tracer tests performed
on any type of reactor. The purpose of an RTD analysis is to determine the mean residence time,
the extent of mixing within the reactor, and to identify unusual operating problems. The
recovery of a given component is determined by these conditions. The extent of mixing within
the reactor is defined as either plug flow, mixed flow, or arbitrary flow. Plug flow represents a
perfect reactor with no mixing; what goes into the reactor leaves the reactor as the same
concentration. Mixed flow represents a reactor where mixing is taking place (Levenspiel 1972).
Figure 2.3 represents the type of reactors and the associated mixing which takes place with each
reactor.
12
|
Virginia Tech
|
Figure 2.3 - Reactor mixing classifications
A perfectly plugged flow is rare in most reactors. Because of this, reactors are measured
by the degree of mixing within the reactor. This mixing variable is defined as the Peclet number,
or mixing intensity. The higher the Peclet number, the more plug flow the reactor is. In the case
of an aggregate stockpile, it is believed that little mixing occurs within the pile. Therefore, the
stockpile is assumed to operate in a plug flow manner. While within the pile, very little mixing
is taking place, the introduction of rocks to the pile may cause immediate rock separation, or
mixing. This paper proves to verify both cases.
When determining the mean residence time and Peclet number of a stockpile, the
following equation is used:
C 1 1 t/2
t exp
C 2 t/ /Pe 4 t/ /Pe
o
Where, C = Species concentration at time t
t
C = Normalizing number (original concentration)
o
t = Current time
τ = species retention time
Pe = Peclet Number (mixing intensity)
This equation was used to determine RTD of a generic reactor, or system. In the case of
a stockpile, a generic reactor was decided upon because there was no real previous knowledge of
internal flow. Once concentrations at various times were determined, the residence time and
Peclet Number calculated were able to generate an understanding of how particles reacted in the
stockpile (Levenspiel 1972).
13
|
Virginia Tech
|
Chapter 3: Residence Time Distribution Analysis of an Aggregate
Stockpile
Brian Parker
Virginia Tech Mining and Minerals Engineering
100 Holden Hall
Blacksburg, VA 24060
Abstract
Much of the stockpile research of today focuses on how to improve efficiencies associated with
stockpiles by simply having a better understanding on blending techniques. In many aggregate
mines, understanding timing and flow of a pile is just as important for timed sampling and
process adjustments. This paper studies the results found from RTD analyses performed on
three separate stockpiles. These results proved that within an aggregate stockpile, rocks
discharge from a stockpile similar to that of a plug flow system. However, it was also
determined that conveyor loading techniques can cause separation of a particular group of rocks
prior to discharge. Early separation of rocks lengthened the discharge distribution time, thus
leading to a lower Peclet number more representative of a mixed system. Mean residence times
found in each analysis were deemed useful for this particular stockpile, but could not be used to
represent retention for every aggregate stockpile. It is recommended that future physical flow
research perform a larger number of RTD analyses in order to develop discharge rate equations
useful for any size and type of aggregate stockpile.
14
|
Virginia Tech
|
1. Introduction
In the aggregate industry, stockpiling is the primary method of storing rock. Both
product and plant rock are stored in stockpiles. Stockpiles are used over other storage methods
because of the various required loading techniques and the ability to actually store the rock in
this fashion. Silos, and other storage devices, are an unnecessary added cost and are needed only
if the material to be stored is toxic, flows freely, or requires a special discharge technique. While
stockpile storage of aggregate material is preferred, little is known on particle flow through a
stockpile when discharged from beneath the pile. This is essential in obtaining proper samples
and in understanding how particles move through a stockpile.
While there is limited research on stockpiles and their internal flow, there is much
research involving quality control. Quality control is important in controlling product and
throughput for a plant. For example, stockpile performance numbers captured by a quality
management program were used to help make proper adjustments as shown in P. Keleher’s paper
(Keleher, Cameron et al. 1998). J.E. Everett (Everett and Kamperman 2000)discussed how
simple adjustments to discharge rates and pickup timing can improve mine life and quality of
product. Another example is seen in M.G. Nelson’s paper where push piles are used to blend
different grades of phosphate by correctly using a real-time analyzer(Nelson and Riddle 2004).
These are just a few examples of the research associated with blending and quality control.
