University
stringclasses 19
values | Text
stringlengths 458
20.7k
|
|---|---|
Colorado School of Mines
|
Leach pads are constructed from impermeable (relatively impermeable) layers of
prepared soils, clays and geotextiles such as high-density polyethylene (HDPE), very-
low-density polyethylene (VLDPE) or polyvinyl chloride (PVC) (Strachan, C. and
Dorey, R., 1988). A pad may be placed on relatively flat topography, on a hillside or in a
valley depending on it’s proximity to the mine. Lupo, J.F., 2006, Thiel, R. and Smith,
M., 2004, review several of the geotechnical factors involved in designing heap leach
pads, alternative geosynthetic liner systems, leach solution drainage piping and
geomembrane compatibility. In all cases an underdrain system over the pad liner system
and under the overliner material is designed and placed to provide hydraulic conduits for
the pregnant solution out from under the heap (Dorey, R., van Zyl, D.J.A and Kiel, J.,
1988).
Heaps are formed by placing run-of-mine ore, or ore that has been crushed and/or
agglomerated and/or chemically treated to achieve desired leach characteristics, on top of
a leach pad (Dorey, R., van Zyl, D.J.A and Kiel, J., 1988, Chamberlain, P.G and Pojar,
M.G, 1984). Muhtadi, O. A., 1988, describes various heap building techniques and
solution application scenarios; however, there is no mention of selective ore placement as
an alternative to the random placement of mined ore.
With increased use of heap leach operations worldwide there is a need to better
understand the parameters, processes and reactions involved. Theoretical simulation and
modeling of heap dynamics and leach extraction is useful in estimating leach results
without actually performing testwork. Multiple models have been constructed to better
understand the heap leach process (Bartlett, R.W., 1991). Prosser, A.P., 1989 presents a
heap leach simulation model that may be used during process development, results from
which are ultimately compared to results from column leach tests and actual operations.
O'Kane, M. et. al, 1999, present a framework for improving the ability to understand and
predict the performance of heaps. Metsim is a brand of simulation software that may be
used to model heap leach operations. Swanson, V.F., 1994, demonstrates its use in
evaluating many heap leach operating parameters. Similarly, Ogbonna, N., 2005,
12
|
Colorado School of Mines
|
describes heap modeling using Heapsim. A heap leach model considering unidirectional
flow of lixiviant (as used in this dissertation) is described by de Andrade Lima, L. R. P.,
2004.
The use of an exponential decay curve (as used in this dissertation) is presented
by Menne, M.D. and Muhtadi, O.A., 1988, Quast, K.B., 2000, wherein the residue of
leached ore is a function of the original ore concentration, a first order rate constant and
the cumulative leach liquor quantity recovered from the base of a heap. In that only
consistent leach application rates are considered in this dissertation, recovered leach
liquor is directly proportional to leach time, thereby allowing estimates of residual ore
grade to be prepared on the same basis. The leach curves used in this dissertation are
similar to all other leach curves presented in the literature.
The author has combined power and exponential curves to represent leach
extraction in the five scenarios presented herein. Any other leach curves that may be
developed theoretically or experimentally may be directly applied without detriment to
the modeling and concept of selective ore placement. Miller, G., 1998, provides several
leach curves for a copper leach system. Hurfst, 1989, used linear leach curves in her
analysis as presented by Prosser, A.P., 1988, which, although simplified for this example,
may accurately represent the leach profile of some ores.
2.3 Enabling Technologies
Components and activities related to mine and process operations and their
relationship to heap building and leach extraction are graphically depicted in Figure 2.1 -
Current Use of the Components, Parameters, Modeling and Ore Characteristic Analyses
Available for Selectively Placing Ore on a Heap. Important to consider is that a
13
|
Colorado School of Mines
|
Block modeling is discussed in detail in Hustrulid and Kuchta, 1998, p. 214;
Knudsen, H. P., p. 293; and Stanley, B. T., p. 43. Block modeling and mine planning
information is typically generated and stored using mine modeling and planning software
such as these listed in Table 2.3 - Mine Planning Software and Companies with Internet
Addresses.
Table 2.3 - Mine Planning Software and Companies with Internet Addresses
Surpac - http://www.surpac.com/.
Gemcom - http .7/www. gemcomsoft ware, com/.
Vulcan - http://www.vulcan3d.com/.
Runge - http://www.runge.com/.
Mintec - http://www.mintec.com/. and
Mincom - http://www.mincom.com/
These mine software packages typically contain ore block modeling and
visualization tools, design tools, planning tools and operations tools, all with substantial
data storage and manipulation capabilities.
Results from metallurgical tests that may be performed on drill cuttings from
production blastholes is a final measure of ore characterization that should be considered
before ore placement. When ore grade is the characteristic upon which selective
placement is based, the ore block grade generated during block modeling should be
amended as appropriate with the block grade estimate obtained after analysis of the
production drill cuttings (Miller, G., 1998). If, however, the ore characteristic(s) upon
which selective placement is determined will not be estimated during tests of the
production drill cuttings, or if the time to perform such tests is too lengthy to allow the
incorporation of such results before placement, then the characteristic(s) determined
during block modeling should be used.
16
|
Colorado School of Mines
|
With regard to the mine planning tools referenced above, use of selective ore
placement techniques should be considered and may ultimately create variations within
the mine planning process. If a more desirable mine block delivery sequence is shown to
improve ultimate leach extraction, then the mine should make adjustments during the
planning phase to provide the best sequence of ore blocks to the heap as is practicable
(Miller, G., 1998).
2.3.2 Rapid Differential Global Positioning Systems (GPS) Survey
Use of the GPS is the only means by which mine operations may effectively and
economically place ore at specific locations on a heap in real time. Zoschke, L.T., 2000,
describes the use of GPS in open pit mining operations and various modifications to GPS
so as to achieve required accuracy. Figure 2.3 - Tracking Mine Haul Trucks Using GPS,
graphically depicts the use of the GPS and component interaction with a haul truck. A
mine applying GPS would commonly have each machine in the mine fleet, including
drills, shovels, loaders, dozers and trucks, outfitted with the appropriate equipment.
17
|
Colorado School of Mines
|
Caterpillar promotes MineStar as.
“...an integrated mining information system. Through a series of
component programs for production reporting, truck assignment, health
monitoring and fleet analysis, MineStar links information gathered from
machines in the field and linking it to the office business enterprise
systems. MineStar provides the tools and services needed to help the mine
get the most from their mining assets while lowering the cost per ton.”
MineStar Material Tracking and MineStar Machine Tracking are both specific
Caterpillar proprietary GPS related softwares already in use at mine sites worldwide.
These technologies coupled with the mine software previously noted are the enabling
technologies that allow the application of selective ore placement techniques.
Intellimine is a product of Modular Mining, a subsidiary of Komatsu. The
installation and implementation of Intellimine is divided into the areas of system audit,
engineering analysis, performance analysis and corporate benchmarking, essentially
providing the same basic functions as those described in the CAT MineStar system (a
competitive product) described above.
2.4 Advanced Selective Placement Opportunities
Three fundamental areas, which may be improved to enhance selective placement
opportunities, are discussed below.
2.4.1 Ore Characterization
Quantitatively and qualitatively, ore characteristics may be beneficial or
deleterious to the leach extraction process and represent inherent qualities of an ore that,
19
|
Colorado School of Mines
|
without exception, must be taken into consideration during mining and processing
operations (Miller, G. 1998). Methods and procedures have been developed to analyze
and evaluate ores so as to metallurgically take advantage of favorable ore characteristics
while compensating for those that are deleterious to the leach extraction process
(Malhotra, D. 2006). Bottle roll, agglomeration and column leach tests are most
commonly performed during the development of heap leach processes for specific ores.
It is these types of tests that must be frequently and consistently performed so as to
properly characterize ores before placement on a heap. Seldom is it possible to
completely remedy problems caused by such fundamental parameters (Malhotra, D.,
2007).
Ore characterization techniques must be accurate, timely and representative to be
of use to selective ore placement operations. There are many references to tests that may
be conducted to determine ore characteristics (Malhotra, D., 2006). McClelland, G.E.
and McPartland, J.S., 1990, describe the usefulness of metallurgical testing and
comparisons between such tests to the results obtained during operations. Miller, G.,
2003 specifically describes geotechnical parameters as determined in the laboratory and
their correlation to those observed during heap operations.
Additional ore characteristic tests need to be developed that may be performed
relatively quickly so that results may be considered before ore is placed. Correlation to
ore characteristics used during block modeling would be made and ore placement
locations could be finalized. An example of such a test is the Field Leach Test (FLT)
used by the U.S. Geological Survey to simulate the chemical reactions that occur when
rocks and soils are leached by water. The 5-minute FLT gives similar results to the 18-
hour Synthetic Precipitation Leaching Procedure (SPLP) used by the U.S. Environmental
Protection Agency (USEPA, 2002).
The importance of testing and the extent to which testing may be conducted may
be best summarized by reviewing the active leach program at Kennecott’s million-ton test
heap (Schlitt, W.J., 2006). Of the 332 million metric tonne resource, 307 million metric
20
|
Colorado School of Mines
|
tons are represented by the testwork. Extreme care has been taken through all aspects of
this test to portray actual leach conditions, as they will occur when leaching this material.
Of interest to the author is the excellent ore characterization and staking that was
performed during the construction of the “full size commercial leach panel”. This
demonstrates that such careful classification by ore characteristic and stacking on a leach
pad is within current full-scale operational capabilities. Further, the use of heap base and
interlift: collection channels demonstrates the capability of leach operations to better
sample solutions coming from specific lift sections than is commonly performed today.
Several fundamental ore characteristics that commonly affect metallurgical
performance of ores during the leach processes are summarized below.
2.4.1.1 Ore Grade Ore grade is the single most important ore
characteristic in any mining operation and is given in units of percent metal or grams of
metal per tonne material (ore or waste). Ore grade, as an ore characteristic, is used from
initial exploratory investigation and valuation of a deposit though development,
operations and reclamation. Forensic analyses of mine wastes, milled tailings and/or
leached residues for metal grade are often conducted to evaluate the possibility of
secondary recovery from these materials (Ruzycki, A., 1992). The author concludes that
ore grade is often the only parameter consistently and repeatedly applied to direct mining
operations and process controls. Multiple techniques are available to determine ore grade
including wet chemistry (SGS Group, 2007), fire assay (Bugbee, E.E. 1940) and
radiometric techniques (Macdonald, W.G., 1997).
2.4.1.2 Mineralogy/Petrology and Liberation Mineralogy and petrology
are the sciences by which rock masses both barren and mineralized are classified.
Petrology and mineralogy of host rock and it’s mineralization attempt to quantify the
21
|
Colorado School of Mines
|
fundamental characteristics of ores upon which development and operating plans are
based to mine and process the material. Through an understanding of the geology in and
around ore deposits, it is possible to determine how valuable minerals are disseminated
and how the characteristics of those minerals may be best exploited when encountered
during process operations (Peters, W.C., 1977). Mineralogy also describes deleterious
mineral constituents and the similarities, differences and associations that may exist
between these and the other minerals. Often such information can be used during process
design and operations to assist in separating the valuable from the gangue minerals.
Understanding the occurrence and association of minerals in an ore matrix allows better
decisions to be made regarding the extraction and recovery processes. Baum, W., 1990,
and Gasparrini, C. 1984, describe the usefulness of mineralogical evaluation of ores.
Mineralogy (as enabled by ore microscopy) allows examination and evaluation of
mineral sizes, structures, shapes, geochemical compositions and associations of the
contained valuable and gangue mineral constituents. Particle size at which liberation
occurs (of the valuable minerals from gangue minerals) and the type of liberation (being
intergranular or transgranular) may be determined through such examination. Kelly, E.G.
and Spottiswood, D.J., 1982, briefly describe several automated techniques that may be
used for the characterization of particles and their potential impact on mineral processing
operations.
2.4.1.3 Fracture Orientation/Frequency. Rock Hardness and Rock Quality
The hardness of a rock, as it relates to the ability of a lixiviant to penetrate and flow
through it, the planes in which the rock fractures, and the orientation and intersection of
those fracture planes relative to the volumes within the rock bearing the valuable mineral
ultimately define the size to which such rock must be fractured to achieve acceptable
economic leach extraction. Mineral bearing rock typically needs only to allow lixiviant
to access the mineral surface for the leach reaction to occur (Spottiswood, D.J., 1982).
22
|
Colorado School of Mines
|
Rock quality designation (RQD) as determined by a geologist tending a drilling
rig is the first systemized accounting of rock hardness as core drilled material is
recovered and logged (LeRoy, L.W. and LeRoy, D.O., 1977). Little fracturing typically
represents hard competent rock and is reflected by a high RQD value. Substantial
fracturing represents poor rock competency and is reflected by a low RQD. Leach
extraction does not require substantial particle liberation for the process to be effective
(such as with flotation and gravity concentration processes.) Low RQD values would
commonly indicate that the rock should be easily leachable without substantial size
reduction requirements. Low RQD values may also indicate that a larger percentage of
material will be made up of fines, often a sign of clays with the potential to affect heap
permeability.
Caterpillar describes Aquila Drill Management products, which...
.. .provide basic systems from production monitoring and automated rock
recognition to more advanced control and high-precision GPS-based
guidance systems. Drill Manager empowers the operator with information
to make intelligent decisions, work faster and with fewer errors.
Productivity is improved through optimized drilling procedures
(Caterpillar, 2007).
The Aquila Drill Management system would allow correlation to the RQD
estimated during block modeling. If tied to a particle size analyzer (Kennedy, J.M.,
1994), real time estimates of the amount of fines and ultimately estimates of heap
permeability (Bartlett, R.W., 1997) may be made available to management in real time
allowing selective placement of the ore on the leach pad.
Various blasting techniques, blasting agents and crushing methods have been used
to improve leach extraction by acting on these fundamental ore characteristics.
Production drilling blast patterns with less burden and spacing and the same hole
diameter increase powder factor yielding greater fragmentation and finer blasted material.
The explosive may be also changed or modified to increase or decrease the shock and/or
23
|
Colorado School of Mines
|
gas volume induced during blasting, again with changes to the resultant blasted product
(Passamaneck, 2002).
Kennedy, J.M., 1994, presents work on a practical technique for determining the
size distribution of rock on blasted benches, waste dumps and heap leach sites.
Photometric techniques allow real time estimates of size distribution potentially useful in
determining an ore’s permeability before reaching the leach pad.
Knecht, J., 1994, Patzelt, N. et al., 1995, describe the use of high pressure
grinding rolls (HPGRs) in the minerals industry. The ability of HPGRs to fracture rock
matrix along it’s mineralized borders while maximizing microfractures in the ore has
been shown to improve leach extraction. Newmont Boddington, and Twin Creeks
operations have realized these benefits (MPD conference, 2002).
2.4.1.4 Rock Permeability Host rock permeability as achieved after
fragmentation and placement to specification on a heap is critical to the even distribution
and percolation of leach solution through ores (van Zyl, D.J.A., 1988). Permeability is
measured in meters/second2 or Darcy’s (Darcy’s commonly used in petroleum
engineering.) This process of fluid flow in and out of rock is known as “imbibition and
drainage.” Too low a permeability may substantially slow the rate at which leach
solutions may be applied to an ore ultimately increasing the leach duration necessary to
achieve economic extraction. Since there are additional costs associated with additional
leach time, low permeability may directly impact an ores value. Variable permeabilities
in mined ore can also pose substantial deleterious affects by allowing acceptable leach
solution percolation and metal extraction in some areas (volumes) of a heap while
reducing leach solution flow and metal extraction in others (Orr, S., 2002).
Agglomeration before placement on a heap is the typical means by which low
permeability is addressed (Heinen, H.J., McClelland, G.E. and Lindstrom, R.E., 1979).
Agglomeration may be achieved by adding water or water and an agglomeration agent
24
|
Colorado School of Mines
|
such as cement at specific points on the conveyor line. Pug-mills are used to intimately
mix ore with water and agglomerating agents when more stringent treatment is required
to increase heaped ore permeability.
2.4.1.5 Deleterious Materials Deleterious materials contained in an ore
deposit may be organic or inorganic in nature. Carbonaceous material intimately
associated with gold ore can rob gold from a cyanide solution after the gold has been
extracted from the rock. If only a small amount of carbonaceous material is contained in
a gold ore, a carbon adsorption circuit instead of a Merrill-Crowe circuit may be used
therewith allowing metallurgical activated carbon to compete with natural - lower
activity carbon for the gold-cyanide molecule.
Copper in gold ores under leach can consume up to four (4) times its
stoichiometric content through the stepwise formation of Cu(CN)4 , (CRC, 1990).
Recently a few methods have been developed for handling copper cyanide problems in
gold process circuits. One method involves preferential copper stripping from activated
carbon using a cold caustic strip before the hot caustic/cyanide strip used to recover the
gold (Bentzen, E., 2006). The other method involves the use of reverse osmosis to
separate the copper cyanide from the gold cyanide molecules (Young, T, 2006). Gangue
acid consumers such as dolomite that are basic in nature (i.e., geochemically generate
solutions with pH greater than 7) can consume sulfuric acid during the leaching of copper
ores. In all cases, overall extraction and/or recovery process are degraded to some extent
due to deleterious elements contained in the ore.
25
|
Colorado School of Mines
|
2.4.2 Selective Ore Placement and Decision Process Improvement
Timely analysis of ore characteristics beginning during the mine planning stage
and continuing through mining operations to final ore placement allow leach parameters
to be improved for selectively placed ore volumes. Selective ore placement, as prepared
by the author herein, may include segregated, averaged and/or selectively combined ore
columns based on goal seek optimization as determined from specific inherent ore
characteristics as they may occur when arranged in ore columns.
Miller, G., 1998, observes that when the leaching of a lift section commences,
large volumes of ore termed “management units”, that are formed by the accumulation of
ore on the leach pad over several weeks (if not months), are simultaneously put under
leach. Miller describes the wide variability of ore characteristics between the many ore
blocks within such a large volume. Ores blocks placed randomly form ore columns with
various combinations of ore characteristics, many of which will not achieve the same
ultimate leach extraction when uniformly placed under identical leach conditions.
Marsden, J.O., 1993 describes forensic testwork on a heap that contained several higher
grade zones, equivalent to 4% of the total ore on the pad, that could benefit from
additional leaching. The author has personally seen similar situations at the Green
Springs Mine and at the Alligator Ridge Mine in White Pine County, Nevada.
Miller continues to discuss the permanence of “as placed” material and the limited
number of “after the event” management tools available to the operator with reference to
the possibility of selectively placing material as an additional means of improving leach
extraction. These alterable leach parameters include lixiviant application rate and
sequencing, total lixiviant applied, lixiviant chemical make-up and concentration, leach
duration and ancillary heap modification such as installed forced aeration. The
unalterable leach parameters, no longer available to mine management after lift section
placement, include further comminution through blasting or crushing, agglomeration and
mixing with additional reagents and modification to ultimate heap height. The author
26
|
Colorado School of Mines
|
sees selective ore placement as an attempt to increase leach extraction by compensating
for unalterable conditions while taking advantage of the alterable leach parameters.
