Sunday, 13 March 2011 16:36

Detection of Gases

Rate this item
(0 votes)

All who work in underground mines should have a sound knowledge of mine gases and be aware of the dangers they may present. A general knowledge of gas detection instruments and systems is also necessary. For those assigned to use these instruments, detailed knowledge of their limitations and the gases they measure is essential.

Even without instruments, the human senses may be able to detect the progressive appearance of the chemical and physical phenomena associated with spontaneous combustion. The heating warms the ventilating air and saturates it with both surface and integral moisture driven off by the heating. When this air meets colder air at the ventilation split, condensation occurs resulting in a haze and the appearance of sweating on surfaces in the returns. A characteristic oily or petrol smell is the next indication, followed eventually by smoke and, finally, visible flames.

Carbon monoxide (CO), which is odourless, appears in measurable concentrations some 50 to 60 °C before the characteristic smell of a spontaneous combustion appears. Consequently, most fire detection systems rely on the detection of a rise in carbon monoxide concentration above the normal background for the particular part of the mine.

Sometimes, a heating is first detected by an individual who notices a faint smell for a fleeting instant. Thorough examination of the area may have to be repeated a number of times before a measurable sustained increase in the concentration of carbon monoxide can be detected. Accordingly, vigilance by all those in the mine should never be relaxed and a prearranged intervention process should be implemented as soon as the presence of an indicator has been suspected or detected and reported. Fortunately, thanks to considerable progress in the technology of fire detection and monitoring made since the 1970s (e.g., detector tubes, pocket-sized electronic detectors, and computerized fixed systems), it is no longer necessary to rely on the human senses alone.

Portable Instruments for Gas Detection

The gas detection instrument is designed to detect and monitor the presence of a wide range of gas types and concentrations that could result in a fire, an explosion and a toxic or oxygen-deficient atmosphere as well as to provide early warning of an outbreak of spontaneous combustion. Gases for which they are used include CO, carbon dioxide (CO2), nitrogen dioxide (NO2), hydrogen sulphide (H2S) and sulphur dioxide (SO2). Different types of instrument are available, but before deciding which to use in a particular situation, the following questions must be answered:


  • Why is the detection of a particular gas or gases required?
  • What are the properties of these gases?
  • Where and in what circumstances do they occur?
  • Which gas detecting instrument or device is most suitable for those circumstances?
  • How does this instrument work?
  • What are its limitations?
  • How should the results it provides be interpreted?


Workers must be trained in the correct use of portable gas detectors. Instruments must be maintained according to the manufacturer’s specifications.

Universal detector kits

A dectector kit consists of a spring-loaded piston- or bellows-type of pump and a range of replaceable glass indicating tubes that contain chemicals specific for a particular gas. The pump has a capacity of 100 cc and can be operated with one hand. This allows a sample of that size to be drawn through the indicator tube before passing to the bellows. The warning indicator on the graduated scale corresponds to the lowest level of general discolouration, not the deepest point of colour penetration.

The device is easy to use and does not require calibration. However, certain precautions are applicable:

  • Indicator tubes (which should be dated) generally have a shelf-life of two years.
  • An indicator tube may be re-used ten times provided there has been no discolouration.
  • The general accuracy of each determination is usually within ± 20%.
  • Hydrogen tubes are not approved for use underground because of the intense heat developed.
  • A “pre-tube” filled with activated charcoal is required when estimating low levels of carbon monoxide in the presence of diesel exhausts or the higher hydrocarbons that may be present in afterdamp.
  • Exhaust gas should be passed through a cooling device to make sure the temperature is below 40 °C before passing though the indicator tube.
  • Oxygen and methane tubes are not approved for use underground because of their inaccuracy.


Catalytic-type methanometers

The catalytic-type methanometer is used in underground mines to measure the concentration of methane in the air. It has a sensor based on the principle of a network of four resistance-matched spiral wires, usually catalytic filaments, arranged in a symmetrical form known as a Wheatstone-bridge. Normally, two filaments are active and the other two are passive. The active filaments or beads are usually coated with a palladium oxide catalyst to cause oxidation of the flammable gas at a lower temperature.