Other research is usually associated with displacement of piles and the results associated
with displacement. B. A. Quinn (Quinn and Partridge 1995) had an article which more focused
on displacement of a stockpile due to fines. He performed lab tests to determine whether
excessive fines and lack of medium size particles can cause a decrease in stockpile slope
stability. Jenike & Johnson (Carson, Royal et al.) has done research focusing on segregation of
stockpiles leading to rat-holing and sifting. Segregation is an issue when loading onto aggregate
stockpiles due to the excessive amounts of fines that load near the center of the pile. Also,
overall safety and control of how particles enter and leave a pile are better understood with this
research.
All of the research available today shows the importance of stockpile management but
lacks the understanding of how particles flow within the pile. In addition, there is very little
experimental research associated with stockpile flow. Research that could prove the amount of
mixing occurring within the pile would allow for better understanding on what to expect during
discharge. In the case of aggregates, sampling the discharge of stockpiles is important for
quality control and when specific tests are needed to rate performance. Sometimes stockpile
adjustments can be very difficult to time and can skew data or cause excessive wait time in order
to assure a sample of a specific rock. There are also times where various changes in rock
15
|
Virginia Tech
|
hardness and size can throw off a plant or cause damage to various parts of a plant. The addition
of experimental stockpile flow research would provide a better understanding of the mixing and
flow rate occurring within an aggregate stockpile.
Although, much of the research of today shows the importance of stockpile management
and control, it lacks important internal flow research that could better explain the movement of
rocks within a discharging stockpile. Provided what we know today, a better understanding of
how particles react upon discharge of a stockpile would allow for better models and more
efficient research. The purpose of this paper is to discuss the results obtained while performing
an RTD analysis on an aggregate stockpile. These results will provide a better understanding on
the degree of mixing within an aggregate stockpile during discharge. Also, mean residence times
of various aggregate stockpiles will be evaluated in hopes of determining a better understanding
of average retention times within stockpiles. It is believed these results will assist in
understanding the above stated issues while also providing new knowledge on internal flow
characteristics of an aggregate stockpile.
2. Analysis of a Stockpile Reactor
In this paper, residence time distribution analyses were performed on many stockpiles.
Residence time distributions (RTD’s) are the results determined by tracer tests performed on any
type of reactor. The purpose of an RTD analysis is to determine the mean residence time, the
extent of mixing within the reactor, and to identify unusual operating problems. The recovery of
a given system is determined by these conditions. These conditions are important in properly
understanding the internal flow and mixing taking place within a stockpile.
Within an aggregate stockpile, three types of mixing are likely to take place: Plug flow,
mixed flow, or arbitrary flow. Plug flow represents a perfect reactor with no mixing; what goes
into the reactor leaves the reactor as the same concentration. Mixed flow represents a reactor
where mixing is taking place (Levenspiel 1972). Figure 3.1 represents the type of reactors and
the associated mixing which takes place with each reactor.
16
|
Virginia Tech
|
Figure 3.1 - Reactor mixing classifications
A perfectly plugged flow is unlikely to occur within a stockpile. Because of this, the
results taken from the stockpile are measured by the degree of mixing within the stockpile. This
mixing variable is defined as the Peclet number, or mixing intensity. The higher the Peclet
number, the more plug flow the reactor is. In the case of an aggregate stockpile, it is believed
mixing occurs differently depending on the location of interest within the stockpile. Therefore,
the stockpile is assumed to operate in an inconsistent manner. It is also assumed that the
introduction of rocks to the pile may cause immediate rock separation, or mixing. This paper
proves to verify both cases.
When determining the mean residence time and Peclet number of a stockpile, the
following equation is used:
C 1 1 t/2
t exp
C 2 t / / Pe 4 t / / Pe
o
Where, C = Species concentration at time t
t
C = Normalizing number (original concentration)
o
t = Current time
τ = species residence time
Pe = Peclet Number (mixing intensity)
This equation was used to determine the RTD of a generic reactor, or system. In the case
of a stockpile, a generic reactor was decided upon because there was no real previous knowledge
of internal flow. This equation was combined with concentration results at different time periods
in order to determined reliable flow data. This was done by assuming original Peclet number
and residence times. The resulting concentration, based on these guesses, was then compared to
17
|
Virginia Tech
|
the original stockpiles concentration. The Microsoft Excel solver function was then used to
adjust mean residence time and Peclet number accordingly until the sum of squares was
minimized. These results were then used to describe the flow of the aggregate stockpile.