There is a difference; however, between the author’s concepts and methods
regarding selective ore placement and the method by which Miller plans to account for
and reduce the variability of ores delivered to the pad. Miller’s concept is to alter the
mine plan such that the mine sequence will deliver a more homogenous ore for the
construction of an entire “management unit” (i.e., lift section). In this case; however,
even though the mine plan will revolve around ore homogeneity in sufficient quantity to
construct and entire lift section, the ore will still be placed randomly on the heap. The
author’s plan is to selectively place ore material on the heap through which ore columns
are formed in manageable volumes to which target leach parameters may be applied.
Miller also discusses the need to continuously update the database of
metallurgical information from block modeling through production leaching so as to
obtain an accurate reflection of the results obtained when specific leach parameters are
applied to the characterized management leach unit.
Urmie Hurfst wrote on the topic of selective ore placement for her 1989 Masters
Thesis, “Evaluation of Stacking Alternatives in Gold Heap Leach Operations”, T3711 at
the Colorado School of Mines under the direction of Dr. William Hustrulid. In it, the
concept of averaging ore grades over the entire extent of a heap is discussed. Ms. Herfst
correctly concurs with Miller that stockpiling with selective withdrawal and placement is
the only method by which averaging of each ore column can be fully optimized. Both
Hurfst and Miller submit that stockpiling heap leachable ore is not a practical alternative.
Ms. Hurfst’s work did not consider segregated placement of ore blocks by grade as
delivered directly from the mine nor other averaging and/or optimizing techniques within
the practical capabilities of a typical mining operation as developed herein.
Applied operations research techniques must be involved in mine planning and
ore block sequencing and location processes if selective ore placement is to truly be
optimized. Linear programming techniques (Johnson, T.B., 1984) are commonly used
27
|
Colorado School of Mines
|
when the objective is to blend or average ores with different characteristics for mill feed
thereby reducing fluctuations in the milling circuit. These mathematics; however, best
apply only when ore volumes may be divided into portions such that the combinations
thereof meet predetermined constraints. Integer programming techniques (Woolsey,
R.E.D., 2002) are most applicable for optimizing the placement of discreet selective
mining units (SMUs) which must be excavated and placed as entire single units as
proposed herein. Further, it is imperative that the selective ore placement technique be
seamlessly integrated into the operation if it is to be accepted by mine management and
provide value to the mine owners (Woolsey, R.E.D., 1999).
An application of selective placement similar in many ways to that proposed is
already in operation in the agricultural industry. The John Deere GPS field navigation
system for agricultural application is called GreenStar (John Deere Co., 2007).
Agricultural planning software using GreenStar includes the Apex Farm Management
Software with components of AutoTrac, Parallel Tracking, Field Doc, Map Based
Prescriptions and Combined Yield Systems for monitoring and mapping crops and the
characteristics associated therewith. The similarities to the associated MineStar systems
are striking; however, it is currently only the John Deere system that takes advantage of a
GPS directed farm fleet to selectively identify areas with specific crop characteristics and
selectively place amendments as needed on precisely located areas of the agricultural
plot.
2.4.3 Advanced Leach Proceedures
Several leach process variables, beyond selective placement, that may be
modified to optimize leach extraction include agglomeration, lixiviant application type,
rate and concentration; and ore crush size/type. Although not investigated in this
dissertation, the author references these advanced techniques both for their contribution
28
|
Colorado School of Mines
|
to enhanced leach extraction and as a demonstration of the lengths to which current
thinking has advanced and the effort that has been (and is commonly) expended to
employ such techniques. Selective placement would prove less expensive, simpler to
implement and easier to control than many of these techniques.
Staging application rates so as to apply greater rates on fresh ores then
successively tapering off to lower rates on ores already under leach for some time period
has been previously used at the Coeur Rochester Mine (Marsden, J.O., 1993). The author
considers this as an advanced leach procedure still in development with much still to be
learned about imbibition and drainage as it relates to heap permeability and leach
extraction.
Orr, S., 2002, discusses enhanced heap leaching and shares insights into the heap
leach process with regard to hydraulic parameters and fluid flow. These heap parameters
and fluid flows within a heap can be modeled using the Multiple Resolution Décision-
Support System (Meystel and Albus, 2002). Orr indicates that the structure of heaps and
dumps, especially with regard to leach solution flow, is determined and affected by every
stage of their construction from blasting to crushing (when not placed as run-of-mine) to
conveying and stacking. Leach extraction may be improved through a better
understanding of the flow of gasses and liquids within a heap during leach operations.
Such understanding is then applied during the mining, construction and operation of
future heaps.
Advanced leach procedures include alternative leach chemistries that may be
applied to enhance leach extraction. Multiple lixiviant alternatives exist and may be
modified both in chemical content and concentration so as to further enhance leach
extraction if applied to volumes of ore selectively placed. The author has listed and
briefly describes several alternative lixiviants and modifiers that may be used during gold
leach extraction. Other chemistries and lixiviants would equally apply to the leaching of
other metals.
29
|
Colorado School of Mines
|
Halogens including bromine and iodine have also been shown to for complexes
with gold and silver (DeVries, F.W and Hiskey, J.B, 1992).
Calcium Peroxide has been used successfully to improve silver leach extraction
from ores at CRA’s Hidden Valley property, New South Wales, in which the silver is
mineralogically associated with sulfides (Roden, S.J., Chouzadjian, K.A., Bums, R. S.,
Smith, T. J., and Spieth, MA., 1990). Hydrogen peroxide is also known to enhance gold
and silver leach extraction and reduce cyanide consumption with the addition of oxygen
to the heap (NcNeille, A. and Damon, L.L, 1992; Lee, V. et al., 1989).
Thiourea, CS(NH^, has been demonstrated to leach gold and silver in acidic
conditions rather than the basic conditions required for cyanide leach (Hiskey, J.B., 1984,
1981). Thiourea is favored in applications where sulfides may need to be oxidized before
leach extraction to liberate the precious metals. Thiourea has also been shown to achieve
a leach rate 10 to 12 times faster than cyanide with greater selectivity and lower toxicity.
Thiosulfate, S2O3"2, forms complexes with gold, silver, copper, mercury, iron,
nickel, cobalt lead and cadmium (Sillen and Martell, 1964). IN 1858, sodium thiosulfate
was used in the Patera Process to leach gold and silver (Liddell, 1945).
Malononitrile, (% (€% , has been demonstrated to dissolve gold under alkaline
conditions forming either the organic complex Au[CH(CN)2)2] 2 or the common cyanide
complex after hydrolyzing (Heinen, H.J., et al. 1970).
Kozin, L.F. and Melekhin, V.T., 2004 present alternative lixiviant chemistries
including thiocarbamide solutions containing Fe(III) ions, hypochlorite-chloride,
bromine-bromide, iodine-iodide and copper-thiosulfate solutions.
McClelland, G.E. and Eisele, J.A., 1982, present concepts for improvements for
the recovery of silver and gold from low-grade resources that focus on agglomeration of
ores with various amendments and, in particular, Portland cement. Marsden, J.O., 1993,
makes reference to the Brewer Gold Mine in South Carolina where a polymeric
agglomeration aid was successfully used to reduce cement additions.
30
|
Colorado School of Mines
|
CHAPTER 3
RESEARCH CONCEPT AND METHOD
The author proposes that heap leach extraction from a single leach pad can be
significantly improved through the selective placement of ores with specific inherent and
propagated metallurgical characteristics within distinct heap volumes. Such placement
would allow similar ores to be selectively leached under improved, if not optimized
process conditions.
This section of the dissertation, Chapter III, develops a base case including tables
and figures that represent current heap leach methodology. The next sections, Chapters
IV and V, present the results obtained through selective ore placement by comparison to
this base case.
Selective placement allows ores to be segregated, averaged or placed within a lift
section volume so as to achieve some quantitative or qualitative goal. The objective is to
improve leach extraction by tuning leach parameters based on distinguishable inherent
ore characteristics as determined through metallurgical testwork. Before placement, ores
may be treated to enhance amenability to leach extraction by a wide variety of processes
and materials (such as by crushing, agglomeration on chemical/biological treatment)
(McClelland, GE., 1988), or may be hauled and dumped directly on the heap as run-of-
mine (ROM) ore (Muhtadi, O.A., 1988). A wide variety of processes are also available to
treat ores once placed including variations in chemical lixiviant constituents, strengths,
31
|
Colorado School of Mines
|
temperatures, application rates, frequencies, repetitiveness and durations; all of which
maybe varied over time (Kozin, L.F., 2004, Orr, S., 2002).
Improved leach extraction is the goal of using such techniques and is the primary
outcome to which all other techniques are measured and compared. Additional costs
associated with such techniques, if incurred, must be considered in determining the best
economic alternative to pursue.
A series of five hypothetical scenarios involving open pit mining coupled with a
heap leach operation were developed to evaluate the potential for improving heap leach
extraction through selective ore placement techniques. Grade, tonnage and deleterious
mineral content was obtained for ore blocks within a real gold mine currently under
development including mine bench maps showing plan views of the material to be mined.
This data was used in conjunction with mathematical models typical of open pit
mine/heap leach process combinations to develop the scenarios.
Gold ore cutoff grade was arbitrarily designated by the author in this base case to
be 0.023 toz Au/ton, a value that would be determined internally by an operator based on
capital costs, operating costs, project risks and expectations. At current gold prices near
$580 per toz Au, ore at this cutoff grade has a value of $13.34 per ton. Cyanide soluble
copper is the deleterious constituent that occurs naturally with the ore. There is no cutoff
grade nor economic value for deleterious copper; however, the model considers that
levels of cyanide soluble copper that result in cyanide consumptions above 2.0 lbs
CN/ton ore reduce leach extraction from ore columns and cause significant problems
within the recovery circuit.
The five alternative mine operating and heap placement scenarios were developed
to simulate placement and leaching of ores on a heap leach pad. The first scenario
demonstrates random placement and leaching of ores in a Lift Section, the current
method by which heaps are constructed and operated.
32
|
Colorado School of Mines
|
The second, a segregated placement scenario, demonstrates the opportunity for
increased leach extraction (in comparison to the random placement scenario) when ores
with similar characteristics are selectively placed on a heap, each being processed using
conditions most favorable for that ore. In this scenario, ores are characterized and
segregated by grade into three equal groups and placed adjacent to one another on the
same pad.
The third scenario, again using segregated placement, was prepared by sorting
ores by grade into two equal groups rather than three, thereby allowing comparisons to be
made between the benefits in leach extraction when more or less segregation is
implemented.
The fourth, a placement scenario involving both the placement of ore on a heap
segregated by gold grade followed by averaging of the segregated rows of ore columns
by cyanide consumption (lbs CN/ton ore) due to deleterious cyanide soluble copper
Cusoi%, was developed to demonstrate additional leach extraction improvements that
could be achieved when selective ore placement techniques are applied congruently.
The fifth scenario involves averaging cyanide consumption over an entire Lift
Section by Lift Section Rows thereby substantially reducing the possibility of creating
ore columns with cumulative deleterious copper contents higher than the maximum
acceptable (2.0 lbs/ton ore) due to random placement.
In Scenarios 1, 2, 4 and 5, a heap design is contemplated that requires 243 ore
blocks per Lift Section to be delivered every 40 days. Lift Sections in these scenarios are
square, 9 blocks wide x 9 blocks long x 3 levels deep when complete allowing easy
division into 3 equal portions. Scenario 3 contemplates a heap design using 240 of the
same 243 blocks used in the other scenarios, less three blocks nearest the average grade
allowing direct comparison to be made to the other scenarios with very little error. The
Lift Section in this scenario is 8 blocks wide x 10 blocks long x 3 levels deep for easy
division into two halves. The grade and deleterious material characteristics of each ore
33
|
Colorado School of Mines
|
block are known before excavation and are tracked through placement on the heap. A
mine plan to support the heap design was developed resulting in a sequenced schedule of
6 ore blocks per day that would be delivered to the heap for placement.
Proforma income statements were developed for the various mine and heap-
operating scenarios allowing the economic impacts of randomly and selectively placed
heaps to be considered. The only line item, which would change significantly, is leach
extraction and those profits related thereto. Except for slight increases in process
operating costs when lixiviant application is increased, there are essentially no expensed
cost differences between the various scenarios. Ore production, waste production, ore
grade, capital cost, operating cost, refining cost, refining schedule (regarding penalty
elements), appreciation, depletion, amortization, project indirect costs and sustaining
capital remain essentially constant for all 5 scenarios presented. For this reason,
differences in leach extraction and estimated expensed dollar values associated with such
differences for the various scenarios will form the entire basis of comparison. Other line-
items that may change slightly to better support selective placement are highlighted on
the proforma income statements and are discussed individually.
3.1 Ore Deposit - Block Model Data
Ore blocks located on a specific mine bench were obtained from a block model
developed for a real gold deposit. The two characteristics associated with each ore block
are gold grade (toz Au/ton) and cyanide consumption (lbs CN/ton ore).
34
|
Colorado School of Mines
|
3.2.2 Mine and Heap Design Criteria
A summary of specific information pertaining to the mine, heap and leach process
is contained in Table 3.4 - Pertinent Mine and Heap Leach Design Criteria. As shown,
the in-situ ore blocks are cubic and 25 feet long on each edge in situ. Ore density is 181
lbs ore / ft3. One in situ ore block has a volume of 15,625 ft3 and contains 1,414 tons of
ore. With an estimated swell factor of 30%, the broken ore from one selective mining
unit (SMU) increases from 15,625 ft3 to 20,313 ft3 after blasting.
3.2.3 Mine Operations
With regard to mine operations, a typical scenario may include production
blasthole drilling using an Atlas-Copco DML drill with a single pass capability of 31 feet.
This would comfortably provide the 25-foot long blastholes and up to 6 feet of sub
drilling as required. A typical powder factor would be 0.4 to 0.7 lbs ANFO per ton ore
depending on rock hardness, rock quality designation (RQD), blast design and the desired
final broken rock size.
CAT 992G - 15 cubic yard loaders working in conjunction with CAT 777D - 100
ton haul trucks, or some combination of similar and compatibly sized equipment, would
be appropriate for use in excavating with this size of operation. Typically, at least one
CAT D9R bulldozer would operate in the pit and one CAT D8R or smaller bulldozer
would operate on the heap to assist with the plug dumping of the trucks (Hustrulid, W.
and Kuchta, M., 1998). A CAT 16G grader would be used to maintain the haulage roads.
In addition, balanced complements of additional and ancillary equipment, including fuel
42
|
Colorado School of Mines
|
Table 3.4 - Pertinent Mine and Heap Leach Design Criteria (continued)
Pertinent Heap Parameters -9x9
SMU’s to Heap - Daily # 6.075
ROM Ore to Heap - Daily tons 8,590
Lift Section Tonnage tons 343,602
Lift Section Height ft 20
Level Height ft 6.67
Lift Section Plan View Shape plan view shape square
Dimensions of Lift Section ft x ft x ft 20x497x497
Plan View Area of Lift Section ft2 246,803
Pertinent Heap Parameters -8x10
SMU’s Mined - Daily # 6
ROM Ore Mined - Daily tons 8,484
Lift Section Tonnage tons 343,602
Lift Section Height ft 20
Level Height ft 6.67
Lift Section Plan View Shape plan view shape rectangular
Dimensions of Lift Section ft x ft x ft 20 x 442 x 552
Plan View Area of Lift Section ft2 243,984
trucks, lube trucks, tire trucks, water trucks, maintenance vehicles and light towers would
be used to operate the mine.
The ability of the haul trucks to adequately maneuver on the heap so as to
selectively place the ore was considered. As presented in Table 3.4, the heap has a square
plan view shape, 497 feet per side. The concept presented herein relies on dividing the
Lift Section into thirds, each rectangular section being 497’ long and 165’8” wide
(49773).
This width will more than adequately allow the CAT 777D haul trucks to turn
within its turning radius of 83’0” (CAT Performance Handbook, Edition 35). Even the
44
|
Colorado School of Mines
|
largest CAT haul truck, the CAT 797B (payload target weight of 380 tons) with a machine
clearance turning circle of 138’3”, would have adequate room to negotiate this Lift
Section width.
The forgoing information regarding mine equipment is critical to actual mine
operations; however, provided that Lift Sections can be constructed at the rate discussed,
further detail regarding mine equipment is inconsequential. Ultimately, fleet sizing is
determined from mine economics developed with consideration for the total amount of
material to be excavated and the mine production schedule to achieve this level of
production.
3.2.4 Foregoing Mine Excavation and Heap Placement Assumptions
A geometric consideration must be made regarding the edge effects of mining
adjacent blocks so as to adequately define the concept presented herein. During
excavation, portions of adjacent ore and waste blocks are inherently and inadvertently
excavated with ore blocks being mined at that time. Fortunately, as ore is excavated
along a bench (that is primarily made up of ore), the grades of ore in adjacent blocks are
not as variable as the variability between blocks separated by some distance. The
quantity of material in the blocks actually removed relative to the quantity of material
being commingled from adjacent blocks and the true change in actual excavated block
grade that occurs due to such commingling suggests that the blocks can essentially be
mined selectively (enter the term “Selective Mining Unit” or SMU) (Hustrulid, W., and
Kuchta, M., 1998). In mine engineering practice it is common for the SMU to purposely
be configured the same size as an ore block.
Similarly, portions of adjacent ore blocks being placed on a heap are inherently
and inadvertently commingled, forming a parallelogram slightly overlapping with
45
|
Colorado School of Mines
|
previously placed Lift Sections due to the angle of repose of rock as it is dumped and
dozed for placement. This overlapping; however, is relatively insignificant compared to
the volume of material being placed within the Lift Section. After ore placement and
upon commencement of lixiviant application, natural dispersion of leach solution around
a Lift Section beyond the perimeter due to hydrostatic head further reduces the
significance of this geometry (Orr, S., 2002).
The various analyses presented in this dissertation consider and accept that
commingling of adjacent blocks occurs in the pit and on the heap during actual mine
operations. However, for the analyses presented herein, it is assumed that the integrity of
the selective mining unit is maintained from mine excavation through placement on the
Lift Section and that the ore blocks are stacked vertically and at right angles throughout.
In professional mine engineering practice, in addition to other geologic safeguards
used to prevent the overestimation of block grade, it is common to assume that during
excavation, 3wt% of the ore will inadvertently be rejected as waste and that an additional
3wt% of the ore will be replaced by waste (Jolk, R.W., 2006). This essentially has an
overall effect of reducing ore grade by 3% and is often so applied during formal mine
reserve calculations.
3.2.5 Heap Lift Section Design
Ore blocks are configured in a square pattern 9 blocks on a side (9 blocks wide x
9 blocks long - 81 blocks/level), 3 levels deep (81 blocks/level x 3 levels = 243 blocks/lift
section) making up one Lift Section. This Lift Section configuration is ultimately
achieved in scenarios 1, 2, 4 and 5 regardless if random or selective placement scenarios
are employed. A schematic showing an exploded view of the stacked ore blocks as
46
|
Colorado School of Mines
|
described above is shown in Figure 3.4 - Exploded View of Random Stacked Ore Blocks
on Levels 1, 2 and 3, composing a Lift Section. The blocks are conditionally formatted
so as to visually discern blocks with gold grades from 0.0000 to 0.0295 toz Au/ton, from
0.0295 to 0.0395 toz Au/ton and above 0.0395 toz Au/ton.
Once complete, a Lift Section is put under leach for a total of 40 days. Leach
solution is pumped to the top of the heap at a rate of approximately 987 gallons per
minute, 24/7 to maintain the 0.004-gpm/ft2-application rate on the 246,803 ft2 of Lift
Section top surface.