Methane in the atmosphere reaches the sample chamber either by diffusion through a sintered disc or by being drawn in by an aspirator or internal pump. Pressing the operating button of the methanometer closes the circuit and the current flowing through the Wheatstone-bridge oxidizes the methane on the catalytic (active) filaments in the sample chamber. The heat of this reaction raises the temperature of the catalytic filaments, increasing their electrical resistance and electrically unbalancing the bridge. The electric current that flows is proportional to the resistance of the element and, hence, the amount of methane present. This is shown on an output indicator graduated in percentages of methane. The reference elements in the Wheatstone-bridge circuit serve to compensate for variations in environmental conditions such as ambient temperature and barometric pressure.

This instrument has a number of significant limitations:

  • Both methane and oxygen must be present to get a response. If the oxygen level in the sample chamber is below 10%, not all the methane reaching the detector will be oxidized and a false low reading will be obtained. For this reason, this instrument should not be used to measure methane levels in afterdamp or in sealed off areas where the oxygen concentration is low. If the chamber contains pure methane, there will be no reading at all. Accordingly, the operating button must be depressed before moving the instrument into a suspected methane layer in order to draw some oxygen-containing air into the chamber. The presence of a layer will be confirmed by a greater than full scale reading followed by a return to scale when the oxygen in consumed.
  • The catalytic type of methanometer will respond to flammable gases other than methane, for example, hydrogen and carbon monoxide. Ambiguous reading, therefore, may be obtained in post-fire or explosion gases (afterdamp).
  • Instruments with diffusion heads should be sheltered from high air velocities to avoid false readings. This may be accomplished by shielding it with a hand or some other object.
  • Instruments with catalytic filaments may fail to respond to methane if the filament comes in contact with the vapours of known poisons when being calibrated or used (e.g., silicones in furniture polish, floor polish and paints, phosphate esters present in hydraulic fluids, and fluorocarbons used as the propellant in aerosol sprays).
  • Methanometers based on the Wheatstone-bridge principle may give erroneous readings at variable angles of inclination. Such inaccuracies will be minimized if the instrument is held at an angle of 45° when it is calibrated or used.
  • Methanometers may give inaccurate readings at variable ambient temperatures. These inaccuracies will be minimized by calibrating the instrument under temperature conditions similar to those found underground.


Electrochemical cells

Instruments using electrochemical cells are used in underground mines to measure oxygen and carbon monoxide concentrations. Two types are available: the composition cell, which responds only to changes in oxygen concentration, and the partial pressure cell, which responds to changes in the partial pressure of oxygen in the atmosphere and, hence, the number of oxygen molecules per unit of volume.

The composition cell employs a capillary diffusion barrier which slows the diffusion of oxygen through the fuel cell so that the speed at which the oxygen can reach the electrode depends solely on the oxygen content of the sample. This cell is unaffected by variations in altitude (i.e., barometric pressure), temperature and relative humidity. The presence of CO2 in the mixture, however, upsets the rate of oxygen diffusion and leads to false high readings. For example, the presence of 1% of CO2 increases the oxygen reading by as much as 0.1%. Although small, this increase may still be significant and not fail-safe. It is particularly important to be aware of this limitation if this instrument is to be used in afterdamp or other atmospheres known to contain CO2.

The partial pressure cell is based on the same electrochemical principle as the concentration cell but lacks the diffusion barrier. It responds only to the number of oxygen molecules per unit volume, making it pressure dependent. CO2 in concentrations below 10% have no short-term effect on the reading, but over the long term, the carbon dioxide will destroy the electrolyte and shorten the life of the cell.