3. Residence Time Distribution, RTD, Analysis
3.1. Analysis Introduction
An RTD analysis was performed on three aggregate stockpiles over three days. Each
stockpile had a different shape and each was discharged at a different flow rate. The first RTD
analysis was performed on tracer rocks located at a selected position on the pile. The purpose of
this test was to see how the rocks separated during discharge. The next two tests allowed for
random introduction of particles by means of conveyor. These tests showed how particles
separated when being introduced randomly to the pile and how it may have adjusted the Peclet
number. The following procedure describes how each test was set up and completed.
3.2. Analysis Setup Procedure
Correct sampling and analysis procedures were essential in obtaining proper RTD’s.
There has been little to no RTD analytical work done on stockpiles, so the setup was designed
solely on similar tracer tests and reasonable estimations. In order to perform an RTD analysis
tracer rocks were to be located in the pile and monitored when leaving the pile. Capturing this
information was essential in understanding the mean retention time in the pile and in
understanding how rocks perform when flowing through the pile.
3.2.1. Determining a Location
Preparation of stockpile analysis began by first establishing a mine at which to perform
the RTD analysis. After much consideration of many aggregate mines, it was decided to focus
on Luckstone mines due to previous work experience with the company and the company’s
understanding on the importance of aggregate research. Because Virginia Tech had previously
done research with Luckstone’s Bealeton, Virginia, mine for a Mine-to-Mill project, this mine
was deemed the best option for the analysis. Also, rocks from a previous test at the mine had
been stored at Virginia Tech and could be used for this analysis. Overall, the mine was a perfect
fit for the type of project environment needed for a reliable analysis.
18
|
Virginia Tech
|
3.2.2. Tracer Rock Setup
After choosing the Bealeton mine as the analysis location, the rocks, stored at Virginia
Tech’s office, had to be sorted through in order to choose a viable tracer rock size class. This
was decided based on what size rock would be presented to a crusher and what rock is targeted
as product. In order to produce product, larger rocks from the pit must be crushed and sized in
order to meet product specifications. Understanding this, the decision was made to use rock
sizes larger than the secondary crusher setting but rocks after primary crushing. Rock sizes were
limited to large rocks subsequent to primary crushing mainly because these rocks were the first
presented into a stockpile at the facility, and were of a reasonable size. Considering Luckstone’s
operating parameters, it was decided that two size rocks be used as tracers: 3 to 3.5 inch rocks, or
“small” rocks, and +6 inch rocks, or “large” rocks. These rocks remained larger than the closed-
side setting of the secondary crusher.
Once rocks sizes were decided upon, the quantity of tracer rocks to use in the analysis
was then needed. In order to represent the stockpile, enough tracer rocks had to be used so that
the results could be deemed reliable. In this case, the number of rocks needed to be chosen so
that each could be accounted for when leaving the pile. Most tracers used today tend to be
radioactive and are measured based on the radioactive intensity of the exit stream. A stockpile,
however, cannot be treated this way considering the size and consistency of the material being
studied. Because of this, the amount of tracer rocks needed was decided upon based on predicted
results. Estimating that at least 20% of the rocks will be unaccounted for, it was decided that 50
rocks be used as the tracer sample. Each size class (3-3.5 in. and 6+ in.) will consist of 25 rocks
and will be located together on the pile. Losing 10 of the 50 rocks will not hurt the analysis
enough for the results to be unusable. In fact, 40 rocks were believed to be plenty to determine
mean retention time and mixing intensity.
After determining the size classes and quantity of rocks to use in the analysis, a method
of tracing the rocks was needed. It was first believed that a small transponder could be located
within the rock and the rock could be monitored electronically as it entered and exited the pile.
This method was set up and used during the first analysis. Setting up this method consisted of
drilling small 0.25 inch holes approximately 1.25 inches deep into each rock. A single
transponder was then placed into the hole of each rock followed by a silicon gel application to
both protect the transponder and keep it in place. While this approach worked well in a lab
setting, problems arose while attempting to track the transponders while in the field. It was
believed interference either from the belt idlers or the speed of the belt itself was preventing the
transponders from being read by the antennas placed over the product belt. Because there was
reason to believe this issue may occur, a backup method was used to trace the rocks as they
19
|
Virginia Tech
|
exited the pile. In order to distinguish the rocks as they were leaving the pile, a visual spray
painting method was set up. Each rock in each size class was spray painted in order to
differentiate it from other rocks. The spray painted rocks were noticeably different in color from
the standard gray rocks and were now ready for the stockpile analysis. The spray paint applied
to each rock allowed for each rock to be traced as they exited the pile, regardless of whether or
not the transponders worked correctly. Because of this, transponders were only used during the
first analysis; no resulting data collected using the transponder tracing method has been provided
within this paper.