X > 0.0395
0.295< x <0.0395 all in toz Au/ton ore
X < 0.0295
Figure 3.4 - Exploded View of Random Stacked Ore Blocks on Levels 1, 2 and 3
3.3 Lift Section Construction - Random Placement
As ore blocks are sequentially delivered from the pit, the heap is constructed by
end dumping 100 tons of ore at a time from the 777D mine trucks (sometimes also
termed “plug dumping”). The D8R dozer operating on the heap works to level the plug
dumped piles forming a flat surface. Ore blocks are placed in complete rows of 9 blocks
47
|
Colorado School of Mines
|
Table 3.5 - Block Removal Sequence and Placement in Rows of 9 Blocks per Row
Placement Block Block Placement Block Block Placement Block Block
9 Blocks Removal Gold 9 Blocks Removal Gold 9 Blocks Removal Gold
per Row Sequence Grade per Row Sequence Grade per Row Sequence Grade
1 1 0.052 1 28 0.025 1 55 0.067
2 2 0.044 2 29 0.028 2 56 0.058
3 3 0.058 3 30 0.027 3 57 0.031
4 4 0.035 4 31 0.024 4 58 0.035
5 5 0.042 5 32 0.039 5 59 0.036
6 6 0.046 6 33 0.045 6 60 0.079
7 7 0.038 7 34 0.051 7 61 0.068
8 8 0.038 8 35 0.040 8 62 0.056
9 9 0.032 9 36 0.038 9 63 0.060
1 10 0.034 1 37 0.053 1 64 0.035
2 11 0.047 2 38 0.049 2 65 0.042
3 12 0.044 3 39 0.042 3 66 0.042
4 13 0.027 4 40 0.037 4 67 0.067
5 14 0.036 5 41 0.059 5 68 0.072
6 15 0.039 6 42 0.059 6 69 0.078
7 16 0.039 7 43 0.052 7 70 0.082
8 17 0.027 8 44 0.043 8 71 0.075
9 18 0.035 9 45 0.039 9 72 0.074
1 19 0.034 1 46 0.084 1 73 0.023
2 20 0.029 2 47 0.065 2 74 0.024
3 21 0.029 3 48 0.062 3 75 0.025
4 22 0.024 4 49 0.056 4 76 0.023
5 23 0.025 5 50 0.042 5 77 0.023
6 24 0.028 6 51 0.028 6 78 0.024
7 25 0.023 7 52 0.031 7 79 0.025
8 26 0.045 8 53 0.087 8 80 0.028
9 27 0.040 9 54 0.069 9 81 0.025
49
|
Colorado School of Mines
|
2Au + 4CN" + 0 2 + 2H20 = 2Au(CN)"2 + H202 + 20H‘ (2)
These reactions result in very stable gold-cyanide complexes and occur
preferentially in weak alkaline solutions at pH above 10.
Mathematical formulas were developed by the author to correlate the
concentration of residual gold in ore at any leach time given the initial concentration of
gold and the specific leach rate constant for the hypothetical scenarios presented herein.
The formula (equation (5)) expressing equally combined exponential fit (equation (3))
and power fit (equation (4)) curves simulate metallurgical leach extraction from a lift
section providing the basis upon which leach extraction is modeled. The residual gold
grade in ore “C (toz Au/ton ore)”, after a specific number of leach days “t (days)” when
starting with a head grade of “Co (initial toz Au/ton ore)” is defined as the average of the
concentrations from equation (3) and (4) resulting in equation (5).
the exponential fit equation: C=C0e'kt, (3)
and the power fit equation: C=C0tb, (4)
resulting in the combined fit equation: C=Co(e"^ + tb )/2 (5)
The hypothetical exponential and power curve leach rate constants k and b have
values of 0.04 days'1 and - 0.43 days'1 (minus 0.43 days'1) respectively. Figure 3.5 - Time
vs. Residual Gold Grade (toz Au/ton) Using the Combined Fit Equation, Starting with
Minimum, Average and Maximum Grade Ores, Residual Grade Capped at 0.007 toz
Au/ton ore, depicts the reduction in ore grade (i.e., depicts the residual grade) with time
in ores with various initial grades (i.e., head grades) when leaching commenced.
Hypothetical minimum grade capping allows leach extraction to continue with
time following this combined curve until a minimum residual gold grade of 0.007 toz
55
|
Colorado School of Mines
|
Au/ton is reached. No further leach extraction occurs after the minimum capped grade is
reached.
It should be noted that the metallurgical heap leach extraction curve represents the
overall rate of extraction from a heap and not the dissolution rate of gold in cyanide
solution. Included in this rate are all geotechnical and metallurgical rates for all ore
blocks taken as a group, commencing with percolation of lixiviant through the heap to the
ore particle, imbibition of lixiviant into the particle, reaction of the gold with the
lixiviant, drainage of pregnant solution out of the particle and percolation of pregnant
solution to the base of the Lift Section.
The leach rate at any point in time during leach extraction is the timed rate of
change or mathematical derivative (equation (6)) of the leach curve function.
leach rate : dC/dt = (C0/2)(-ke'kt + bt15’1) (6)
An important observation is how leach rate, after any multiple of days, is still
directly related to the initial ore grade unless leaching has continued to the point at which
the minimum residual grade is achieved.
The curve derived above is used as the leach extraction curve in all modeling of
the gold leaching scenarios presented and represents expected overall metallurgical leach
extraction from all ore blocks combined and exploited during typical mining operations.
This combined curve is presented graphically in Figure 3.5 - Time vs. Residual Gold
Grade (toz Au/ton) Using the Combined Fit Equation, Starting with Lower, Average and
Higher Initial Ore Grades, Residual Grade Capped at 0.007 toz Au/ton ore. It is
important to note that other extraction curves, real or hypothetical, could be used in place
of the curve derived above for any of the scenarios presented herein.
56
|
Colorado School of Mines
|
0.1400
S 0.1200
c
c
1 0.1000
2
ô 0.0800
TJ
5
° 0.0600
2
o
ï 0.0400
(0
3
%
0.0200
S
0.0000
0 10 20 30 40 50
Time - Days
Figure 3.5 - Time vs. Residual Gold Grade (toz Au/ton) Using the Combined Fit
Equation, Starting with Lower, Average and Higher Initial Ore Grades, Residual Grade
Capped at 0.007 toz Au/ton ore
Figure 3.6 - Time vs. Gold % Leach Extraction Using the Combined Fit Equation,
Starting with Lower, Average and Higher Initial Ore Grades, Residual Grade Capped at
0.007 toz Au/ton ore, presents the same information but with the dependent variable
being % leach extraction rather than residual gold grade. An important result is that all of
the ores follow identical recovery curves; however, the ores with average and higher
initial gold grades are not limited at the end of the 40 day leach cycle by the capped
minimum residual gold grade whereas the ore with lower initial gold grade is limited by
this constraint. The ores with average and higher initial gold grades would ultimately
become limited to the 0.007 toz Au/ton cap with continued leaching.
57
|
Colorado School of Mines
|
CHAPTER 4
RESULTS FROM LEACH EXTRACTION MODELING OF
SEGREGATED GOLD GRADE PLACEMENT,
SCENARIOS 1, 2 AND 3 RESULTS
Results obtained from the development and evaluation of the first three heap leach
scenarios (Scenarios 1, 2 and 3) described in Chapter III is presented here in Chapter IV.
These include extraction results for 1) the random placement scenario; 2) the placement
scenario in which ore blocks are segregated and selectively placed by gold ore grades into
three groups (Tertiles) through sorting, and 3) the placement scenario in which ore blocks
are segregated by gold ore grades into two groups (Splits) through sorting,
4.1 Scenario 1 - Random Placement of a 9 x 9 Lift Section
A combined exponential / power model leach curve was presented in Chapter 3,
which mathematically represents leach extraction and residual gold grade that would be
achieved during a 40 day leach cycle. Application of this model provides an estimation
of residual gold grade within an ore block for each ore block in a Lift Section at specific
points in time during the leach extraction process. This model was applied to the
randomly placed ores presented in Appendices E, F and G
59
|
Colorado School of Mines
|
Table 4.1 - Randomly Placed Ores - 9 x 9 Lift Section, Residual Gold Grade (toz Au/ton
ore) After 40 Days of Leach, Minimum Residual Grade Capped at 0.007 toz Au/ton
Leach Days =______40
9X9 RANDOM HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Lift Section Column 1 2 3 4 5 6 7 8 9
Rows
1 0.008 0.008 0.009 0.007 0.008 0.008 0.007 0.007 0.007
2 0.007 0.008 0.008 0.007 0.007 0.007 0.007 0.007 0.007
3 0.007 0.007 0.007 0.007 0.007 0.007 0.007 0.008 0.008
4 0.007 0.007 0.007 0.008 0.008 0.009 0.008 0.007 0.009
5 0.011 0.009 0.008 0.007 0.011 0.012 0.010 0.008 0.007
6 0.014 0.012 0.012 0.010 0.008 0.010 0.014 0.016 0.012
7 0.010 0.009 0.007 0.008 0.010 0.017 0.015 0.012 0.010
8 0.007 0.008 0.011 0.014 0.015 0.012 0.012 0.010 0.010
9 0.008 0.008 0.010 0.009 0.008 0.007 0.008 0.008 0.008
Column Ave 0.0087 0.0083 0.0088 0.0088 0.0090 0.0099 0.0099 0.0092 0.0087
Tertile Ave 0.0086 0.0092 0.0093
Lift Section Ave 0.0090
designates x < 0.00837 toz Au/ton, x is residual grade
4.2 Scenario 2 - Segregated Placement of a 9 x 9 Lift Section
The proposed alternative to the random placement of ore blocks presented in
Chapter III considers segregating ore blocks of different grades into three separate groups
within a Lift Section. Each “Tertile” would contain equal numbers of blocks selectively
placed in separate portions of the Lift Section. This produces a low-grade group - Tertile
1, a mid-grade group - Tertile 2 and a high-grade group - Tertile 3.
61
|
Colorado School of Mines
|
Table 4.2 - Block Removal and Segregated Heap Placement Sequence in Rows of 3 Ore
Columns per Tertile Row, 3 Tertile Rows per Lift Section Row - Level 1
81 Low 81 Mid 81 High
Grade Grade Grade
Sorted Sorted Sorted
Placement Block Block Placement Block Block Placement Block Block
3 Blocks Removal Gold 3 Blocks Removal Gold 3 Blocks Removal Gold
per Section Sequence Grade per Section Sequence Grade per Section Sequence Grade
1 13 0.027 1 4 0.035 1 1 0.052
2 17 0.027 2 7 0.038 2 2 0.044
3 20 0.029 3 8 0.038 3 3 0.058
1 21 0.029 1 9 0.032 1 5 0.042
2 22 0.024 2 10 0.034 2 6 0.046
3 23 0.025 3 14 0.036 3 11 0.047
1 24 0.028 1 15 0.039 1 12 0.044
2 25 0.023 2 16 0.039 2 26 0.045
3 28 0.025 3 18 0.035 3 27 0.040
1 29 0.028 1 19 0.034 1 33 0.045
2 30 0.027 2 32 0.039 2 34 0.051
3 31 0.024 3 36 0.038 3 35 0.040
1 51 0.028 1 40 0.037 1 37 0.053
2 73 0.023 2 45 0.039 2 38 0.049
3 74 0.024 3 52 0.031 3 39 0.042
1 75 0.025 1 57 0.031 1 41 0.059
2 76 0.023 2 58 0.035 2 42 0.059
3 77 0.023 3 59 0.036 3 43 0.052
1 78 0.024 1 64 0.035 1 44 0.043
2 79 0.025 2 83 0.030 2 46 0.084
3 80 0.028 3 84 0.036 3 47 0.065
1 81 0.025 1 85 0.036 1 48 0.062
2 82 0.024 2 88 0.031 2 49 0.056
3 86 0.026 3 89 0.036 3 50 0.042
1 87 0.026 1 90 0.039 1 53 0.087
2 92 0.024 2 91 0.030 2 54 0.069
3 93 0.029 3 94 0.034 3 55 0.067
63
|
Colorado School of Mines
|
other tertiles (in the case of the 8 x 10 Lift Section, one Split must be leached
independently of the other). Other than the ability to sequence leach solution application
over tertiles in progressive Lift Sections, such a leach solution application system is
essentially identical to one that is capable of only servicing an entire Lift Section at a
time, the current practice in industry. The modified leach solution application system
would be rotated or switched onto fresh ore three times more frequently when treating the
tertiles in comparison to leaching an entire Lift Section at a time. The area placed under
leach; however, would be reduced to that covering one tertile, i.e., one-third the area of
an entire Lift Section.
This leads to the development of a “Tertile Leach Day” which is, by definition,
equivalent to one-third of a Lift Section Leach Day, providing one day of lixiviant to only
one tertile. The leaching of one complete Lift Section therefore requires 120 Tertile
Leach Days, which is equivalent to 40 Lift Section Leach Days.
Ultimately, the leaching of tertiles compared to the leaching of entire Lift Sections
allows for individual treatment of ores segregated into tertiles while physically involving
essentially the same equipment and operating labor. The author has termed this technique
“Variable Leach Duration Sequencing (VLDS)”.
Table 4.5 - Segregated Placement - 9 x 9 Lift Section Residual Gold Grade (toz
Au/ton ore) After 29, 39 and 52 Days of Leach in Tertiles 1 through 3, Respectively,
Grade Capped at 0.007 toz Au/ton, summarizes the results obtained from leaching the
segregated placement - 9 x 9 Lift Section for the days indicated.
70
|
Colorado School of Mines
|
Table 4.5 - Segregated Placement -9 x9 Lift Section Residual Gold Grade (toz Au/ton
ore) After 29, 39 and 52 Days of Leach in Tertiles 1 through 3, Respectively, Grade
Capped at 0.007 toz Au/ton
Leach Days = t1 = 29 t2= 39 t3= 52
9X9 SORTED HEAP PLACED LEFT TO RIGHT, BACK TO FRON1F, LAYER ON LAYER
Lift Section Column 1 2 3 4 5 6 7 8 9
Rows
1 0.007 0.007 0.007 0.008 0.007 0.007 0.010 0.009 0.011
2 0.008 0.007 0.007 0.007 0.007 0.007 0.012 0.011 0.009
3 0.007 0.007 0.007 0.008 0.007 0.007 0.007 0.008 0.010
4 0.007 0.008 0.008 0.007 0.008 0.007 0.013 0.013 0.011
5 0.007 0.007 0.007 0.007 0.008 0.007 0.009 0.009 0.010
6 0.007 0.007 0.007 0.007 0.007 0.007 0.012 0.009 0.008
7 0.007 0.007 0.007 0.007 0.007 0.007 0.008 0.011 0.010
8 0.007 0.007 0.007 0.007 0.007 0.007 0.010 0.010 0.008
9 0.007 0.007 0.008 0.008 0.008 0.007 0.009 0.010 0.010
Column Ave 0.0073 0.0072 0.0074 0.0074 0.0074 0.0073 0.0099 0.0100 0.0096
Tertile Ave 0.0073 0.0073 0.0099
Lift Section Ave 0.0082
designates x < 0.00837 toz Au/ton residual ore grade
A comparison between leaching the randomly placed ores for 40 days and
leaching the segregated ores for 29, 39 and 52 days in Tertiles 1 through 3 respectively is
presented in Table 4.6 - Random versus Segregated Leach Comparison, Residual Gold
Grade (toz Au/ton ore) - 9 x 9 Lift Section, Tertiles leached for 29, 39, and 52 Days vs.
40 Day Lift Section Leach. This comparison between Segregated and Random Lift
Sections indicates that the leaching of segregated ores attains a 9.69% lower overall
residual gold grade with an overall extraction increase of 2.13%. A substantially greater
number of Tertile Columns in Tertiles 1 and 2 are leached below the 40-day residual gold
grade of 0.00837 toz Au/ton when the ores are segregated and leached using VLDS
versus random placement and 40 days of leaching on an entire Lift Section. Further, note
the relatively high residual gold grade that remains in Tertile 3.
71
|
Colorado School of Mines
|
Table 4.6 - Random versus Segregated Leach Comparison, Residual Gold Grade (toz
Au/ton ore) -9x9 Lift Section, Tertiles leached for 29, 39, and 52 Days vs. 40 Day Lift
Section Leach
9X9 SORTED hEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = t1= | 29 t2= 39 t3= 52
Column Ave 0.0073 0.0072 0.0074 0.0074 0.0074 0.0073 0.0099 0.0100 0.0096
Tertile Ave 0.0073 0.0073 0.0099
Lift Section Ave 0.0082
I
9X9 RANDOM HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = 40
Column Ave 0.0087 0.0083 0.0088 0.0088 0.0090 0.0099 0.0099 0.0092 0.0087
Tertile Ave 0.0086 0.0092 0.0093
Lift Section Ave 0.0090
designates x < 0.00837 toz Au/ton residual ore grade
The higher-grade ores in Tertile 3 are separate from those in Tertiles 1 and 2 and
can therefore be leached for some additional time as necessary to achieve improved
economic recovery as determined by mine management. Given the same economics (i.e.,
discounted cash flow rate of return) as appropriate for leaching of the randomly placed
ores after 40 days, leaching should continue in Tertile 3 until a leach rate of 0.000135 toz
Au/ton day is reached. This rate occurs at approximately 58 Tertile Leach Days. To
obtain the additional Tertile Leach Days (58 - 52 = ), total lixiviant to the tertile would
6 6
need to be increased by approximately 5% / 120 Tertile Leach Days increase) to allow
(6
58 rather than 52 Tertile Leach Days of Tertile 3.
Additional calculations were performed to investigate the same scenario if 64
Tertile Leach Days were used, to bring the final residual gold grades in Tertile 3 nearer
those in Tertiles 1 and 2. This would represent a 10% increase in lixiviant to the tertile.
Table 4.7 - Random versus Segregated Leach Comparison, Residual Gold Grade (toz
72
|
Colorado School of Mines
|
Au/ton ore) - 9 x 9 Lift Section, Tertiles leached for 29, 39, 52, 58 and 64 Days, vs. 40
Day Lift Section Leach, compares these scenario to the 40 day leach of randomly placed
ores.