The following conditions affect the reliability of oxygen readings produced by partial pressure cells:

  • Altitude and barometric pressure: The trip from the surface to the bottom of the shaft would increase the oxygen reading by 0.1% for every 40 m travelled. This would also apply to dips, encountered in the underground workings. In addition, the 5 millibar normal daily variations in barometric pressure could alter the oxygen reading by as much as 0.1%. Thunderstorm activity could be accompanied by a 30 millibar drop in pressure that would cause a 0.4% drop in the oxygen reading.
  • Ventilation: The maximum ventilation change at the fan would be 6-8 inches water gauge or 10 millibar. This would cause a drop of 0.4% in the oxygen reading going from the intake to the return at the fan and a drop of 0.2% in travelling from the furthest face from the pit bottom.
  • Temperature: Most detectors have an electronic circuit that senses cell temperature and corrects for the temperature effect on the sensor output.
  • Relative humidity: An increase in relative humidity from dry to saturated at 20 °C would cause approximately a 0.3% decrease in the oxygen reading.


Other electrochemical cells

Electrochemical cells have been developed which are capable of measuring concentrations of CO from 1 ppm to an upper limit of 4,000 ppm. They operate by measuring the electric current between electrodes immersed in an acidic electrolyte. CO is oxidized on the anode to form CO2 and the reaction releases electrons in direct proportion to the CO concentration.

Electrochemical cells for hydrogen, hydrogen sulphide, nitric oxide, nitrogen dioxide and sulphur dioxide are also available but suffer from cross-sensitivity.

There are no commercially available electrochemical cells for CO2. The deficiency has been overcome with the development of a portable instrument containing a miniaturized infrared cell that is sensitive to carbon dioxide in concentrations up to 5%.


Non-dispersive infrared detectors

Non-dispersive infrared detectors (NDIRs) can measure all gases that contain such chemical groups as -CO, -CO2 and -CH3, which absorb infrared frequencies that are specific to their molecular configuration. These sensors are expensive but they can provide accurate readings for gases such as CO, CO2 and methane in a changing background of other gases and low oxygen levels and are therefore ideal for monitoring gases behind seals. O2, N2 and H2 do not absorb infrared radiation and cannot be detected by this method.

Other portable systems with detectors based on thermal conduction and refractive index have found limited use in the coal mining industry.

Limitations of portable gas detection instruments

The effectiveness of portable gas detection instruments is limited by a number of factors:

  • Calibration is required. This normally involves a daily check on zero and voltage, a weekly span check and a calibration test by an authorized external authority every 6 months.
  • Sensors have a finite life. If not dated by the manufacturer, the date of acquisition should be inscribed.
  • Sensors can be poisoned.
  • Sensors may suffer from cross-sensitivity.
  • Overexposure may saturate the sensor causing its slow recovery.
  • Inclination may affect the reading.
  • Batteries require charging and regular discharging.


Centralized Monitoring Systems

Inspections, ventilation and surveys with hand-held instruments often succeed in detecting and locating a small heating with limited makes of CO before the gas is dispersed by the ventilation system or its level exceeds the statutory limits. These do not suffice, however, where a significant risk of combustion is known to occur, methane levels in the returns exceed 1%, or a potential hazard is suspected. Under these circumstances, continuous monitoring at strategic locations is required. A number of different types of centralized continuous monitoring systems are in use.

Tube bundle systems

The tube bundle system was developed in Germany in the 1960s to detect and monitor the progress of spontaneous combustion. It involves a series of as many as 20 plastic tubes made of nylon or polyethylene 1/4 or 3/8 of an inch in diameter that extend from a bank of analysers on the surface to selected locations underground. The tubes are equipped with filters, drains and flame traps; the analysers are usually infrared for CO, CO2 and methane and paramagnetic for oxygen. A scavenger pump pulls a sample through each tube simultaneously and a sequential timer directs the sample from each tube through the analysers in turn. The data logger records the concentration of each gas at each location and automatically triggers an alarm when predetermined levels are exceeded.