3.2.3. Equipment Preparation
Once the tracer rocks were ready for the field analysis, visual and survey equipment was
collected for use. The visual equipment consisted of a Sony TRV-900 minidv camera with
tripod, which would be used to visually record the outflow of rock from the pile. This camera
was the premium 3CCD (charged coupled device) digital camcorder which was capable of
monitoring the three primary colors. This camera was chosen because of its technology and its
ability to capture colors well. Survey equipment consisted of MDL LaserAce 300, which is a
laser surveying system. This system was capable of obtaining distances and angles without the
use of a tripod. This would be used to determine the stockpiles height and basic dimensions
before discharge.
3.3. On-site Analysis Setup
Over the course of seven months, three separate stockpile analyses were completed at the
Luckstone quarry in Bealeton, Virginia. The purpose of the first analysis was to determine how
particles separated during discharge when placed at the same location on the pile. The second
and third analysis both allowed random introduction of tracer rocks into the stockpile. The
ultimate purpose of all three experiments was to determine the degree of mixing within the
stockpile while discharging and the mean residence time.
3.3.1. Analysis Setup 1
On October 21, 2005, tracer rocks and analysis equipment were transported to the
Luckstone mine in Bealeton, Virginia in preparation for the RTD analysis. Upon arrival at the
mine, managers and employees were notified of the work that was going to be done and of the
20
|
Virginia Tech
|
assistance that may be needed. Once understood, tracers and equipment were prepped for the
analysis.
With the assistance of managers and employees, a stockpile was chosen for use in the
analysis. It had been previously determined that a primary stockpile would be used, but there
were two on site that could be used. A stand alone stockpile separate from the standard primary
stockpile was chosen to be the best option. This pile was used because its size was smaller than
Luckstones other primary piles and had one drawpoint. The size made it favorable for placing
the tracer rocks at a set location on the pile while the single drawpoint made the pile consistent to
that of the PFC simulation.
Once the pile was chosen, the tracer rocks were ready to be located on the pile. It was
decided that in the first analysis the rocks were to be located together at a central location at the
very top of the pile. While the pile was not completely full, there was still a peak at the top of
the pile where the tracer rocks could be placed. In order to place the tracer rocks, a manlift
hoisted the rocks up to the pile where the rocks could be located safely. Once the rocks were
placed on the pile, the manlift was lifted higher so that the stockpile could be surveyed. Figure
3.2 shows placement of the rocks at the peak of the stockpile. Figure 3.3 shows a side view of
the pile with the tracer rocks sitting at the top of the pile.
Figure 3.2 - Placement of rocks at peak of stockpile
21
|
Virginia Tech
|
Figure 3.3 - Side view of pile (tracer rocks located at peak)
3.3.2. Analysis Setup 2 and 3
Analyses 2 and 3 were setup similarly on April 19, 2006 and May 23, 2006 at the
Luckstone mine in Bealeton, Virginia. Like the first analysis, the manager and employees were
notified of the situation and were prepared to assist in the exercises. Upon arrival at the mine,
the same stockpile used in the first analysis was filled to near-full capacity and surveyed.
Unlike the first analysis, these analyses allowed for random introduction of tracer rocks
to the pile. This was done by adding the tracer rocks randomly to the conveyer which feeds the
stockpile. Once the conveyor was loaded with the tracer rocks, the conveyor was turned on and
the rocks were dumped onto the pile. This method best represented how aggregate rocks were
presented to a stockpile. Doing this allowed for rocks to react realistically when hitting the pile
and allowed for a random tracer rock distribution. Figure 3.4 shows the rocks on the conveyors
just before being presented to the pile while Figure 3.5 shows how the rocks reacted after hitting
the pile.