Table 4.7 - Random versus Segregated Leach Comparison, Residual Gold Grade (toz
Au/ton ore) -9 x 9 Lift Section, Tertiles Leached for 29, 39, 52, 58 and 64 Days vs. 40
Day Lift Section Leach
9X9 SORTED HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = Tertile 1 = 29 Tertile 2 = 39 Tertile 3 = 52
Column Ave 0.0073 0.0072 0.0074 0.0074 0.0074 0.0073 0.0099 0.0100 0.0096
Tertile Ave 0.0073 0.0073 0.0099
Lift Section Ave 0.0082
Leach Days = Tertile 1 = 29 Tertile 2 = 39 Tertile 3 = 58
Column Ave 0.0073 0.0072 0.0074 0.0074 0.0074 0.0073 0.0091 0.0091 0.0088
Tertile Ave 0.0073 0.0073 0.0090
Lift Section Ave 0.0079
Leach Days = Tertile 1 = 29 Tertile 2= 39 Tertile 3 = 64
Column Ave 0.0073 0.0072 0.0074 0.0074 0.0074 0.0073 0.0085 0.0084 0.0082
Tertile Ave 0.0073 0.0073 0.0084
Lift Section Ave 0.0077
9X9 RANDOM HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = 40
Column Ave 0.0087 0.0083 0.0088 0.0088 0.0090 0.0099 0.0099 0.0092 0.0087
Tertile Ave 0.0086 0.0092 0.0093
Lift Section Ave 0.0090
designates x < 0.00837 toz Au/ton residual ore grade
Comparison of the Segregated to Random Lift Section with yet an additional
6
days of leaching in Tertile 3 beyond the originally scheduled 52 day leach duration (to 58
days total), indicates that segregation with additional leaching attains an overall residual
gold grade approximately 12.9% lower than random placement with an overall leach
extraction increase of 2.8%. An additional 12 days of leaching in Tertile 3 beyond the
73
|
Colorado School of Mines
|
Table 4.11 - Segregated Placement -8x10 Lift Section Residual Gold Grade (toz Au/ton
ore) After 27 and 53 Days of Leach in the Low-Grade and High-Grade Splits,
Respectively, Grade Capped at 0.007 toz Au/ton
8X10 SORTED HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
ON LAYER
Leach Days = Split 1 = 27 Split 2 = 53
Lift Section Column 1 2 3 4 5 6 7 8
Rows
1 0.009 0.009 0 008 0.007 0.007 0.007 0.007 0.007
2 0.008 0.008 0.007 0.007 0.007 0.007 0.007 0.007
3 0,008 0.008 0.008 0.009 0.009 0.010 0.009 0.008
4 0.008 0.008 0.008 0.007 0.008 0.007 0.008 0.009
5 0.008 0.007 0.008 0.008 0.008 0.008 0.007 r 0.008
6 0.008 0.009 0.009 0.008 0.007 0.009 0.010 0.008
7 0.009 0.009 0.008 0.008 0.011 0.013 0.013 0.014
8 0.008 0.008 0.008 0.008 0.014 0.013 0.010 0.010
9 0.008 0.009 0.008 0.008 0.007 0.007 0.007 0.007
10 0.008 0.009 0.009 0.008 0.007 0.007 0.007 0.007
Column Ave 0.008 0.008 0.008 0.008 0.009 0.009 0.009 0.008
Tertile Ave 0.0081 0.0086
Lift Section Ave 0.0083
designates x < 0.00837 toz Au/ton residual ore grade
A comparison between leaching the randomly placed ores for 40 days and
leaching the segregated ores for 27 and 53 days in the Lift Section Splits is presented in
Table 4.12 - Random versus Segregated Leach Comparison, Residual Gold Grade (toz
Au/ton ore) - 8 x 1 0 Lift Section, Splits Leached for 27 and 53 Days, vs. 40 Day Lift
Section Leach.
Comparison of the Segregated to Random Lift Section for the Split scenario
indicates that segregation attains an overall residual gold grade approximately 4.62%
lower than random placement with an overall leach extraction increase of 0.98%. An
important observation is that more Split Lift Section Columns are leached below the 40-
day residual gold grade of 0.00837 toz Au/ton when the ores are segregated and leached
80
|
Colorado School of Mines
|
using VLDS then when random placement and a subsequent 40 days of leaching is
employed.
Table 4.12 - Random versus Segregated Leach Comparison, Residual Gold Grade (toz
Au/ton ore) -8x10 Lift Section, Splits Leached for 27 and 53 Days, vs. 40 Day Lift
Section Leach
8X10 SORTED HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = t1 = 27 t2= 53
Column Ave 0.008 0.008 0.008 0.008 0.009 0.009 0.009 0.008
Tertile Ave 0.0081 0.0086
Lift Section Ave 0.0083
8X10 RANDOM HEAP PLACED LEFT TO RIGHT, BACK TO FRONT, LAYER ON LAYER
Leach Days = 40
Column Ave 0.009 0.010 0.009 0.009 0.008 0.008 0.007 0.008
Tertile Ave 0.0096 0.0079
Lift Section Ave 0.0087
designates x < 0.00837 toz Au/ton residual ore grade
The reduced enhancement in leach extraction due to segregation into splits rather
than tertiles (halves rather than thirds) was anticipated given that better metallurgical
results are commonly obtained when ores are selectively treated for their specific
characteristics (Miller, G, 1998). Segregating into tertiles rather than splits allows more
selective treatment of the ores thereby comparatively improving on overall leach
extraction.
81
|
Colorado School of Mines
|
5.1 Scenario 4 - Tertile Averaging of Cyanide Consumption in Ore Columns
Segregated by Gold Grade
A method to reduce the possibility of inadvertently stacking higher-grade
deleterious copper (i.e., with higher cyanide consumptions) in ore columns when
segregating gold grades for placement on a heap was developed. An integer
programming method to determine the optimal sequential placement of ore blocks to
achieve the best overall average cyanide consumption (lbs CN/ton ore) within the Tertile
Rows was formulated.
In every Tertile Row, each 3 Levels deep, there are 9 ore blocks making up 3 ore
columns. This is depicted in Figure 5.1 - Generic Tertile Rows, Location of Blocks.
Xjj Block Location
X X X
31 32 33
X X X
21 22 23
Xu X X
12 13
Figure 5.1 - Generic Tertile Rows, Location of Blocks
In Figure 5.1, the ore columns, from the bottom-up, consist of blocks Xu, X and
21
X ; blocks X , X and X ; and blocks X , X , and X . The overall objective is to
31 12 22 32 13 23 33
determine the best location to place each block in each ore column so as to yield ore
column cyanide consumptions (in lbs CN/ton ore) closest to the overall Tertile Row
average cyanide consumption, thereby uniformly minimizing the possibility of negatively
influencing gold leach extraction by preventing high concentrations of cyanide soluble
copper from occurring in any one ore column. The integer programming formulation for
this problem is given in Table 5.2 - Formulation for Tertile Row Averaging.
85
|
Colorado School of Mines
|
The objective function, to minimize Mi + M2 + M3 +.... +M36, Where Mi is
defined as the root-mean-square of the differences between the ore column averages and
the Tertile Row average for each permutation of the “ith” ore column of 9 stacked ore
blocks in a Tertile Row. This is subject to constraints allowing only 3 ore columns in a
Tertile Row, no duplicates allowed.
Mi can be found by first calculating ore column cyanide consumption averages for
all possible ore column combinations followed by solving for the lowest root-mean-
square average of the difference between the ore column averages and the Tertile Row
averages. The number of possible combinations of ore blocks that must be considered is
3 x 3 x 3 = 27. However, given that the blocks cannot be placed on any other level than
the level on which they sequentially occur as delivered from the mine, and that the
arrangement of the blocks on the lowest level is inconsequential to the averaging that is
to occur through optimized placement of the blocks in the upper 2 levels, the number of
possible permutations is 3! for the second level times 3! for the third level; the total
number of permutations therefore being 3 ! X 3 ! = 36.
86
|
Colorado School of Mines
|
- Averaged Cyanide Consumption (lbs CN/ton ore) by Tertile Row. Table 5.3 of the same
title presents a summary of the Tertile Row Averages from this Appendix.
The number of ore columns expected to consume between 1.8 and 2.0 lbs CN/ton
ore due to deleterious cyanide soluble copper have increased from 25 in both the random
and segregated (by gold grade) placement scenarios to 35 after averaging. However, the
number of ore columns expected to consume between 2.0 and 2.2 lbs CN/ton ore have
decreased from 12 in the random case and 13 in the segregated (by gold grade) case to 5
after averaging. Further, after averaging, there are no ore columns with a greater
consumption than 2.2 lbs CN/ton ore as encountered in the random placement scenario.
Averaging in this way allows leach solution to be applied consistently on the tertiles
without concern of exhausting cyanide concentration, due to high deleterious cyanide
soluble copper content, toward the bottom of an ore column during leaching. Averaging
thereby also prevents large unwanted increases in copper cyanide concentrations in the
pregnant solution feed to the recovery plant.
90
|
Colorado School of Mines
|
CHAPTER 6
SUMMARY OF THE RESEARCH RESULTS
A model of a generic gold mining operation was developed based on ore grades
from an actual deposit. The objective in developing this model was to prepare a basis
upon which comparisons in metallurgical leach extraction could be made between ores
placed randomly and ores placed selectively. A substantial amount of information
regarding mine and process operations was not included in the model providing focus on
the placement of ore blocks while preventing dilution of critical mine and heap criteria
with superfluous detail. All necessary and sufficient design criteria and information is
provided to make the appropriate comparisons.
Presented in this model is a mine plan for one bench level, a mine schedule
detailing ore production, mine equipment identification, mine operation plans, heap leach
design criteria, heap construction plans, leach extraction analysis and the salient ore
characteristics of gold grade and cyanide consumption. Constraints applied to gold grade
and cyanide consumption form the basis upon which selective placement was determined.
It should be noted that these mine and heap design, construction and operating
parameters would commonly be determined during mine feasibility studies and final
engineering including the results from all appropriate geotechnical and metallurgical
testing for a specific deposit and location. Although the parameters in this model appear
to be simplified and generalized, they are actually to the level of detail used by industry
to design and construct such operations.
97
|
Colorado School of Mines
|
Development of the five scenarios in the previous chapters has yielded results
demonstrating benefits attributable to selective ore placement and ore treatment during
heap construction and leaching operations, respectively.
6.1 Scenario 1
A representative Lift Section of a heap was modeled by randomly placing 243 ore
blocks in a configuration 9 rows (Lift Section Rows) long, 9 columns (Lift Section
Columns) wide, and 3 levels deep. The average gold ore grades of the ore columns
formed by this Lift Section construction technique are given in Table 3.6 - Average Ore
Column Gold Grades (toz Au/ton) of Blocks Randomly Placed on the Lift Section 9
Blocks per Row, 9 Rows and 3 Levels Deep. Metallurgical leach extraction curves were
developed and used to estimate residual ore grade on this Lift Section after 40 days of
leaching. Leach extraction model results indicate leach extraction of 78.1% of the gold in
the Lift Section and an overall leach rate of 0.00014 toz Au/ton day at cessation of
leaching at 40 days.
6.2 Scenario 2
A Lift Section identical in size and shape to that in Scenario 1 is considered in
Scenario 2 wherein the same 243 ore blocks are segregated into 3 equal Tertile groups by
gold grade. As was expected, leaching of the entire Lift Section for 40 days resulted in
the same leach extraction as that achieved by the Lift Section in which the blocks were
placed randomly.
Segregated placement presents the advantage of allowing discreet leaching to
occur for each specific tertile group. Exploiting this advantage, it was demonstrated that
leach extraction in Tertiles 1 and 2 (the low-grade and mid-grade ores, respectively), was
98
|
Colorado School of Mines
|
as complete after 29 and 39 days of lixiviant application, respectively, as would be
achieved from these ore grades after a full 40-day leach period. Further, reducing the
total leach days for these two lower grade Tertiles allowed for an additional 12 days of
lixiviant application (for a total of 52 Tertile Leach Days) on Tertile 3. On this basis, the
concept of variable leach duration sequencing (VLDS) was developed in which each of
the tertiles are leached for different durations depending on their respective leach
requirements. The leaching of Tertiles 1 through 3 for 29, 39 and 52 days, respectively,
resulted in an overall leach extraction of 80.2%, an improvement of 2.1% over the leach
extraction achieved from the randomly placed Lift Section.
Beyond this improvement, it was shown that the ore blocks in Tertile 3 still
contained sufficient gold grade for economic leaching to continue beyond 52 days, an
opportunity that would not be available if the ores had been placed randomly. Extending
the leach period in Tertile 3 by an additional 6 and an additional 12 Tertile leach days
improved overall leach extraction to 80.9% and 81.4%, respectively. In this way, an
overall improvement in leach extraction of 3.3 % was achieved by segregated placement
and VLDS.
6.3 Scenario 3
A Lift Section containing 240 ore blocks placed in a configuration 8 rows (Lift
Section Rows) long, 10 columns (Lift Section Columns) wide, and 3 levels deep was
modeled to approximate the previously modeled Lift Section containing 243 ore blocks
described in Scenario 2. The three ore blocks removed from the original group of 243
had gold grades nearest the mean (two immediately below and one immediately above).
The objective of developing the 8 x 10 Lift Section was to compare segregation and leach
extraction from the Splits to segregation and leach extraction from the Tertiles.
99
|
Colorado School of Mines
|
Leach extraction from the 8 x 10 Lift Section was the same at 40 days as that
obtained from the 9 x 9 Lift Section. Leaching of the lower and higher grade splits for 27
and 53 days, respectively, resulted in a 1.0% leach extraction increase to 79.1%; a smaller
increase in comparison to the 2.1% leach extraction increase to 80.2% achieved by
segregated placement and leaching of the tertiles. This result demonstrates that
additional segregation from Splits to Tertiles subsequently followed by VLDS for the
segregated ore columns provides additional extraction. More generally, the greater the
opportunity to treat a specific ore characteristic, the greater the ultimate opportunity for
enhanced return on the effort invested.
6.4 Scenario 4
Deleterious cyanide soluble copper, inherent to the gold ore, can consume
significant amounts of cyanide (lbs CN/ton ore) during the leach extraction process.
Although acceptable up to levels of 2.0 lbs CN/ton ore consumption, levels above this
can reduce leach extraction by weakening the concentration of the lixiviant and can cause
copper loading problems during metal recovery in the process plant. Cyanide
consumption, like gold grade, is an ore characteristic not constant throughout the ore
blocks in the modeled deposit. Cyanide consumption distribution statistics for the ore
blocks were developed and presented in Table 3.3. As ore blocks are placed three levels
deep forming ore columns during Lift Section construction, there is a possibility of
inadvertently forming several ore columns high in cyanide consumption.
An integer programming technique to reduce the random possibility of placing
several ore blocks inherently high in cyanide consumption on top of one another was
developed for application to the tertile rows (three blocks wide). A Lift Section identical
to that in Scenario 2, wherein the 243 ore blocks were to be segregated by gold grade for
placement into 3 equal tertile groups, was further evaluated for cyanide consumption.
100
|
Colorado School of Mines
|
The segregated placement by gold grade into the tertiles was maintained while cyanide
consumption was averaged through selective placement across the tertile rows. This
prevented the inadvertent random formation of ore columns substantially higher than
average in cyanide consumption.
Applying this technique increased the number of ore columns expected to
consume between 1.8 and 2.0 lbs CN/ton ore 40% from 25 to 35 after averaging in both
the random and segregated (by gold grade) placement scenarios. The number of ore
columns expected to consume between 2.0 and 2.2 lbs CN/ton ore; however, decreased
by 60% from 12 in the random case and 13 in the segregated (by gold grade) case to 5
after averaging. There were no ore columns with greater than 2.2 lbs CN/ton ore cyanide
consumption after averaging the tertile rows as were encountered in the random
placement scenario.
6.5 Scenario 5
The need to average deleterious content in ore columns through selective
placement may be required regardless if ores are to be segregated or not. Formulation of
an integer programming method to average across entire Lift Section Rows (9 blocks
wide) was considered. It was quickly realized; however, that the effort needed to
determine the optimum solution would be substantial while the useful value of such a
solution would be no better than accepting a slightly less optimal solution; the ability and
accuracy of estimating cyanide consumption in actuality being far less than perfect.
A simple heuristic approach was developed to determine near optimal averaged
placement wherein the Lift Section Row composed of 3 levels of 9 ore blocks each (27
ore blocks total) to be placed on top of one another are sorted in ascending order. The
level with the greatest range in cyanide consumption values remains in the sorted
101
|
Colorado School of Mines
|
sequence generated and will be placed in this order. The level with the second largest
range of cyanide consumption is sequentially reversed (sequencing this row in
descending order) and will thereby be placed with block values from high to low cyanide
consumption, diametrically opposed to the placement in the first level from low to high.
The cyanide consumption values of these two rows are then added. The final level to be
considered, with the smallest range of cyanide consumption, is added where necessary to
minimize the range of cyanide consumption values when all three blocks making up each
ore column are averaged. In this way a lift section row is averaged across all nine ore
columns. It is important to recall during this solution that block placement may be
manipulated within the rows but not between levels, these being dictated by the mine
block removal sequence.
A substantial improvement in averaging cyanide consumption in ore columns
across the entire Lift Section was achieved through the application of this technique. The
number of ore columns expected to consume between 1.8 and 2.0 lbs CN/ton increased
from 25 in both the random and segregated placement scenarios to 34 after averaging
across the 9 Lift Section Rows. However, there are now no ore columns expected to
consume greater than 2.0 lbs CN/ton ore after Lift Section averaging, a total elimination
of the 13 ore columns exhibiting this amount of cyanide consumption when placed
randomly. Considering that 5 ore columns remained above 2.0 lb CN/ton cyanide
consumption in the optimal solution obtained for tertile row averaging while none remain
here, it incorrectly appears that this solution outperforms the optimal solution. The
averaging across all nine ore columns in a Lift Section Row; however, is a different
problem than averaging of tertile rows, the former allowing more flexibility to achieve a
better average.
These various techniques and analysis may be applied to any combination of
selective ore placements, ore characteristics, leach extraction rates and lixiviant
variations. More accurate information regarding these characteristics and parameters
delivers more accurate results.
102
|
Colorado School of Mines
|
CHAPTER 7
ADVANTAGES OF SELECTIVE ORE PLACEMENT APPLIED
DURING HEAP CONSTRUCTION AND OBSERVATIONS
PERTAINING THERETO
Selective placement of ores on a heap leach pad is the final opportunity mine
operations have to assist and enhance the leach extraction process before actual leach
operations commence. The concept of selective ore placement contrasts current state-of-
the-art practice whereby ores are placed randomly on a heap in the order received from
the mine at locations decided by the mine equipment operators “on the fly". Use of
selective placement has the potential to improve heap leach consistency, predictability
and overall metallurgical leach extraction. This is accomplished by precision placement
of ores graded by characteristic in appropriate groups and combinations for targeted
leaching opportunities with alternative lixiviant suites and concentrations. Further,
selective placement, when applicable and properly employed, reduces both solid and
liquid inventory by providing a more efficient first leach cycle resulting in higher overall
leach extractions.
All components necessary to enable selective placement are commercially available
and often already in use at mine sites. Until the early 1990’s, conventional mine survey
using theodolite and stadia rod was employed exclusively to spatially locate ores within
an open pit. With launch and orbit insertion of the 24th and final Navigation Satellite
Timing And Ranging Global Positioning System (Navstar) satellite on January 17th,
103
|
Colorado School of Mines
|
1994, the global positioning system (GPS) became fully operational. The system was
made available for commercial use in 1996.
GPS allows precision real time location of equipment and ore blocks within an open
pit mine to accuracies within 2 cm horizontal and 10 cm vertical (Leica, 2007). Further,
GPS is commonly installed on each piece of ore contacting equipment in a mine fleet and
can be used to direct such a fleet in real time during mine operations. GPS is the
enabling technology allowing selective placement to occur simply and economically in
real time.
Mine planning has also advanced substantially in the late 20th century since the
advent of the computer and the development of mine planning software. Until this time,
drill data interpretation, block modeling and resource/reserve estimation were performed
through rigorous and repetitive calculation in conjunction with hand drawn graphic
support including statistical analysis, geometric modeling and cartographic presentation
techniques. Mine planning software today provides quick, accurate and affordable
methods to store and analyze data, model a deposit, optimize a reserve, prepare a mine
plan and track the location of ore and waste materials as mining progresses when used in
conjunction with GPS. Proper application of geostatistics and optimization routines
allow mine operators to evaluate a multitude of physical mine plans and operating
scenarios before mining commences, ultimately allowing the most profitable mining
opportunity to be identified and pursued.