This system has a number of advantages:

  • No explosion-proof instruments are required.
  • Maintenance is relatively easy.
  • Underground power is not required.
  • It covers a wide range of gases.
  • Infrared analysers are usually quite stable and reliable; they maintain their specificity in a changing background of fire gases and low oxygen atmospheres (high concentrations of methane and/or carbon dioxide may be cross-sensitive to the carbon monoxide reading in the low ppm range).
  • Instruments can be calibrated on the surface, although calibration samples of gases should be sent through the tubes to test the integrity of the collection system and the system for identifying the locations where particular samples originated.


There are also some disadvantages:

  • The results are not in real time.
  • Leaks are not immediately apparent.
  • Condensation may collect in the tubes.
  • Defects in the system are not always immediately apparent and may be difficult to identify.
  • The tubes may be damaged by blasting or in a fire or an explosion.


Telemetric (electronic) system

The telemetric automatic gas monitoring system has a control module on the surface and intrinsically safe sensor heads strategically located underground which are connected by phone lines or fibre-optic cables. Sensors are available for methane, CO and air velocity. The sensor for CO is similar to the electrochemical sensor used in portable instruments and is subject to the same limitations. The methane sensor works through the catalytic combustion of methane on the active elements of a Wheatstone-bridge circuit which can be poisoned by sulphur compounds, phosphate esters or silicon compounds and will not work when the oxygen concentration is low.

The unique advantages of this system include:

  • The results are available in real time (i.e., there is rapid indication of fire or a build-up of methane).
  • Long distances between the sensor heads and the control unit are possible without compromising the system.
  • Sensor failure is recognized immediately.


There are also some disadvantages:

  • A high level of maintenance is required.
  • The sensor range for CO is limited (0.4%).
  • The variety of sensors is limited; there are none for CO2 or hydrogen.
  • The methane sensor is subject to poisoning.
  • In situ calibration is required.
  • Cross-sensitivity may be a problem.
  • There may be a loss of power (e.g., >1.25% for methane).
  • Sensor life is limited to 1 to 2 years.
  • The system is not suitable for low oxygen atmospheres (e.g., behind seals).


Gas chromatograph

The gas chromatograph is a sophisticated piece of equipment that analyses samples with high degrees of accuracy and that, until recently, could only be fully utilized by chemists or specially qualified and trained personnel.

Gas samples from a tube bundle-type of system are injected into the gas chromatograph automatically or they can be manually introduced from bag samples brought out of the mine. A specially packed column is used to separate different gases and a suitable detector, usually thermal conductivity or flame ionization, is used to measure each gas as it elutes from the column. The separation process provides a high degree of specificity.

The gas chromatograph has particular advantages:

  • No cross-sensitivity from other gases occurs.
  • It is capable of measuring hydrogen.
  • It is capable of measuring ethylene and higher hydrocarbons.
  • It can accurately measure from very low to very high concentrations of most of the gases that occur or are produced underground by a heating or a fire.
  • It is well recognized that modern methods of combating fires and heatings in coal mines may be most effectively implemented on the basis of interpretation of gas analyses from strategic locations in the mine. Accurate, reliable and complete results require a gas chromatograph and interpretation by qualified, experienced and fully trained personnel.


Its disadvantages include:

  • The analyses are relatively slow.
  • A high level of maintenance is required.
  • The hardware and the controls are complex.
  • Expert attention is required periodically.
  • Calibration must be scheduled frequently.
  • High methane concentrations interfere with low level CO measurements.

Choice of system

Tube-bundle systems are preferred for monitoring locations that are not expected to have rapid changes in gas concentrations or, like sealed areas, may have low oxygen environments.

Telemetric systems are preferred in locations such as belt roads or on the face where rapid changes in gas concentrations may have significance.

Gas chromatography does not replace existing monitoring systems but it enhances the range, accuracy and reliability of the analyses. This is particularly important when determination of the risk of explosion is involved or when a heating is reaching an advanced stage.