22
|
Virginia Tech
|
Figure 3.6 - Still visual taken from video
3.4. Stockpile Analysis
The stockpile feeder and conveyor were started once the setup was completed. The
camera was also started while employees and helpers were located next to the belt in order to
monitor rock flowing on the conveyor. When tracer rocks passed, they were noted and a brief
time was recorded. This was carried until the pile was exhausted and the conveyor belt ran
empty. When the pile was empty, cumulative times, flow, and tonnage values were recorded.
3.5. Stockpile Analysis 1 Results
Final results show that the stockpile feeders ran for approximately 0.82 hours until the
30.9 foot stockpile was completely exhausted. During this time the total rock discharged
averaged 563.4 short tons per hour, tph. This resulted in a total of 462.0 short tons over the
course of the analysis. These results were measured by the online control system used to
monitor rock flow through the plant.
The study of video analysis results proved that 45 of the 50 rocks were accounted for
during the analysis. This 10% error was less than the 20% anticipated before the test began.
However, all of the rocks that went unaccounted for were in the “small” size class. This is
understandable considering the smaller rocks could have been trapped under larger rocks and
visually unable to detect. It was not believed that the lost smaller rocks affected the final results.
24
|
Virginia Tech
|
Table 3.1 shows a summary of the tracer rocks and the time range that they left the
stockpile. Ranges were used to classify when rocks left the pile so that a mean retention time
and diffusivity of the pile could be classified. This information was used to develop the model
seen in Figure 3.7. This plot represents the short time period where the concentration of rocks
leaving the system was high. As mentioned earlier, 5 rocks are not included in the small rock
count column due to being unaccounted for.
Table 3.1 - Analysis 1 rock count summary
25
20
15
10
5
0
0 10 20 30 40 50
-5
Minute Range
#
noitartnecnoC
kcoR
Combined
Large
Small
60
Figure 3.7 - Rock concentration over time for analysis 1
In order to determine the mean retention time and how the flow of the pile was in accordance to
where the rocks were located, the following equation was used:
C 1 1 t/2
t exp
C 2 t/ /Pe 4 t/ /Pe
o
Where, C = Species concentration at time t
t
C = Normalizing number (original concentration)
o
t = Current time
τ = species retention time, mean residence time
Pe = Peclet Number (mixing intensity)
25
|
Virginia Tech
|
Figure 3.8 - Rock concentration over time for analysis 2
The same procedure that was used to determine the results for analysis 1 was used on
analysis 2. The following table shows the results obtained from the data collected during this
analysis:
Table 3.3 - Calculated RTD analysis 2 results (time in minutes)
Mean Residence Time 18.25
Peclet Number 112.94
Normalizing Number 2.17
Weighted sum-of-squares 29.79
The weighted sum-of-squares was significantly higher here than what was calculated for
analysis 1. This occurred due to the higher number of outlier rocks that discharged later into the
analysis. The Peclet number also dropped in analysis 2 because of the larger time range at which
the tracer rocks left the pile. The majority of the tracer rocks discharged the stockpile in analysis
1 between the 9:20 to 12:00 minute range. Analysis 2 showed that conveyor method of
introducing the tracer rocks to the pile caused an immediate separation in random directions.
The majority of the tracer rocks in analysis 2 discharged between 14:00 to 24:00 minutes, as seen
in Figure 3.8. The larger time range proved that the system in its entirety, conveyor loading
method included, has some mixing during the loading and unloading process. However, the
results do show that the majority of the rocks were grouped to a single range.
The mean residence time in this analysis was also higher than that of analysis one. This
was expected due to the significant decrease in discharge flow rate during the second analysis.
27
|
Virginia Tech
|
The same procedure used to determine the results for analyses 1 and 2 was completed on
analysis 3. The following table shows the results obtained from the data collected during this
analysis:
Table 3.4 - Calculated RTD analysis 3 results (time in minutes)
Mean Residence Time 20.20
Peclet Number 793.30
Normalizing Number 1.64
Weighted sum-of-squares 37.41
As with analysis 2, the weighted sum-of-squares was significantly higher for analysis 3
than what was calculated for analysis 1. Once again, this occurred due to the higher number of
outlier rocks that discharged later into the analysis. In this particular case, there were more
outliers that sat farther away from the primary discharge range between 18 and 22 minutes.
Because of this, the sum-of-squares was even higher than that of analysis 2.
Unlike analysis 2, the Peclet number actually increased higher than that of analysis 1.