Several techniques with potential to improve leach extraction from heap leach
operations through selective placement and treatment of ores were investigated. Design
criteria for pertinent components of an open pit mine, heap leach pad and process plant
were developed, allowing various selective ore placement and leach extraction scenarios
to be modeled. Ore block grade and size data for a mine bench was obtained for a real
yet undisclosed gold deposit in the American Southwest. A mine schedule and ore block
removal/heap delivery sequence was developed from this bench data. Heap lift section
configurations containing 240 to 243 blocks were developed in which the ore
104
|
Colorado School of Mines
|
characteristics of gold grade (toz Au/ton ore) and cyanide consumption (lbs CN/ton ore)
due to deleterious cyanide soluble copper were spatially retained for each ore block
within the model. Combined exponential and power decay functions representing gold
leach extraction curves were developed from which residual gold grades in the ores could
be determined for any given time within the leach cycle.
Five scenarios were developed using this model wherein specific consideration was
giving to ore placement and resulting leach extraction. The first three scenarios
compared leach extraction results from randomly placed ores with various grades to those
ores segregated by grade into two and three portions (splits and tertiles). The last two
scenarios focused on the averaging of deleterious constituents specifically within the
segregated groups of ore and within an entire lift section.
Varying the duration of lixiviant application to the segregated portions of the lift
section, as appropriate for the gold grade contained within that portion, increased
modeled leach extraction. The term “variable leach duration sequencing or VLDS” was
coined to represent this process. Leach duration was proportionally extended for higher-
grade heap volumes and decreased for lower-grade heap volumes to the extent that leach
rate for the specific heap volume at cessation of leaching that volume was approximately
equal to that of the overall leach rate for the randomly placed heap at 40 days. In this
way, leach extraction from any heap volume resulting from segregated ore placement was
always maintained at or above the economic bounds accepted for a randomly placed
heap. Further, VLDS provides a hydrostatic head in adjacent portions of ore under leach,
thereby driving solution away from longer leached ore and forcing the fresh solution to
collection at the base of the heap. This is different than typical leaching in which there is
no resisting hydrostatic head, allowing fresh leach solution to percolate back into
previously leached material.
Beyond allowing equal amounts of lixiviant to be applied in both the random and
segregated scenarios, it was demonstrated that residual ore grades in the higher-grade
heap volumes were sufficient to allow additional leaching to be performed on an
105
|
Colorado School of Mines
|
economically viable basis. Additional leaching of these higher-grade heap volumes was
again justified by accepting that the final leach rate for a segregated heap volume could
be equivalent to the overall leach rate for the randomly placed heap at 40 days. It is
important to observe that although division of the lift section into thirds (tertiles) was the
method used herein, division into halves (splits) already showed improvement. Further
improvements may be achieved by segregation of ores into even more groups provided
that mine and heap leach operations are not logistically impacted to the point at which the
cost of segregating and leaching the ores is greater than the value of the increased
extraction.
The problem of unintentionally constructing lift sections with multiple ore columns
bearing inordinately high deleterious cyanide soluble copper content was overcome by
averaging the cyanide consumption characteristic of the ore through selective placement
during lift section construction. This was performed for both the tertile rows in the
segregated heap volumes and for entire lift section rows for the case when segregation by
gold grade was not used. A substantial reduction in the number of ore columns with
average cyanide consumption above 2.0 lbs CN/ton ore was achieved. As expected,
cyanide consumption averages increased slightly in many of the ore columns as the ore
blocks with characteristically higher cyanide consumption values were redistributed
throughout all the ore columns within the lift section.
Averaging of the deleterious cyanide consumption may be considered analogous to
the blending of ores in milling operations; blending being a long understood and well
established technique in the industry. Conversely, segregation of ores by characteristic
may have beneficial applications in milling operations. Commonly, ores are blended in
milling operations to provide consistent feed for the subsequent physical and/or chemical
extraction and recovery processes employed after mineral liberation. There are; however,
occasions when, due to the established mine plan, there is insufficient availability of ores
with specific characteristics to blend, thereby forcing mill operations to process "off-
blend” material not optimal for the current operations and processes. By segregating ores
106
|
Colorado School of Mines
|
with such characteristics, milling operations may be adjusted to optimize recovery from
each specific “off-blend” ore mixture, possibly providing a higher overall extraction
and/or recovery alternative. Unfortunately, much more effort is typically required to
modify milling operations thereby precluding such efforts in favor of increased
negotiations with mine operations to provide a more favorable blended suite of ores to the
mill.
There is an obvious but less quantifiable aspect to segregating by ore characteristic.
In the case of ore grade, as developed herein, the high-grade portion of ore is segregated
into a volume one-third that of the entire heap volume allowing more aggressive leach
extraction techniques to be employed on this material. Although not discussed or
considered herein, there are several other techniques that may be applied during the leach
cycle beyond additional leach time. These other techniques include chemical
modifications such as increased reagent concentration and/or additions/modifications to
the lixiviant content and geotechnical modifications such as increased lixiviant
application rate and or surging of lixiviant to promote imbibition and drainage. The use
of surfactant to modify lixiviant surface tension is a chemical modification that affects
heap geotechniques.
Selective placement of ores by ore characteristic on a leach pad requires a concerted
effort by both mine and process personnel. Most mines already use the GPS and mine
software components needed to enable selective placement and could easily adapt these
systems for that purpose. Gold grade and cyanide consumption results derived from
operational sampling and analyses suites that are commonly used to control heap leach
operations would provide a sufficient basis upon which to employ the selective placement
techniques demonstrated herein. Enhanced sampling and analysis techniques should be
developed to complement and/or replace current methods. Such techniques would need
to be both accurate and timely, thereby allowing selective ore placement decisions to be
made in confidence and within the required mine/load/haul/dump time frame. These may
include the use of in-line or on-stream x-ray diffraction, radioactive density
107
|
Colorado School of Mines
|
measurement, photographic particle size analyses, mineralogical examination, quick
permeability tests, reagent shake tests, gravity tests, other wet chemistry tests and
possibly determination of rock quality designation (RQD), as may be provided real time
by Caterpillar’s Aquila Drill Management System, when correlated to drill logs and the
geologic record. Accurate results from such analysis, if consistently obtainable within
the limited time allowed, would provide even greater opportunities to improve leach
extraction when selective placement techniques are used. Further, once selective
placement is adopted and in place, there are essentially no additional costs associated
with using this technique.
Selective heap placement scenarios beyond simple segregation and averaging may
further improve heap extraction. Material may be stockpiled, reclaimed and delivered to
a leach pad so as to perform similar optimization on crushed ore products. Other
techniques may include segregating ore blocks into four or more groups (rather than three
or less groups demonstrated herein) depending on the operating objectives and
constraints. Opportunities include the placement of low permeability material all one
area, possibly with the lowest grade material. Alternatively, it may be determined that
sequentially placing low permeability materials in a checkerboard design allows for the
best solution flow and leach extraction.
Regardless of the improvements achieved through selective placement and targeted
leaching, the selective ore placement process is still dependent to a great extent on the
sequence in which mine blocks are delivered to the heap. Advanced understanding of ore
block characteristics and their interaction during the leach cycle could provide useful
information to mine management whereby a mine plan may be developed dedicated to
providing an ore block sequence that supports selective placement. Such interaction
between mine and heap operations would provide the basis wherein truly optimal leach
extraction may be achieved.
It must be recognized that a Lift Section is only a small portion of a heap and that
several lifts are ultimately placed on top of one another to form the heap. Some
108
|
Colorado School of Mines
|
CHAPTER VIII
CONCLUSIONS AND RECOMMENDATIONS
Selective placement of ores during heap construction has the potential to improve
heap leach consistency, predictability and overall metallurgical leach extraction. When
applicable and properly applied, selective placement reduces both solid and liquid
inventory by providing a more efficient and higher yielding first leach extraction cycle.
Equipment, systems and software currently used by the mining industry fully support
selective placement techniques. Enabling technologies include use of the global
positioning system to direct the mine fleet from excavation through placement of ores on
a leach pad and advanced mine software capable of manipulating and storing information
about ore characteristics and location, both before excavation and after placement.
The hypothetical design criteria developed for pertinent components of an open pit
mine to be operated in conjunction with a heap leach pad and process plant were used to
model various selective ore placement and leach extraction scenarios. This included the
development of a mine plan and ore block excavation and heap delivery schedule. An
overall mathematical heap leach extraction profile was also developed to reflect the
residual gold grades in the ores during leaching. In all cases, actual data could be directly
substituted for the hypothetical and results relating to specific ores and leach parameters
could be quickly developed.
To evaluate the effectiveness of selective placement in comparison to random
placement with regard improve leach extraction and efficiency several alternate selective
placement scenarios were developed and tested using these design criteria. Alternative
111
|
Colorado School of Mines
|
scenarios demonstrated include segregation of ores by grade and averaging of ores by
their deleterious characteristics.
Leach parameters to improve, if not optimize leach extraction were developed for
consideration for each selectively placed volume of ore. Variable Leach Duration
Sequencing is one such application proposed herein through which higher-grade ores are
treated for longer periods of time than would be afforded during the leaching of randomly
placed ores. It was demonstrated through these evaluations that residual ore grades in the
higher-grade heap volumes were sufficient to allow additional economic leaching thereby
improving overall leach extraction.
Selective placement techniques as applied to deleterious cyanide consumption yielded
a substantial reduction in the number of ore columns that would consume more than the
pre-determined acceptable amount of cyanide. Implications are that selective placement
used in this way would improve leach extraction through lixiviant conservation within
specific ore columns.
Ultimately, selective ore placement must be considered during mine planning so as to
optimize leach extraction by obtaining the best-sequenced ore excavation and placement
scenario possible.
Academically there are multiple opportunities for research stemming from selective
ore placement. These include the development of quick, concise and useful ore
characterization techniques; development of advanced selective placement techniques for
use on ores with specific characteristics; and development of advanced leaching
techniques designed to improve extraction, reduce costs and enhance the consistency and
predictability of heap leach operations.
Development of additional sampling and analysis techniques for gold, copper and
other ores that may complement or replace current methods should be aimed at enhancing
leach extraction and beneficiation processes for specific ores. Accurate results from such
analysis, if consistently obtainable within the limited time allowed, would provide even
greater opportunities to improve leach extraction when selective placement techniques
112
|
Colorado School of Mines
|
are used. Such analyses should be conducted on a frequency predicated by the variability
of the ore. More detail and time consuming analysis should also be performed when it is
determined that the ores to be encountered have characteristically changed enough to
potentially affect metallurgical performance.
Selective heap placement scenarios beyond simple segregation and averaging may
further improve heap extraction. Material may be stockpiled, reclaimed and delivered to
a leach pad so as to perform similar optimization on crushed ore products. Other
techniques may include segregating ore blocks into four or more groups (rather than three
or less groups as demonstrated herein) depending on the operating objectives and
constraints. Opportunities include the placement of low permeability material all one
area, possibly with the lowest grade material. Alternatively, it may be determined that
sequentially placing low permeability materials in a checkerboard design allows for the
best solution flow and leach extraction. It may also be desirable to maximize the
minimum number of blocks meeting some criteria or to minimize the maximum number
of blocks meeting some criteria. The number and variety of potentially applicable
selective placement scenarios to optimize metallurgical performance is boundless.
Advanced leach techniques that may be applied during the leach cycle include
chemical enhancements and modifications such as increased reagent concentration and/or
additions/modifications to the lixiviant content and geotechnical modifications such as
increased lixiviant application rate and/or surging of lixiviant to promote imbibition and
drainage. The use of surfactant to modify lixiviant surface tension is a chemical
modification that affects heap geotechniques.
Computerized operating systems working in conjunction with the GPS should be
developed to store, interpret and integrate the information generated from the mine plan,
block model and last minute metallurgical testwork results obtained from tests performed
on production drill cuttings. A mine operating system component capable of quickly
determining where and when ores exhibiting specific inherent characteristics are to be
placed would be interfaced with the existing system allowing implementation of the
113
|
Colorado School of Mines
|
selective placement process. Lixiviant application and leach extraction would be
monitored to quantify the applied process parameters and substantiate the leach
extractions achieved. Mine planning software, previously accommodating only mine
related information, would track both the spatial locations and characteristics of ore
blocks from mine to heap including all records of lixiviant application and leach
extraction. Topsoil and waste rock block locations, quantities and characteristics would
concurrently be tracked and stored for later use as needed for dealing with reclamation
and site-closure issues. Such mine system integration ultimately allows progressive
optimization steps including selective ore placement to be applied. As with most all
mining industry innovations, it will take the commitment of mine management and
cohesive interaction between mine and process operations to support the use of selective
placement, ultimately leading to improved metallurgical leach extraction.
For a relatively insignificant cost, mine operations have the opportunity to adapt their
current GPS and software technologies so as to integrate selective placement into their
mine planning and operating systems. Financial entities that fund mining ventures have
an opportunity to increase revenues and reduce their risk by supporting the use of
selective placement techniques when applicable. Engineering and equipment sales firms
may enhance their portfolio of capabilities and products used to win sales from mine
operators by providing engineering and equipment geared to support enhanced leach
extraction through selective placement. Ultimately, environmental groups and
government agencies should support the use of selective placement since greater metal
extraction and production efficiency reduces environmental impacts and delivers greater
real value from a given ore deposit.
Regardless of the improvements achieved through selective placement and targeted
leaching, the selective ore placement process is still dependent to a great extent on the
sequence in which mine blocks are delivered to the heap. Advanced understanding of ore
block characteristics and their interaction during the leach cycle could provide useful
information to mine management whereby a mine plan may be developed dedicated to
114
|
Colorado School of Mines
|
ABSTRACT
During emergency mine rescue operations, life-threatening situations and commu
nication problems are frequent and recurring issues. Autonomous robotic vehicles
can assist rescue workers by exploring areas that may be unsafe for rescue teams
to enter and by establishing a wireless communication network. A key trait of an
autonomous vehicle is the ability to map and localize relative to the surrounding en
vironment, using artificial or natural landmarks. One technique for localization in a
mine is to use the mine wall’s unique features as landmarks. Traditionally, a common
sensor choice for detecting these natural landmarks is a scanning laser; however, laser
sensors are unreliable in the conditions that are common in mine emergency situa
tions (e.g. thick smoke). Thus, localization and mapping algorithms using ultrasonic
range finders are examined; these sensors are impervious to thick smoke and can,
therefore, accurately range in emergency mine rescue scenarios. This thesis focuses
on the design of an autonomous robotic vehicle, along with the development of a
localization algorithm using ultrasonic range finders. Simulations of the localization
algorithm lead to the conclusion that localization using ultrasonic range finders is
possible if the mine wall texture exceeds, or is distinguishable from, the sensor noise.
|
Colorado School of Mines
|
CHAPTER 1
INTRODUCTION
Robotics technology has seen many advancements in the past 20 years; com
puters are becoming faster, allowing robots to become more intelligent and more
autonomous. One type of robotics technology is tele-operated vehicles. These vehi
cles travel in areas that are unfavorable to humans, serving as an invaluable tool in
environments ranging from search and rescue in rubble from a collapsed building to
exploring the hostile environment on Mars. Another application of these intelligent
machines can be in emergency response operations in subterranean environments,
where the air may be toxic, the ground may be flooded, or an explosion may be
imminent.
1.1 Motivation
Mining is a dangerous occupation. Since 1976, in the United States alone, there
have been 20 recorded mining disasters (defined as 5 or more fatalities) in both coal
and metal/nonmetal mines, resulting in 215 fatalities [9]. Despite advances in mine
safety, 30% of the fatalities have occurred in the past decade.
A recent example of a mining disaster occurred in 2006, when an explosion trapped
13 miners at the Sago Mine in Sago, West Virginia [27]. Toxic carbon monoxide levels
in the mine delayed rescue attempts until 12 hours after the explosion. Even after
the toxic gases diffused, rescue teams proceeded with caution, as they continually
tested for hazards such as water seeps, explosive gas concentrations, and unsafe roof
conditions. Due to multiple safety concerns, rescue workers took an additional 30
hours after the initial 12 hour delay to reach the 12 victims and single survivor,
located approximately two miles from the mine entrance [26]. On average, rescue
operations proceeded at a rate of 250 feet per hour (two miles/12 hours), leaving
1
|
Colorado School of Mines
|
room for improvement.
In dangerous situations such as these, robotic vehicles can be invaluable tools for
exploring in areas that may be unsafe for rescue teams to enter; twenty-four hours
after the initial explosion at the Sago Mine, the Mine Safety and Health Adminis
tration (MSHA) deployed an explorer robot. Equipped with a gas analyzer and live
video feedback, the explorer robot could have expedited rescue efforts if it had not
encountered technical difficulties [37].
Even in less-dangerous situations, robots can work alongside rescue workers, as
sisting in communications or carrying equipment. According to the Sago Mine Ac
cident Investigation Report, communication was a major problem plaguing rescue
operations: “The handheld radios become less reliable as the distance between users
is increased or when the users are not in direct line of sight of each other” [26]. Indeed,
radio frequency power attenuation around bends in a mine can be more than 20 times
greater than line-of-site attenuation [22]. Communication was such a problem that
losing radio contact “caused the outby team member to repeatedly wade through the
water while trying to maintain communication with the inby rescuers and outby res
cuers” [26]. Robotic vehicles can assist rescue workers by establishing an automatic,
self-configuring communication network [22].
In this thesis, we explore one aspect of the various technologies needed to demon
strate an effective robotic communication network for a mine emergency rescue sce
nario. In the remainder of the chapter, the overall project is described, then a poten
tial scenario outline to resolve the communication issue is proposed. Then, the next
section introduces an important aspect needed to implement the scenario outline:
localization. The final section of this chapter provides an outline of the thesis contri
butions, which includes part of the vehicle design, as well as a localization algorithm.
2
|
Colorado School of Mines
|
1.2 MineSENTRY Description
A project in the Center for Automation, Robotics, and Distributed Intelligence
(CARDi) Research Center, Mine Safety and Rescue Through Sensing Networks and
Robotics Technology (MineSENTRY) is a research program at the Colorado School
of Mines, with a focus on developing emergency first response robotic vehicles for
mine rescue efforts and wireless communication in subterranean environments.
1.3 Scenario Outline
In the event of a mining disaster, such as an explosion or a cave-in, MineSENTRY
proposes a wireless tethering solution to explore the mine and re-establish communica
tion. To illustrate the wireless tethering, consider the scenario depicted in Figure 1.1.
Figure 1.1: Autonomous Mobile Radio (AMR) Tethering [24]
A Bobcat front-end loader leads a caravan of Autonomous Mobile Radios (AMRs) into
the mine. Each AMR is equipped with an array of sensors for autonomous navigation,
along with a radio capable of relaying signals from a base station command center to
the Bobcat leader. Using these vehicles, an operator can clear blockages from a safe
location by tele-operating the Bobcat using live video feedback.