Sampling considerations

  • The siting of sampling points at strategic locations is of major importance. The information from a single sampling point some distance from the source is only suggestive; without confirmation from other locations it may lead to over- or underestimation of the seriousness of the situation. Consequently, sampling points to detect an outbreak of spontaneous combustion must be sited where heatings are most likely to occur. There must be little dilution of flows between the heating and the detectors. Consideration must be given to the possibility of the layering of methane and warm combustion gases which may rise up the dip in a sealed area. Ideally, the sampling sites should be located in panel returns, behind stoppings and seals, and in the main stream of the ventilation circuit. The following considerations are applicable:
  • The sampling site should be set at least 5 m inbye (i.e., toward the face of) a seal because seals “breathe in” when the atmospheric pressure rises.
  • Samples should be taken from boreholes only when they breathe out and when it can be ensured that the borehole is leak free.
  • Samples should be taken more than 50 m downwind from a fire to ensure mixing (Mitchell and Burns 1979).
  • Samples should be taken up the gradient from a fire near the roof because hot gases rise.
  • Samples should be taken inbye a ventilation door to avoid leakage.
  • All sampling points should be clearly shown on maps of schematics of the mine ventilation system. Taking gas samples underground or from surface boreholes for analysis at another location is difficult and error prone. The sample in the bag or container must truly represent the atmosphere at the sampling point.


Plastic bags are now widely used in the industry for taking samples. The plastic minimizes leakage and can keep a sample for 5 days. Hydrogen, if present in the bag, will degrade with a daily loss of about 1.5% of its original concentration. A sample in a football bladder will change concentration in half an hour. Bags are easy to fill and the sample can be squeezed into an analysing instrument or it can be drawn out with a pump.

Metal tubes that are filled under pressure by a pump can store samples for a long time but the size of the sample is limited and leakage is common. Glass is inert to gases but glass containers are fragile and it is difficult to get the sample out without dilution.

In collecting samples, the container should be pre-flushed at least three times to ensure that the previous sample is completely flushed out. Each container should have a tag carrying such information as the date and time of sampling, the exact location, the name of the person collecting the sample and other useful information.

Interpretation of Sampling Data

Interpretation of the results of gas sampling and analysis is a demanding science and should be attempted only by individuals with special training and experience. These data are vital in many emergencies because they provide information on what is happening underground that is needed to plan and implement corrective and preventive actions. During or immediately after an underground heating, fire or explosion, all possible environmental parameters should be monitored in real time to enable those in charge to accurately determine the status of the situation and measure its progress so that they lose no time in initiating any needed rescue activities.

Gas analysis results must meet the following criteria:

  • Accuracy. Instruments must be correctly calibrated.
  • Reliability. Cross-sensitivities must be known
  • Completeness. All gases, including hydrogen and nitrogen, should be measured.
  • Timeliness. If real time is not possible, trending should be carried out.
  • Validity. Sample points must be in and around the site of the incident.


The following rules should be followed in interpreting gas analysis results:

  • A few sampling points should be carefully selected and marked on the plan. This is better for trending than taking sample from many points.
  • If a result deviates from a trend, it should be confirmed by resampling or the calibration of the instrument should be checked before taking action. Variations in outside influences, such as ventilation, barometric pressure and temperature or a diesel engine running in the area, are often the reason for the changing result.
  • The gas make or mixture under non-mining conditions should be known and allowed for in the calculations.
  • No analysis result should be accepted on faith; results must be valid and verifiable.
  • It should be borne in mind that isolated figures do not indicate the progress—trends give a more accurate picture.


Calculating air-free results

Air-free results are obtained by calculating out the atmospheric air in the sample (Mackenzie-Wood and Strang 1990). This allows samples from a similar area to be properly compared after the dilution effect from air leakage has been removed.