This higher Peclet number, which represented a more plugged flow, was understandable given
the results of the analysis. Analysis 3 had a higher peak discharge period closer to 20 minutes as
compared to 10 minutes in analysis 1. Figure 3.9 also shows how the plot resembles that of
analysis 1 rather than analysis 2. This basically means that the tracer rocks had to remain closer
together throughout the loading and unloading period for a longer period of time. In analysis 2,
it appeared that the particles separated more during loading, leading to a more dispersed rock
concentration plot. In analysis 3, the particles didn’t separate as much upon loading onto the
stockpile, with the exception of a few that rolled closer to the bottom of the pile and became
outliers. This resulted in a lower discharge rate range and a concentration plot similar to that of
analysis 1.
The mean residence time in this analysis was also higher than that of analyses 1 and 2.
This was expected due to the even lower discharge rate and increased stockpile size. To further
explain what was stated above, the higher mean residence time along with the lower discharge
rate range contributed to the high Peclet number.
29
|
Virginia Tech
|
4. Summary of Results
The results compiled from each of the separate RTD analyses proved different rock flows
and reactions during discharge. The first of the three analyses was completed in order to see how
rocks located tightly together would flow through an aggregate stockpile. The visual results seen
in Figure 3.7 along with a high Peclet number prove that tracer rocks located in a group at the top
of a stockpile tend to discharge in a plug flow manner.
Analysis 2 was completed in order to gain a better understanding on how rock
introduction to the stockpile affected particle separation. When the tracer rocks were introduced
to the stockpile by means of conveyor, it appeared that separation did occur as a few of the larger
tracer rocks traveled down the pile (Figure 3.5). This form of separation resulted in a lower
Peclet number, or a more mixed system. As observed, the mixing was mainly due to the
separation of particles while being introduced to the pile and not due to the discharging of the
stockpile.
Analysis 3 was completed similarly to analysis 2. The purpose of this analysis was
mainly to validate the results collected from analysis 2. Like analysis 2, the tracer rocks
appeared to separate when introduced to the stockpile by means of conveyor (Figure 3.5).
However, the Peclet number in this case increased significantly compared to both analyses 1 and
2. Figure 3.9 shows a much “tighter” range where the majority of the tracer rocks discharged.
The few rocks that did roll down the pile while loading had little effect on the RTD analysis, and
could be considered outliers.
30
|
Virginia Tech
|
5. Conclusions
Based on the results compiled from the Stockpile RTD analyses performed at
Luckstone’s Bealeton, VA mine, many conclusions can be made. Analysis 1 proved that tracer
rocks located together at the top of the pile tend to travel through the pile and discharge together
as a plug flow. While this is true for rocks located at the top of the pile, or within the range
directly above the feeder, it may not be necessarily true for particles located at other locations on
the pile. It also appeared in analysis 1 that the rocks located directly over the feeder tended to
travel faster through the pile. This was noticed when the mean residence time calculated to be
just over 10 minutes. With an entire operating time of 54 minutes, this means that the rocks
located centrally at the top of the pile discharged within 1/5 of the total discharge time.
Analyses 2 and 3 proved that separation does occur during the conveyor loading process.
Even though analysis 3 showed results similar to that of a plug flow stockpile, the process was
still considered somewhat mixed because of the visual separation during loading. However, as
mentioned earlier, analysis 2 had more of a visual separation upon loading to the pile. Other
than this difference, the pile shape and total quantity of rock to be discharged from each pile
were very similar. With this being understood, the rocks behavior when being loaded to the
stockpile can be deemed arbitrary or immeasurable using the type of tests performed on these
analyses. In order to better track overall rock displacement during the loading of the pile, many
more sample rocks would be needed.
Overall, it was determined that the aggregate stockpile discharges similar to that of a plug
flow system. The introduction of rocks to the pile more influences a mixing of the pile than the
discharge of the rocks from the pile. While mean residence times were useful in understanding
the general residence time within each pile, they were not very useful in generating an overall
understanding of retention within the stockpile. It should be suggested that future physical flow
research perform a larger number of RTD analyses in order to develop a correlation between
loading and discharge features. In addition, future research should perform residence time
distribution analyses on various different starting locations on the pile. This could lead to the
development of rate equations based on stockpile height, or volume. Future research may also
want to prove how rocks located within the area directly above the discharge point tend to be
dominant and travel faster through the pile.
31
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.