The tethering occurs in the sequence illustrated in Figure 1.2. In the MineSEN
TRY scenario, the leader would be the tele-operated Bobcat front-end loader, the
AMR is an autonomous vehicle carrying a radio, and the base station could be either
3
|
Colorado School of Mines
|
fixed or mobile near the entrance of the mine. All vehicles begin near the base sta
tion, with identical received signal strength (RSS). When the Leader moves, the RSS
among the radio nodes becomes unbalanced. The AMR detects this imbalance, and
automatically navigates forward until the RSS is approximately equal among radio
nodes. Equal RSS is not necessarily equivalent to equal distance. This tethering
experiment can be extended to any number of AMRs in between the base station and
the Bobcat; the AMRs would still configure themselves for equal RSS among all radio
nodes.
1. START (RSS approx equal between Base, AMR, and Leader)
Base AMR Leader
Station
2. LEADER MOVES (RSS less from AMR to Leader than AMR to base)
Base AMR Leader
Station
3. AMR MOVES (RSS approx equal between Base, AMR, and Leader)
(but might not be equally spaced in distance)
§f<5 <^V<5
Base AMR Leader
Station
Figure 1.2: AMR Tethering Sequence [24]
1.4 Autonomous Mobile Radio (AMR) Localization
In order to navigate in a mine, as depicted in the AMR tethering sequence in
Figure 1.2, autonomous vehicles rely on intelligent computer algorithms that make
decisions based on exteroceptive and propreoceptive sensor readings. One such algo
rithm is a localization algorithm; localization is the problem of estimating the pose
of a robot relative to a map—the map may be known a-priori, or the robot may use
a Simultaneous Localization and Mapping (SLAM) algorithm. The simplest form of
localization is dead reckoning, where a robotic vehicle begins with an initial position
4
|
Colorado School of Mines
|
and propagates that position forward based on odometry and heading information
(from a gyro or steering angle sensor, for example). Dead reckoning has inherent
uncertainty and drift, requiring periodic correction using exteroceptive sensors.
For example, localization in a mine using scanning laser sensors for drift cor
rection has already been demonstrated [20]. However, due to reasons detailed in
Chapter 2, ultrasonic range finders are more appropriate in emergency situations.
Though many studies utilize ultrasonic sensors for SLAM algorithms in an indoor
environment, localization using ultrasonic sensors in a mining environment is a novel
concept. Ultrasonic sensors may be able to replace scanning laser sensors in drift-
correcting localization algorithms, which would prove useful during mine disaster
rescue operations.
1.5 Thesis Contributions
This thesis, along with work by Hulbert [22] and work by Meehan [13], provide the
documentation for the entire MineSENTRY research project; the three theses contain
all the information required for replicating and improving upon the MineSENTRY
experiments. This thesis focuses on a localization algorithm for the AMR using
ultrasonic range finders, including:
1. The electronics and sensors involved in converting a stock EZ-GO golf cart into
an AMR, particularly covering:
(a) Sensors
(b) Circuit Schematics
(c) Wiring
(d) Robotic-level code
2. The development of a computer simulator to test the localization algorithm on
artificial data
5
|
Colorado School of Mines
|
CHAPTER 2
BACKGROUND
Before diving into vehicle construction and localization algorithm development,
previous research efforts are explored. With a focus on vehicle localization algorithm
development, this chapter highlights some of the research efforts in mining rescue
robotic vehicles, as well as localization algorithms in both indoor and mine environ
ments, using both ultrasonic and scanning laser sensors. Furthermore, this chapter
examines the use of ultrasonic range finders versus scanning sensors in Section 2.3.
Beginning with some challenges of a mining environment for robotic vehicles,
this chapter first describes some of the notable vehicles designed to overcome those
challenges. Many of the vehicles were either tele-operated or are not designed for
emergency situations (e.g. exploratory robotic vehicles), so were not equipped with
the appropriate sensors for this project. Other research efforts fill in the gap by testing
various sensors for emergency rescue operations, specifically comparing scanning laser
sensors with ultrasonic range finders.
The remaining sections cover the previous efforts on localization and mapping
algorithms. First, the algorithms that utilize scanning laser sensors in a mining envi
ronment are explored, followed by the algorithms using the same sensor in an indoor
environment. The last section describes a SLAM algorithm for ultrasonic sensors in
an indoor environment. Each particular sensor and environment combination pro
vides unique challenges, but none satisfy the constraints of this particular project:
ultrasonic sensors in an underground mine environment. Combined, the various re
search efforts have the potential to produce a robotic vehicle platform capable of
traversing a mine and localizing using sensors that are appropriate for a mine rescue
operation.
7
|
Colorado School of Mines
|
2.1 Mining Environment
Mines present a particular challenge for robotic vehicles. Many mines have un
compacted surfaces and large rocks, which may be insurmountable obstacles for small
vehicles. Even for larger vehicles, the uneven terrain causes errors in dead reckoning
sensors, such as wheel odometry; tire slipping and rocks result in an overestima
tion of distance traveled, and tire skidding results in an underestimation of distance
traveled. Many autonomous vehicles incorporate additional sensors such as Global
Positioning Satellite (GPS) receivers or magnetometers to correct the dead reckoning
drift. However, both GPS receivers and magnetometers are useless in mines; iron
deposits common in many mines skew the earth’s magnetic field and satellite signals
cannot penetrate into most mines. Furthermore, water and mud wreak havoc on elec
tronics and sensors, and smoke and toxic gases can be present during mine disasters,
particularly after explosions.
Coal mines present an additional challenge; the dust in the air is flammable, so a
small spark from onboard electronics may trigger an explosion. Particularly important
in coal mines, mining regulations require that all electronics on a vehicle entering a
mine must be either intrinsically safe (defined by the International Electrical Code)
or in an explosion-proof container (as defined by the National Electric Code).
No vehicle is suitable for all mining environments and disasters; size is one of
many dichotomous factors. Large vehicles cannot enter a mine after a roof collapse,
and a small vehicle may have difficulty traversing the rough terrain. Other problems
come into play, such as explosion-proofing, water-proofing, and using sensors robust
to mud, smoke, water, and other conditions common in a mine disaster scenario.
2.2 Mine Rescue Vehicles
Despite the numerous challenges, there have been several attempts at using robots
during mine rescues. For example, the Mine Safety and Health Administration’s
|
Colorado School of Mines
|
(MSHA) ANDROS Wolverine robot (Figure 2.1), nicknamed V2, was deployed dur
ing the Sago Mine rescue operation. M2 is a bomb squad robot customized to be
mine permissible, with explosion-proof motors, navigation and surveillance cameras,
lighting, atmospheric detectors, vision capability, a two-way communication device,
and a manipulator arm [33].
Figure 2.1: Romotec ANDROS Wolverine Robot [33]
At 50 inches tall, and weighing 1200 pounds, V2 can only be deployed via surface
entry. During the Sago mine rescue, the robot traveled approximately 5600 feet before
it ran off the rail, causing the left tread to fall off and a deflated left rear wheel. The
robot failed party because of Human-Robot Interaction (HRI) issues; a single human
operator had difficulty navigating the robot [25].
On the opposite end of the robot size spectrum is the Inuktun Mine Cavern
Crawler (MCC), designed for borehole entry (Figure 2.2). During a mine disaster,
drilling boreholes is common practice in order to determine the air quality and insert
a borehole camera. At approximately eight inches wide, the MCC is a small tethered
robot, capable of fitting into boreholes measuring 8| inches and larger. The MCC
was deployed after a cave-in at the Crandall Canyon Mine, located 120 miles south of
Salt Lake City, Utah. The MCC encountered multiple problems when attempting to
enter the mine; namely, physical barriers blocked the robot, and water, mixed with
9
|
Colorado School of Mines
|
Groundhog navigates by analyzing local three-dimensional scans with regard to
traversable paths, and two-dimensional scan matching keeps the vehicle localized.
Under operator control, Groundhog successfully mapped multiple mines including the
Bruceton Research Mine and the Florence mine, both near Pittsburgh, Pensylvania.
However, during its maiden voyage under autonomous navigation at the Mat hies mine
(also near Pittsuburgh), Groundhog experienced software difficulties. After failing to
establish wireless communication, the wireless link was manually reset and the vehicle
was driven out of the mine under manual control. Though not perfect, Groundhog is
a promising step toward autonomous navigation and Simultaneous Localization and
Mapping (SLAM) in mines.
Despite numerous achievements, none of these robotic vehicles were completely
successful. The world of robotics in mine rescue, particularly autonomous rescue
vehicles, is still new, with plenty of room for improvement. These robotic vehicles
highlight the need for sensors robust to water and other environmental variables,
mobile power and wireless communication development, and improved robot-human
interaction (such as autonomy).
2.3 Sensor Selection
One challenge that was not addressed in the previous section was equipping
the MineSENTRY AMR with sensors appropriate for emergency situations. The
GUARDIANS (Group of Unmanned Assistant Robots Deployed In Aggregative Nav
igation by Scent) project addresses the sensor selection dilemma; they developed a
swarm robotics technique to assist firefighters [31], whom experience similar condi
tions to those present in a mine disaster such as thick smoke, toxic gases, falling
material, and possible explosions. The small swarming robots use sensors robust to
thick smoke to complement the laser sensors; experiments from this project confirm
that laser range finders are prone to failure in the presence of smoke, depending on
11
|
Colorado School of Mines
|
the particulate concentration [30]. Figure 2.4 is one of the experiments.
Distance Measurement
2.5
0.5
0 1 2 3 4 5 6 7
Samples * m‘
Figure 2.4: Laser Range Finder Measurements in Smoke [30]
During this experiment, an object was placed two meters away. Starting at the one
minute mark, smoke was injected into a container surrounding the laser range finder.
During the remaining 28 minutes of the test, the smoke was released through an
opening, increasing the visibility in the test chamber. Even though the target object
remained a constant two meters away, the range readings dropped dramatically after
introducing smoke. Accurate range measurements did not return until the smoke
dissipated, confirming that smoke strongly affects a laser range finder’s ability to
make accurate measurements.
Another study expanded on the GUARDIANS project findings, directly compar
ing a sonar range finder to a laser range finder, as shown in Figure 2.5. Range is
recorded as a function of smoke density, with a target placed two meters away. The
results support findings from the GUARDIANS experiments; the presence of smoke
yields inaccurate range measurements from laser sensors. Although ultrasonic sensors
are less precise, the experiments also confirm that smoke does not significantly affect
ultrasonic range finders.
12
|
Colorado School of Mines
|
S a
1
0
0 10 15 20 25 30 35 40 45 50 55 60
Smoke density (%)
Figure 2.5: Sonar and Laser performance with increasing smoke density [34]
The current industry standard for corridor detection and localization on mining
equipment is a scanning laser sensor. However, as the aforementioned experiments
confirms, thick dust and smoke can block the 800 nm wavelength light used in these
sensors. Since thick dust and smoke are often present in emergency situations, scan
ning laser sensors are not suitable for this project. On the other hand, ultrasonic
sensors use a much longer wavelength than laser sensors (8 x 106 nm compared to 800
nm), and therefore, can still accurately range in thick smoke and dust.
Another characteristic of laser sensors is they have a narrow beam width; a laser
sensor measures the distance to a single point (of negligible area). While a point
measurement is advantageous in many regards, it can cause problems. For example,
a slight vehicle tilt can alter readings from a laser. In contrast, ultrasonic sensors have
an adjustable beam width (by replacing the sensor, from about 15° to 45°) and return
a single measurement from a relatively large target area—the very nature of ultra
sonic sensors provide a low-pass filter, robust to vehicle tilting, but correspondingly,
resolution is reduced.
Ultrasonic ranging sensors have numerous advantages over laser ranging sensors;
however, ultrasonic sensors are not without their disadvantages. One problem with
the sonar sensors is that their accuracy is on the order of a few centimeters (depending
13
|
Colorado School of Mines
|
on the frequency used), which causes problems when trying to build an accurate
environment model. Another problem with the sonar sensors is they have a limited
angle of incidence: angles greater than the critical angle tend to reflect away from
the sensor, yielding unreliable readings in indoor environments. However, in theory,
the mine environment helps alleviate this issue, because the diffuse mine walls scatter
acoustic waves, allowing the sensor to easily detect the wall (see Section 3.2.4).
Having discussed previous research efforts on emergency rescue vehicles and sensor
selection, the remaining sections highlight the localization algorithms that run behind
the scenes, utilizing ultrasonic range finders and scanning laser sensors to control
various vehicle platforms in either an indoor or underground mine environment. The
sensor and environment combinations are delineated into four categories, each with an
associated localization algorithm: scanning lasers in a mining environment, scanning
lasers in an indoor environment, ultrasonic range finders in an indoor environment,
and ultrasonic range finders in an underground mine environment. The former three
scenarios are discussed in the next section and the reminder of this thesis is dedicated
to the fourth.
2.4 Previous Localization and Simultaneous Localization and Mapping
(SLAM) Overview
Localization has two distinct parts: reference guidance and dead reckoning [4].
Reference guidance refers to algorithms, such as a Kalman filter or particle filter
techniques, that use exteroceptive sensor data (e.g. laser or ultrasonic range finders)
to search the surrounding environment for landmarks (unique environment features).
The major drawback of reference guidance involves data association, or linking the
exteroceptive data with the correct landmark. Dead reckoning comes to the rescue by
giving an initial position estimate, eliminating the improbable landmarks. From there,
the exteroceptive algorithms finely tune the robot’s position, eliminating the drift
14
|
Colorado School of Mines
|
from dead reckoning. Thus, reference guidance and dead reckoning algorithms are
complementary. Exteroceptive choice becomes important; reference guidance works
best with precise and accurate data, though even the best sensors are ineffective if
they are unreliable or fail to function correctly in their expected environment.
General SLAM algorithms are prevalent; however, localization algorithms specific
to a mine environment are somewhat less common. We are unaware of any work on
using ultrasonic rangers for robot localization in a mining environment. Some papers
use an artificial infrastructure, such as beacons [14] [29], reflectors (at LKAB’s Kiruna
Mine) [38], or surveying lasers (built into the mine) [35] for localization. However, an
artificial infrastructure does not exist in all mines, and even if one does exist, it may
not be intact after a mine disaster. To resolve these issues, many papers developed
localization algorithms that use natural mine features, such as walls [20], intersections
[1], and geometric beacons [17].
2.5 Map Building and Localization Using Scanning Laser Sensors in a
Mine Environment
When smoke, dust, or other visible particulates are not a concern, many research
efforts use scanning laser sensors, which supply plenty of data for scan matching and
localization algorithms. Scanning laser sensors are very precise, so they can capture
fine details of the surrounding environment; laser precision allows for algorithms such
as the Iterative Closest Point (ICP) matching algorithm. Though these sensors are not
used on the MineSENTRY project, some of the algorithms still apply. The following
research efforts use multiple scanning laser sensors for exteroceptive sensors, either
mounted front-and-back or orthogonally on the robotic vehicle.
A research paper from Makela describes a means of localization in a mine [20] us
ing scanning laser sensors; Load-Haul Dump (LHD) vehicles autonomously navigate
through a mine using the wall’s profile to correct for dead-reckoning drift. The algo
15
|
Colorado School of Mines
|
rithm uses a scanning laser to measure the wall profile and compare that profile to
a pre-made “environment model,” which is built from range data compiled while an
operator manually drives a route. Assuming the wall profile is uniquely identifiable at
every location (a “fingerprint”), wall profile measurements can be used to determine
and correct vehicle location within a reasonable uncertainty using a Kalman-type al
gorithm. Similar LHDs automate Sandvik Tamrock’s test mine in Tampere, Finland
[2].
The Groundhog vehicle (Section 2.2) also uses laser scan matching for localiza
tion. However, Groundhog does not use an a-priori map: instead, it has the additional
difficulty of localizing while simultaneously creating a map of the mine (SLAM al
gorithm) without using odometry—localization is entirely based on scan matching.
To accomplish this task, Thrun et al. used a Bayesian estimation technique, namely
Markov localization and a particle filter [8].
The SLAM algorithm begins with a scan matching algorithm that compares two
adjacent scans and identifies pairs of overlapping points. The algorithm then com
putes the relative displacement and orientation by minimizing the distance between
all pairs of points. This matching algorithm is not perfect, and typically produces
errors in the form of parallel corridors (Figure 2.6).
Direction of Travel
"3 (Conflict)
Direction of Travel
Figure 2.6: Scan Matching, before erroneous path detection [36]
The parallel corridor is typical in a large scale cyclic environment, where a robotic
vehicle is prone to having large uncertainty when re-traversing a path already traveled;
the positional uncertainty causes the robot to make the correct data association with
the previous path, so the robot incorrectly assumes a parallel corridor. To account for
16
|
Colorado School of Mines
|
this issue, the Groundhog vehicle imposes Gaussian constraints on the displacement
and orientation computations of each successive scan, a representation known as
Markov random fields. When a conflict arises, such as an erroneous parallel hallway,
an iterative search questions past decisions in data association in order to increase
the probability of successive scans, resulting in the map in Figure 2.7. Hahnel et ah
have a similar approach to address the data association issue in a large scale cyclic
environment; a probabilistic model of the residual errors from scan matching results
reduces uncertainty [11].
Map After Adjustment
Figure 2.7: Scan Matching, after erroneous path detection [36]
In another research project, Bakambu et ah developed a system that has simi
lar navigation capabilities as the Groundhog [1]. Bakambu:s robotic vehicle has two
modes: surveying and navigation. Surveying mode produces two and three dimen
sional maps of the mine using artificial landmarks for localization and two orthogonal
scanning laser sensors for range measurements.
In navigation mode, a user sets a high level mission, in the form of waypoints on
an a-priori map (produced in surveying mode), and a motion planner reaches the
waypoints by translating the high-level missions into a set of consecutive navigation
actions. For localization landmarks, the robotic vehicle uses abrupt changes in the
local structure, such as corridors, intersections, and bays. To match the sensor read
ings to landmarks, Bakambu uses point-to-line-segment matching (similar to Cox
[4]) instead of point-to-point matching (similar to Madhavan et al. [19]). Bakambu
notes that the matching algorithm degenerates in long drifts without distinguishable
features. Thus, matching results from these portions of the mine are ignored. This
17
|
Colorado School of Mines
|
problem is not a major concern for MineSENTRY purposes; radio attenuation occurs
around corners in a mine, so precise vehicle localization in long drifts (in between
nodes or adits) is not critical. In other words, until reaching an intersection or bend,
the AMR would only need to remain centered in the drift while moving forward or
backward, depending on RSS. Deciding which path to take upon reaching an inter
section involves high-level path planning, beyond the scope of this thesis.
In another project using scanning laser sensors, Madhavan and Durr ant developed
an outdoor map-building and localization algorithm that used a combination of an
iterative closest point (ICP) algorithm and an Extended Kalman Filter (EKE), to
gether coined ICP-EKF [19] [18]. The ICP-EKF fuses Ackermann dynamics with
an ICP matching algorithm to localize an autonomous robotic vehicle relative to a
pre-made poly-line map (a map that consists of a series of connected line segments).
Figure 2.8 outlines the process.
Initialize the state estimate
with nominal values
Obtain laser range & bearing
at the current time instant
Predict the vehicle pose
using dead-reckoning
|Establish correspondence Polyline x
I using ICP . Map y
Compute predicted range using the established
correspondence and thus compute the range innovation!I*
f Does the computed innov. h nmoo I| PPrroocceeeedd ttoo nneexxtt sseett ooff llaasseerr rraanngg1 e &
y fall within the 2cgate? J j bearing values of the ith scan
Compute vehicle state
estimate updates using EKF
— I Proceed to scan # (i+1)
Figure 2.8: Flow Diagram of the ICP-EKF Algorithm [18]
18
|
Colorado School of Mines
|
The ICP-EKF process is broken down into several steps. First dead reckoning
predicts the vehicle pose. Then, the ICP algorithm calculates the correspondence,
or how well the map fits the laser scan measurements. While proceeding forward,
the vehicle updates its position by fusing position estimates from Ackermann vehicle
dynamics (dead reckoning) with position estimates from correspondence data in an
EKF algorithm. The algorithm proceeds to the next scan iteration and the whole
process is repeated. See Madhavan et al. [19] for more details on the step-by-step
process.
Individually, Ackermann Dynamics and the ICP algorithm have shortcomings;
Ackermann Vehicle Dynamics is prone to drifting (like any other dead reckoning
method) and the ICP algorithm fails to form correspondences in areas without dis
tinguishable landmarks (such as a long corridor or circular regions of a mine adit).
Fusing the two methodologies together dramatically reduces localization uncertainty,
allowing the robotic vehicle to accurately place itself on a map.
Localization using scanning laser sensors has been widely studied and is generally
considered a solved problem. Missing from the literature are localization algorithms
for sensors besides scanning lasers, for use in environments where laser-based sensors
fail. One such environment is a smoke-filled mine during an emergency rescue opera
tion, driving the need for localization algorithm development using ultrasonics range
finders. Before addressing the issues that arise when using ultrasonic sensors, we first
expand upon the previous section by examining the use of scanning laser sensors in
an indoor environment.
2.6 Localization and Matching Algorithms in Indoor Environments
Many research efforts focus on localization and SLAM algorithms in an indoor
environment. Like sensor selection, environment plays a major role in localization
algorithm development; an algorithm designed for indoors most likely will not work for
19
|
Colorado School of Mines
|
an underground mining environment. Though no research efforts have both ultrasonic
range finders and a mining environment, the algorithms developed and lessons learned
for indoor environments can partially apply to the MineSENTRY project.
2.6.1 Matching Using a Laser Sensor in an Indoor Environment
Adapted by several research efforts, including Bakambu et al. [1], Ingemar J.
Cox was one of the first to develop a point-to-line-segment matching algorithm [4].
Intended for a cost-effective means of navigation in an office environment, Blanche
is a non-holonomic vehicle that uses a single rotating infrared ranger and odometry
data to form a map of the surrounding environment. The robot gathers 180 range
points (per full revolution) from the ranging sensor and estimates the wall position
relative to a universal or local reference frame. Using an a-priori map, the robot can
then match its sensor readings to the map using an iterative least squares matching
algorithm. Beginning with the range data and environment map in Figure 2.9, Cox
developed a matching algorithm to correct the dead reckoning drift.
i
Figure 2.9: Blanche Iterative Least Squares Matching Algorithm Scenario [4]
Dead Reckoning drift has skewed the robot’s position estimate both in both carte
sian position, (x,y), and orientation, 0; rotating the collected points clockwise and
20
|
Colorado School of Mines
|
moving the points down and left would match the wall. To correct the drift, Cox uses
an iterative matching algorithm, broken down into several discrete steps [4]:
1. For each point in the image, find the line segment in the model that
is nearest to the point. Call this the target.
2. Find the congruence that minimizes the total squared distance be
tween the image points and their target lines.
3. Move the points by the congruence found in Step 2.
4. Repeat step 1-3 until the procedure converges. The composite of all
the step 3 congruences is the desired total congruence.
For every point in an image (collected range data), the algorithm finds the closest
line segment in the model (a-priori map). The algorithm then iteratively rotates and
translates the image until the image and the model line up as well as possible. Cox
concedes that the problem with least squares is its sensitivity to outliers. However,
eliminating outliers makes this algorithm robust, noting that “minimizing the absolute
standard deviation would probably lead to a more inherently robust algorithm [found
in [12]]. However, it is computationally more expensive” [4].
An indoor environment assumption imposes a few restrictions not applicable to an
underground mining environment. However, these restrictions mostly apply to map
building. Matching the estimated map to the a-priori map still applies to a mining
environment, and hence, this algorithm can still prove useful for this project, despite
its indoor (and laser) origins.
Before a matching algorithm can be successful, ultrasonic sensors have numerous
disadvantages to overcome; compared to a scanning laser sensor, an array of ultrasonic
sensors produces only a fraction of the data, at a slower rate, and with less precision.
Despite this fact, range measurements from ultrasonic sensors can still be used to
build maps and localize using a wall profile. However, the map produced and data
association have a larger uncertainty compared to those produced by laser range data.
21
|
Colorado School of Mines
|
The following section addresses the issues that arise when using an ultrasonic range
finder for localization.
2.6.2 SLAM Using Ultrasonic Ranging in an Indoor Environment
In an attempt to compensate for the relatively large uncertainty of ultrasonic
sensors, Crowley describes a means of localizing by forming line segments from three
consecutive ultrasonic readings [6]. Each line segment has an associated certainty,
called the ‘certainty factor;’ the certainty factor is initialized to a value of one and
any subsequent sensor data verifying the line segment adds one to the certainty factor
up to a maximum of five. This method eliminates stray sonar readings, but assumes
flat surfaces typically found in indoor environments. This method could be applied
to a mining environment to discover side passages, but would model the long tunnel
as a very long, flat wall, defeating the purpose of localization.
Crowley’s matching algorithm is primarily based on vehicle and line segment
(formed from ultrasonic sensor readings) uncertainties. Initially, for each measured
line segment, Crowley computes its orientation (#), its perpendicular distance to the
origin (c), and uncertainty values for 6 and c (o# and ac). Before matching the line
segment to the composite model, the vehicle position uncertainties are added to the
line segment uncertainties, given by
+ + (2.1)
^c^new ^c, old T > (2-2)
new = ae, old
+
acn
(2-3)
where a = sin(6), b = —cos(6), % is the variance of the vehicle’s distance to the
0
origin, and is the variance of the vehicles orientation. If the observed segment
22
|
Colorado School of Mines
|
matches (within a computed uncertainty) the orientation, alignment, and overlap of
the composite model, the vehicle position and composite model are both updated.
Due to their wide beam width, ultrasonic sensors have more uncertainty in their
orientation than in their actual range measurement. Thus, Crowley simplifies error
propagation by approximating sensor covariance with one value: crw, the variance in
the orientation of the sensor. Crowley determined variance of the particular sensor
he used by experimental calibration: aw = 0.1 H-d* tan_1(5°), where d is the distance
reported from the sensor. In a similar fashion, the uncertainty in vertical displacement
of a range measurement (due to vehicle tilting) can also be included in aw. In other
words, a wide beam width can be advantageous as ultrasonics sensors are robust to
vehicle tilting.
2.7 Other Related Work
A missing element from the previous sections was an explicit method for calculat
ing the measurement covariance of an empirical model (i.e. the matching algorithm)
in a Kalman filter. This last section highlights a method for estimating a model’s
measured covariance, from a rather unexpected source.
Kolter et al. [15] developed a multi-model Linear Quadratic Regulator (LQR)
controller to autonomously slide a vehicle sideways into a parking space. The LQR
control interpolates between two separate models for vehicle dynamics: the first mod
els the vehicle traversing in a straight line and the second models the vehicle while
sliding. The first model is well-described mathematically. However, due to stochastic
environmental parameters (e.g. friction), the second model is extremely difficult to
predict. Thus, an approximate model estimates the states while the vehicle is sliding;
the second model is a learned behavior from a human operator executing a single
sliding maneuver. Individually, each model fails to place the car accurately in the
stunt maneuver. However, when the models are fused together, the car is placed
23
|
Colorado School of Mines
|
consistently within two feet of the desired trajectory.
Fusing the two models in an LQR controller is analogous to fusing measured and
predicted data in an EKF algorithm. An important aspect of empirical modeling (be
it a car drifting sideways or observing a mine wall) is estimating the model’s true
covariance. In practice, the true covariance is impossible to obtain; Kolter et al. ap
proximate the covariance of the second model by computing the error of the model’s
predictions and averaging over a small time window. This technique effectively low
ers the covariance in neighboring points of the trained data set, making the sliding
vehicle more robust to deviations from the training set. Though the multi-model
LQR controller is not a mine rescue vehicle or a SLAM algorithm, the technique for
approximating the covariance matrix can be applied to the measurement update of
the EKF algorithm, making the matching algorithm more robust to sensor noise.
2.8 Background Summary
Despite the higher uncertainty and lower data rate of ultrasonic sensors compared
to laser sensors, their ability to accurately range in smoke, and therefore emergency
situations, overcomes the disadvantages. Though no research focuses on using ultra
sonic sensors for localization in a mine, research efforts that developed localization
algorithms using scanning laser sensors, as well as research for indoor environments
and other related work, can help produce similar algorithms for ultrasonic range find
ers. Before developing the localization algorithm, a vehicle is required for testing; the
next chapter covers the vehicle development.
24
|
Colorado School of Mines
|
CHAPTER 3'
VEHICLE DESIGN
Using some of the ideas from Sections 2.2 and 2.3 in the previous chapter, this
chapter focuses on the electronics and sensors involved in converting a stock EZ-GO
TXT1 golf cart into the Autonomous Mobile Radio (AMR). The mechanical portion
and communication protocols are only briefly discussed: for further information, refer
to Hulbert [13]. For Information on the wireless tethering, refer to Meehan [22].
The overall purpose of the AMR is to relay a radio signal in an underground mine
during an emergency scenario. Thus, the vehicle must be large enough to carry a
radio and camera, rugged enough to traverse a typical underground mine, and have
the battery capacity to last the duration of the mine rescue operation. Furthermore,
the AMR must be capable of navigation without assistance, which requires numerous
sensors and computer algorithms. Each of these design requirements are addressed,
starting with the initial design decisions and moving on towards vehicle construction.
3.1 Design Alternatives
As with any design project, there are numerous design alternatives. Two of the
most significant design decisions involved robotic vehicle size and exteroceptive sensor
choice. The following briefly describes the alternatives to each of the two major design
choices, along with the reasoning behind the final decisions.
3.1.1 Vehicle Size
Because most design decisions depend on vehicle size, it is the first decision to
make. The vehicle size was broken down into two options: large (capable of trans
porting a human passenger) and small (capable of transporting the necessary elec
tronics only). Smaller vehicles have the advantage of fitting in tight places, mobility,
25
|
Colorado School of Mines
|
and portability. However, rocks, boulders, and even cart tracks become major obsta
cles for these smaller vehicles. Additionally, without a tether, these vehicles have a
relatively short range because their ability to store energy (e.g. a battery) is limited.
Larger vehicles can easily overcome some of the obstacles that are problematic for
smaller vehicles. Even without tethering, larger vehicles have a relatively long range,
with the added ability to haul heavy equipment, injured miners, etc. However, these
vehicles can get stuck in narrow passages and areas with low ceilings.
In an early attempt to determine proper size, a small radio control vehicle was
tested at the Edgar Mine. Measuring 20” x 16” x 10” (Length x Width x Height) and
weighing about 10 pounds, the vehicle managed to traverse over most obstacles, but
struggled when it high-centered on a cart rail; the 3” ground clearance was insufficient
for the 6” rails.
Due to the rigors of a mine environment (Section 2.1), and the ability to transport
persons and equipment, larger vehicles are preferred in the mining industry. Thus
the vehicle of choice is a TXT1 EZ-GO golf cart. This vehicle has sufficient room
for necessary electronics and sensors, the radio and camera, six large 6V lead acid
batteries, and 2 passengers.
3.1.2 Sensor Choice
. Another important decision driving the vehicle design is the sensor choice; based
on Section 2.3 in the previous chapter, the choice of sensors for emergency situations
is somewhat limited to ultrasonic sensors. There are a few interesting alternatives,
including a scent sensor (GUARDIANS project [31]); however, this sensor requires a
‘unique’ scent to follow, which is uncommon in a mine. Other algorithms use a hybrid
approach using both scanning lasers and ultrasonic sensors [34]. Though, if smoke
renders scanning laser sensors unusable, a localization algorithm without scanning
lasers (i.e. ultrasonic sensors alone) would be beneficial. Since laser range finders are
26
|
Colorado School of Mines
|
unreliable in adverse conditions, the vehicle must have the ability to range and localize
using ultrasonic sensors alone. Thus the AMR has an array of 14 ultrasonic sensors
surrounding the vehicle, providing both soft bumper and localization capabilities.
Given the design decision to use ultrasonic sensors, and the decision from the previous
section to use a large vehicle, the remainder of the chapter is dedicated to building
the vehicle platform.
3.2 AMR Construction
The MineSENTRY project developed a vehicle for the conditions at the Edgar
Mine in Idaho Springs, Colorado. Historically, the mine produced metals such as
silver, gold, lead, and copper. Today, the mine serves educational purposes, providing
a touring venue for public and school groups, a laboratory for underground research,
and a training facility for future engineers [28]. The mine primarily consists of dry,
loose rock; thus, for a proof-of-concept vehicle, explosions and water-proofing are not
a concern.
This section includes the documentation of the model numbers, calibration, and
placement of the sensors and electronic components used to make the AMR. The
vehicle platform is an EZ-GO TXT1 golf cart; thus, the terms ‘AMR’ and ‘golf cart’
are used interchangeably. The order of this section roughly corresponds to the order
in which the AMR was built. Beginning with the stock golf cart, the first step was to
add hardware, such as actuation and sensors, including sensor characterization. When
all of the hardware was completed, the golf cart was ready for a computer. The next
step was developing the software to run the hardware, which includes the computer
architecture, the microprocessor platform, and the microprocessor code. The final
step was to wire the computer to the hardware components, with careful regard to
isolation, voltage regulation, power distribution, and noise. The engineering safety
controls is the last subsection covered, as numerous safety concerns were discovered
27
|
Colorado School of Mines
|
and corrected as the project progressed.
3.2.1 Actuation
The golf cart required three actuators to become autonomous; the steering, brake,
and throttle actuation were designed so that the AMR could be driven manually at
any point during operation. The steering actuator is a servo motor with a quadrature
encoder linked to the steering wheel via a drive belt. Simply sliding the splined
assembly plate axially engages and disengages the steering servo for autonomous or
manual operation, as shown in Figure 3.1.
V, Servo Motor Engaged ;
(Remote or Autonomous Ntode)
Figure 3.1: Steering Servo Motor Engagement
A linear actuator, located underneath the vehicle, depresses the brake pedal using
analog feedback. The assembly attaching the linear actuator to the brake pedal is
designed such that a human operator can override the brake actuation for emergency
braking. The robotic controller detects the brake override and cuts the throttle. Refer
to Hulbert [13] for more details on the mechanical assemblies.
Throttle actuation is electronic; the on-board golf cart computer reads an analog
signal from an inductive throttle position sensor and adjusts the power accordingly.
28
|
Colorado School of Mines
|
Measuring the voltage across the throttle position sensor terminals yields 13.8 Vdc
at no power and 11.2 Vdc at full power, so interfacing with the on-board computer
is a matter of designing a circuit to match these voltage readings. However, from
the factory, both of the throttle position sensor terminals are ‘floating’ relative to
the positive and negative battery terminals, requiring a more-complicated isolated
control circuitry.
Figure 3.2 shows the throttle position sensor and the microswitch. The throttle
position sensor is built-in to the golf cart and outputs an analog voltage propor
tional to the accelerator pedal position. Adjacent to the throttle position sensor, the
microswitch prevents drive motor power while the accelerator pedal is disengaged.
Throttle q
"XMiproswllch
Inductive Throttle
. Position Sensor
Figure 3.2: Golf Cart Throttle Control Box
A throttle control circuit was designed to meet the voltage and electrical isolation
criteria, while avoiding interference with manual golf cart operation; the circuit design
is centered around a modified buck converter with an isolating transformer. The
circuit schematic can be found in Figure 3.3. This circuit converts a Pulse-Width
29
|
Colorado School of Mines
|
Modulation (PWM) signal from the robotic controller into an analog voltage; a larger
PWM signal produces a smaller analog voltage signal, resulting in more power at the
wheels. This circuit is wired in parallel with the stock golf cart electronics (‘Throttle
Box’), so an operator can command more throttle at any time. If an emergency stop
is required, an operator can use the brake to slow the vehicle; the external brake
depression is detected and the throttle is turned, off.
N,:N2
1 1:1
Throttle
Vinj 11.2V-13.8V
5V Analog Out _ Box
Golf Cart Ground
(Negative Control
Battery Terminal)
Figure 3.3: Throttle Actuation Circuitry
Furthermore, the computer detects throttle-position feedback from a built-in mi
cros witch (see Figure 3.2); power is not activated unless the pedal is depressed, re
quiring bypass circuitry, or a relay. Lastly, the direction (forward and reverse) is
activated by a switch on the golf cart’s front panel, requiring an additional set of
relays. A custom-made circuit board contains the throttle control circuitry, along
with the three optically isolated relays. See Section A.l in the appendix for the over
all circuit schematic and the color code for wiring into the golf cart computer. The
throttle microswitch relay is a single Panasonic Electric Works AQV212 IC (labeled
’AQV212’ on the schematic), and the FWD/REV switch relays are contained in the
single Panasonic Electric Works AQW212 IC (labeled ’AQW212’ on the schematic).
Similar to to the throttle control circuit, all relays are wired in parallel to the existing
switches on the golf cart, allowing an operator to override an open relay at any time.
30
|
Colorado School of Mines
|
The golf-cart computer defaults to reverse if both the forward and the reverse relays
are on simultaneously.
3.2.2 Propreoceptive Sensors
There are a total of six propreoceptive sensors on the AMR: a Hall Effect sensor on
each wheel to measure odometry and calculate vehicle speed, a string potentiometer
on the steering arm to measure the Ackermann steering angle, and a five degree of
freedom Inertial Measuring Unit (IMU) to measure vehicle orientation. Combined,
these sensors provide steering angle, velocity, and yaw, all necessary to compute
Ackermann vehicle dynamics.
The Hall effect sensor on each wheel reads from 12 magnets, making the odometry
resolution 118 mm (tire circumference divided by the number of magnets). The sensor
sends a 3.3V pulse to the microcontroller when a magnet passes, and in turn, the
microprocessor increments the odometer count by 118 mm if the forward switch on
the golf cart is engaged, or decrements the odometer if the reverse switch is engaged.
The speed is then updated with a simple difference equation and a lowpass averaging
filter:
(3.1)
average, n
where n represents a fixed time interval, dn is the current odometry reading in {m}
(the odometry reading is updated after each magnet pass), t is the elapsed time in {s},
S is the vehicle’s speed in {m/s}, and Saverage is a twelfth-order moving average filter
of S. If no magnet pass is detected during the fixed time interval (i.e. the vehicle is
moving slowly or stopped altogether), the numerator of Equation (3.1) becomes zero.
After 12 subsequent time intervals without a magnet detection, Saverage,n becomes
31
|
Colorado School of Mines
|
zero.
For steering angle measurements, a Celesco SP1 string potentiometer is used.
This sensor is fixed to the steering rack and outputs a 0-3.3V proportional to its 0-6”
range, so the steering angle is approximately proportional to the string potentiome
ter’s output voltage; Figure 3.4 is the voltage to steering angle calibration curve. The
conversion from voltage to steering angle is 6 = —31.1 • V + 47. This calibration
was performed using two long straight edges: one aligned with the rear tire and one
aligned with the front tire. The angle between these two straight edges is the steer
ing angle, and observing the voltage for every 5-10 degrees of steering throw gives
the data for a linear regression. Note that a positive steering angle is one that in
duces a counter-clockwise rotation of the vehicle when viewed from above (i.e. left).
Calibration for all analog sensors can be found in Section A.2
String Pot Calibration
50 i
40
30
» 20 —
TJ
a
1 0 -
I
o -
0.50 1.00 2.00 2.50 3.00 3.50
! - i f -
-20
£ —
-31. Ix + 46.958
-30
0.9914
-40
-50 -*•
Voltage Out (V)
Figure 3.4: String Potentiometer Voltage to Steering Angle Calibration
An Inertial Measurement Unit (IMU) detects the vehicle’s acceleration and ro
tation. The IMU is a board from Sparkfun® Electronics. This board incorporates
32
|
Colorado School of Mines
|
units, which have a wider beam width for better obstacle detection, making the EZ1
model best suited as soft bumpers. The sensors oriented towards the wall are LV-
EZ3 units, which have a narrower beam width, and therefore, more accurate ranging
measurements compared to the EZ1; thus, the EZ3 model is well suited for building a
map of the environment. Refer to Table Table 3.1 for a comparison of EZ1 and EZ3
sensors. Furthermore, Section 3.2.4 elaborates on the beam width characteristics.
Table 3.1: List of Sensor Model Numbers
Sonar Sensor Model Beam Width Use
1, 3, 7, 9 Maxbotix LV-EZO Wide Soft Bumpers
All Others Maxbotix LV-EZ3 Narrow Wall Mapping
Each sensor has a 50 ms update rate, so triggering one sensor at a time would
take 0.7 seconds, corresponding to a 1.4 Hz update rate. When controlling a vehicle,
a quicker update rate is more desirable; to speed up the update rate, the 14 sensors
are broken up into four triggering groups. Each triggering group has between two
and four sonars; sensors within each group are triggered simultaneously. Refer to
Table 3.2 for the list of sonars in each triggering group.
Table 3.2: Sonar Triggering Groups
Trigger Group Sonars in Group
1 1, 2, 7, 8
2 3, 4, 9, 10
3 5, 6, 11, 12
4 13 and 14
These triggering groups are daisy chained together. A pulse signal from the robotic
controller sets off the chain. When each triggering group finishes collecting range
measurements, a pulse signal is sent to the next triggering group, until the final pulse
from the fourth triggering group signals the robotic controller to start analog to digital
conversion. Each triggering group takes approximately 50 ms to collect data, making
34
|
Colorado School of Mines
|
the total update rate 5 Hz. To avoid interference, each sensor is oriented 90° apart
from other sensors in the group, as illustrated in Figure 3.5. Measurements from
these sensors are used to build an environment model of mine walls.
In addition to the sonar sensors, four Sharp Infrared long-range sensors are mount
ed at all four corners of the vehicle at 45° angles. These infrared sensors were originally
intended to serve as backups to the diagonal sonar sensors (sensor numbers 5, 6, 11
and 12), and to provide additional range measurements in the event that the ultrasonic
sensors return inconsistent data during testing. However, these sensor were damaged
upon transportation to the mine, so they were not used.
3.2.4 Sonar Characterization
A single wall-profiling EZ3 sensor was tested for its mean and standard deviation
in both indoor and underground mine environments. During each test, the sensor was
■first oriented directly at the wall (0°), at one meter away (measured with measuring
tape). The sonar’s voltage measurements were recorded at 200 Hz for 10 seconds.
The manufacturer’s voltage conversion equation was used to report the range mea
surements: d = Vmeasured ' [512/I/CC] • 0.0254, where Vcc is the power supply voltage
(3.31/), Vmeasured is the sensor’s voltage output, and d is the range measurement in
meters. For each environment, the data was recorded again at two meters away from
the wall, making a total of four experiments: indoors at one meter and two meters,
and in the mine at one meter and two meters.
The mean (/r0) and the standard deviation (ctq) of the range measurements for
the four experiments are reported in Table 3.3. With a mean near the measured
distance from the wall and a standard deviation under 3 cm, the sensors are fairly
accurate and precise. Because of the smoother walls, the standard deviation indoors
was slightly better compared to the mine.
35
|
Colorado School of Mines
|
Table 3.3: Sonar Characterization, # = 0°
Location Distance from Wall {m} /rn {m} cr0 {m}
Indoors 1.0 0.96 0.014
Indoors 2.0 1.9 0.010
Edgar Mine 1.0 1.0 0.026
Edgar Mine 2.0 2.0 0.017
The sonar sensor was also characterized for its critical angle (6critical)- Starting at
0° (perpendicular to the wall), the sensor was rotated 10° at a time until it was no
longer able to detect the. wall, at which point the sensor angle was reduced by 5° and
re-tested. The maximum angle the sensor can be rotated without losing the signal
from the wall is the critical angle {0criticai) . The mean (nc) and standard deviation
(crc) of the range measurements for each scenario at the critical angle are given in
Table 3.4.
Table 3.4: Sonar Characterization, 0 = 6critiCai
Location Distance from Wall {m} ^critical { } Me {m} crc {m}
Indoors 1.0 25 1.0 0.66
Indoors 2.0 25 2.2 0.98
Edgar Mine 1.0 45 1.2 0.47
Edgar Mine 2.0 25 1.9 0.18
The irregularities of the mine wall are nearly the same size as the sonar sensor’s
emitted wave packet, leading to a scattering effect of the acoustic wave. Theoretically,
the scattering effect would allow the sensor to detect the mine wall more easily,
meaning a higher critical angle and a lower standard deviation of the range data
the mine. In practice, the critical angle increased only when near the mine wall
(<lm), and ac was marginally lower in the mine compared to indoors, though ac
was higher across the board compared to ctq (Table 3.3). With the lower <tc and
the potentially higher critical angle, the mine offered a slight advantage in sensor
performance compared to indoors. However, regardless of the sensor environment,
36
|
Colorado School of Mines
|
the ultrasonic sensors are unreliable at angles greater than 25°.
A graphical representation of one sonar characterization experiment is illustrated
in Figure 3.6. This experiment took place at the Edgar Mine with the sonar range
finder rotated through various angles, two meters away from the mine wall. The
dotted lines show 10° increments, the solid black line is a rough approximation of the
mine wall, and the blue dots represent the sonar range data. The sonar range data is
fairly compact when pointed directly at the wall {6 = 0°), but the range measurements
become more widely dispersed as the angle increases. Figure 3.6 clearly delineates the
limit for the sensor angle; beyond 9criticai — 25°, the the sensor has difficulty detecting
the wall, and the range readings vary widely from about two meters to six meters.
These finding fully support the data in Table 3.3 and Table 3.4.
Sonar Characterization
Edgar Mine Army Tunnel
" Mine Wall
" Sonar Data
03
03
E
>
X (meters)
Figure 3.6: Sonar Characterization in the Edgar Mine (Overhead View)
The sensor orientations of Sonars 5, 6, 11, and 12 in Figure 3.5 are 45°. This is the
minimum angle to avoid interference among sensors. Admittedly, the original sonar
characterization took place in a laboratory with numerous acoustically reflectively
surfaces, so the sonars were originally thought to have over a 45° critical angle. The
near-wall critical angle supported this claim, but, as indicated in Figure 3.6, the far
37
|
Colorado School of Mines
|
wall critical angle was found to be far lower (25°). Discovered long after the data was
collected, and requiring a significant hardware change, the rest of this thesis assumes
a 45° critical angle. This issue is addressed in Section 5.4.
The data in the mine was taken across a relatively straight portion of the mine
adit, representing the worst-case scenario; typically, the mine has irregularities that
give the ultrasonic sensors a stronger signal. In fact, the sensor directed toward
one such irregularity at a 40° angle yielded a 0.076m standard deviation from 2.8m
away. However, the sensor lost the signal altogether when rotated by 5° in either
direction. In certain situations, the mine environment can improve the critical angle
and the confidence (standard deviation) of the range measurements, but in general,
the ultrasonics are no more reliable in a mine than they are indoors.
Another characteristic of the ultrasonic sensors is evident in the mean; for certain
scenarios, fic is the same as //0, despite being rotated by 25° or 45°. Due to the
wide beam width, the location of the object (point on the wall) reflecting an echo
is uncertain, causing the range measurement to repeat through sensor rotation. For
example, an ultrasonic sensor with a 10° beam width may detect the same object (and
therefore report the same distance) when rotated from 0° all the way to 20°, creating
a detection pattern in the shape of an arc. In other words, neither the distance nor
the angle are necessarily accurate.
The LV-MaxSonar®-EZ3™datasheet [21] describes the beam width—or rather,
the detection pattern—of the sonar sensors in more detail. Figure 3.7 is the detection
pattern of an EZ3 sensor for a 3.5" diameter rod for 5V (black line) and 3.31^ (red
dots) on a 12" grid. The detection pattern is highly dependent on the reflecting
object, but a 3.5” diameter rod is a fair approximation for a mine wall; the round
surface disperses acoustic waves in several directions and the rod is large enough to
be detectable at distances expected in an underground mine.
For the 3.5" rod, the detection pattern is about 2ft wide, or 4=0.3m from the
38
|
Colorado School of Mines
|
center. Given that the shape and reflectivity of a detected object are typically un
known, the standard deviation of the EZ3 sonar sensors can be over-estimated as
&ezz — 0.3m. Similar to Crowley’s [6] method for approximating the sonar error,
(Jezz is extended in both the axial and perpendicular directions for simplification;
the maximum axial standard deviation (Edgar Mine Location at 25° in Table 3.4)
and the perpendicular standard deviation (sensor beam width, Figure 3.7) are both
near 0.3 m. Thus, the wall-profiling sensors have a 68% confidence interval (±1(Jezz)
that the actual object is within 4=0.3m of the detected position. At 3m away, this
corresponds to about ±5°.
Figure 3.7: Sensor Detection Pattern [21]
3.2.5 Mobile Radio Components
Because the radio is a standalone system, requiring only a power connection to
charge the built-in battery, it was the last hardware component added to the AMR,
so it is briefly described here. Without the mobile radio, the AMR would simply be
’Autonomous.’ The radio workhorse is a Raj ant radio, mounted on top of the AMR’s
electronics housing, shown in Figure 3.8. Each Raj ant radio automatically connects
to nearby Raj ant radios to form a wireless mesh network, robust to the dynamic
nature of the caravan scenario depicted in the introduction to this paper. The nature
of these radios is beyond the scope of this paper; for more information, see the thesis
by Meehan [22].
39
|
Colorado School of Mines
|
Figure 3.8: Raj ant Radio and Mounting Location on the AMR.
This concludes the vehicle platform hardware section, which covered actuation,
sensors, and sensor characterization. At this point, the golf cart can only be controlled
manually. The remaining sections cover switching between manual, remote, and
computer-controlled modes; adding the computer and software, which includes the
computer architecture, the microprocessor platform, and the microprocessor code;
power distribution; and engineering safety controls.
3.2.6 AMR Operation Modes
Transitioning from manual control to computer control requires a switch—in this
case, a software-monitored mechanical mode switch. The AMR has three user-
switchable modes of operation: Manual, Remote, and Autonomous. Manual mode,
as the name implies, is intended for manual, human operation. An operator ensures
that the steering motor is disconnected, and the AMR is ready for manual ’golf cart’
driving mode. To return to the other two modes, a human operator must rotate the
mode switch and re-connect the servo motor to the steering column.
The remaining two modes utilize the on-board computers, which are described in
the following section. Remote mode is intended for driving the AMR using a remote
40
|
Colorado School of Mines
|
transmitter with the controls illustrated in Figure 3.9. The transmitter and receiver
pair is a Spektrum® DX7 7-channel transmitter bound to a Spektrum® AR6000
receiver. The transmitter sends throttle and steering commands to the receiver at
2.4 GHz, and the robotic controller reads the standard servo output from the re
ceiver (see Section A.3 in the Appendix) and adjusts the throttle, brake, and steering
proportionally to the servo pulse width.
Figure 3.9: Radio Transmitter Driving Controls
Entering remote mode, the robotic controller has a specific startup sequence. First,
the robotic controller depresses the brake and locks up until it detects the specific
startup sequence illustrated in Section A.8.2. Then, the robotic controller reads the
steering angle from the string potentiometer and centers the steering. Finally, the
robotic controller returns to its normal operation.
Autonomous mode behaves similarly to remote mode, except the autonomous con
troller sets the steering and brake commands automatically based on sensor readings
and navigation routines. For more information on the autonomous controller, its
navigation routines, and its control algorithms, see Hulbert [13]. Knowing the overall
41
|
Colorado School of Mines
|
function of the autonomous and the robotic controllers, the following section goes
into more detail of each computer and its role in the overall system.
3.2.7 Computer Architecture
The computer architecture consists of three hierarchical layers: a servo controller,
a robotic controller, and an autonomous controller (see Figure 3.10). At the low
est level (far right), the RoboteQ AX3500 servo controller closes the loop around
commanded steering and brake actuator values via PID control. The middle level
consists of a Microchip® dsPIC processor in charge of robotic functions, such as
analog-to-digital conversion, servo set-point calculations, and speed control. Lastly
the autonomous controller (far left) handles upper-level navigation routines [24].
MineSENTRY AMR Vehicle Level Hardware
Digital Input
Estop
SM wo itd ce
h
Digit (a xl 2 )Input, Analog S Ptr oin tg
Dig ta! Input
FWD/
REV
Analog (xS)
5DOF
IMU •^Steering Actuator
A Cut oo nn tro om llo erus RS-232 CR ono tb ro ot li lc
er
RS-232 CoS ne tr rv oo
ller
Digital Encoder
Analog (x4) PWM
Infrared Brake Actuator
Sensors
Analog
Analog (xt
Sonar
Sensors
Battery Analog PWM pSwitch PWM Throttle Actuator
Monitor
Digital Input
Capture (x4i
Hall-
Effect
* Digital Input
Capture
Speed
Figure 3.10: Computer Architecture
The protocol for the communication among the three hierarchical layers is RS232.
The communication between the robotic controller and the servo controller is defined
42
|
Colorado School of Mines
|
by the RoboteQ AX3500 control board command set and configuration, summarized
in Section A.4.
The communication packet structure between the robotic controller and the au
tonomous controller is also in the Appendix (Section A.5 on Page 97 and Section A.6
on Page 99). These packet structures standardize the communication protocol for
error handling, sensor data, and vehicle status.
3.2.8 Microprocessor Platform
Originally, the microprocessor was purchased based on slightly different interface
requirements than depicted above: one RS232 port, one RS485 port, 15 analog to dig
ital converter channels, four digital inputs, one digital output, one input capture pin,
one PWM channel, and one Universal Serial Bus (USB) port. The FLEX platform—
the product of two Italian companies, Evidence Sri and Embedded Solutions Sri—was
one of the only platforms to meet the criteria, particularly because the RS485 and
USB combination is rare among microprocessor development platforms. The FLEX
full board (Figure 3.11), when combined with the multi-bus daughterboard, is fully
customizable with various Controller Area Network (CAN), Universal Asynchronous
Receiver/Transmitter (UART—RS232 and RS485), Serial Peripheral Interface (SPI),
and et her net communication modules.
power supply
9 -36 V
■g - c
CON5 for piggybacking
COM4 COM2
socket for 100 pin
tmmmtm
PIC18
immmmi
6 LEDs
Figure 3.11: FLEX Platform Base Board [7]
43
|
Colorado School of Mines
|
Interrupts are used for important time-dependant events, such as data collection,
RS232 communication, and the emergency stopping procedure. All other tasks, such
as the watchdog timer, command processing (servo set point calculations), and data
processing (conversion to engineering units), are handled using polling.
The code, in its entirety, is in Section A.7 of the Appendix, along with companion
flow charts in Section A.8 to help decipher the important parts of the code. For
information on the code in the autonomous controller, see Hulbert [13].
This concludes the software section, which covered the computer architecture, the
microprocessor platform, and the microprocessor code. The following sections discuss
how the computer and the remaining hardware are tied together.
3.2.10 Power Distribution and Wiring
With the sensors, actuators, drive motor, computers and mobile radio equipment,
power distribution becomes an important issue. To power the wheels and the on-board
computer, the golf cart relies on six 6V lead acid batteries, wired in series, for a total
of 36V. Due to their high capacity, these main batteries also feed the brake actuator
and the steering servo (after passing through an isolated 24V industrial regulator).
Because the main battery bank tends to be electrically noisy while the golf cart motor
is turning, all other components added to the stock golf cart configuration receive
power from an isolated, standalone 12V lead acid battery with voltage regulation for
a 5V and a 3.3V tap. Table 3.5 lists the electronics on the AMR, along with their
power requirements.
Power distribution to all 14 ultrasonic sensors, four IR sensors, four Hall Effect
sensors, and one string potentiometer requires copious amounts of cabling. Thus, two
power distribution boards were designed to reduce the number of cables and clutter
in the main electronics housing. The sensor connector locations and the pinouts for
both boards are located in Section A.9.
45
|
Colorado School of Mines
|
Table 3.5: AMR On Board Electronics Power Requirements
Component Input Voltage Power Source
Golf Cart Computer 36V 36V Main Battery Bank
Golf Cart Drive Motor 36V 36V Main Battery Bank
Brake Actuator 24V 36V Main Battery Bank/ 24V
Regulator
Steering Servo Motor 24V 36V Main Battery Bank/ 24V
Regulator
AX3500 Servo Controller 24V 36V Main Battery Bank/ 24V
Regulator
FLEX Microprocessor Board 12 V 12V Auxiliary Battery
(robotic controller)
Raj ant Radio 12 V 12V Auxiliary Battery
Ethernet Router 12V 12V Auxiliary Battery
Ultrasonic Sensors 3.3V 12V Aux. Battery/ 3.3V Reg
ulator
Infrared Sensors 3.3 V Aux. Battery/ 3.3V Regulator
String Potentiometer 3.3V Aux. Battery/ 3.3V Regulator
Hall Effect Sensors 5V Aux. Battery/ 5V Regulator
Throttle Control Circuitry 5V Aux. Battery/ 5V Regulator
FWD/REV/Thottle Relays 5V Aux. Battery/ 5V Regulator
With up to 15 feet of cable routing analog signals from the sensor to the robotic
control, electromagnetic noise from the golf cart motor became an issue. In an at
tempt to attenuate the noise, the control battery, all power supplies, voltage regu
lating circuits, sensors, and cable shielding were grounded to a single point in the
electronics housing, as shown in Figure 3.13. Despite the star grounding, some noise
is still present, both in the signal and the power supply of the sensors. At the recom
mendation of the manufacturer, a low-pass filter was implemented on each ultrasonic
sensor’s voltage input. The low-pass filter is a first-order Resistor-Capacitor (RC)
filter with R = 100f2 and C = 100/rF, giving a cutoff frequency of 16 Hz.
Still having problems with noisy sensor readings, additional RC filters were in
stalled on all affected sensor signals, including all analog sensor lines and the 4 Hall
Effect sensor lines; with R — 100f2 and C = 1/zF, each line has a filter with a cut
off frequency of 1.6 kHz. This cutoff is sufficiently high to leave the normal sensor
46
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.