The formula is:

Air-free result = Analysed result / (100 - 4.776 O2)

It is derived as follows:

Atmospheric air = O2 + N2 = O2 + 79.1 O2 / 20.9  =  4.776 O2

Air-free results are useful when trending of results is required and there has been a risk of air dilution between the sample point and the source, air leakage has occurred in sample lines, or bag samples and seals may have breathed in. For example, if the carbon monoxide concentration from a heating is being trended, then air dilution from an increase in ventilation could be misinterpreted as a decrease in carbon monoxide from the source. The trending of air-free concentrations would give the correct results.

Similar calculations are needed if the sampling area is making methane: the increase in methane concentration would dilute the concentration of other the gases that are present. Hence, an increasing carbon oxide level may actually show up as decreasing.

Methane-free results are calculated as follows:

Methane-free result = Analysed result / (100 - CH4 %)

Spontaneous Combustion

Spontaneous combustion is a process whereby a substance can ignite as a result of internal heat which arises spontaneously due to reactions liberating heat faster than it can be lost to the environment. The spontaneous heating of coal is usually slow until the temperature reaches about 70 °C, referred to as the “cross over” temperature. Above this temperature, the reaction usually accelerates. At over 300 °C, the volatiles, also called “coal gas” or “cracked gas”, are given off. These gases (hydrogen, methane and carbon monoxide) will ignite spontaneously at temperatures of approximately 650 °C (it has been reported that the presence of free radicals can result in the appearance of flame in the coal at about 400 °C). The processes involved in a classic case of spontaneous combustion are presented in table 1 (different coals will produce varying pictures).

Table 1. Heating of coal - hierarchy of temperatures

Temperature at which coal absorbs O2 to form a complex and produce heat

30 °C

Complex breaks down to produce CO/CO2

45 °C

True oxidation of coal to produce CO and CO2

70 °C

Cross-over temperature, heating accelerates

110 °C

Moisture, H2 and characteristic smell released

150 °C

Desorbed CH4, unsaturated hydrocarbons released

300 °C

Cracked gases (e.g., H2, CO, CH4) released

400 °C

Open flame

Source: Chamberlain et al. 1970.

Carbon monoxide

CO is actually released some 50 °C before the characteristic smell of combustion is noticed. Most systems designed to detect the onset of spontaneous combustion are based on the detection of carbon monoxide in concentrations above the normal background for a particular area of the mine.

Once a heating has been detected, it must be monitored in order to determine the state of the heating (i.e., its temperature and extent), the rate of accelerations, toxic emissions and explosibility of the atmosphere.

Monitoring a heating

There are a number of indices and parameters available to assist planners to determine the extent, temperature and rate of progression of a heating. These are usually based on changes in the composition of the air passing through a suspected area. Many indicators have been described in the literature over the years and most offer a very limited window of usage and are of minimal value. All are site specific and differ with different coals and conditions. Some of the more popular ones include: carbon monoxide trending; carbon monoxide make (Funkemeyer and Kock 1989); Graham’s ratio (Graham 1921) tracer gases (Chamberlain 1970); Morris ratio (Morris 1988); and the carbon monoxide/carbon dioxide ratio. After sealing, indicators may be difficult to use because of the absence of a defined air flow.

No one indicator affords a precise and sure method of measuring the progress of a heating. Decisions must be based on gathering, tabulating, comparing and analysing all information and interpreting it in the light of training and experience.


Explosions are the greatest single hazard in coal mining. It has the potential to kill the entire underground workforce, destroy all the equipment and services and prevent any further working of the mine. And, all this can happen in 2 to 3 seconds.

The explosibility of the atmosphere in the mine must be monitored at all times. It is especially urgent when workers are engaged in a rescue operation in a gassy mine.

As in the case of indicators for evaluating a heating, there are a number of techniques for calculating the explosibility of the atmosphere in an underground mine. They include: Coward’s triangle (Greuer 1974); Hughes and Raybold’s triangle (Hughes and Raybold 1960); Elicott’s diagram (Elicott 1981); and Trickett’s ratio (Jones and Trickett 1955). Because of the complexity and variability of the conditions and circumstances, there is no single formula that can be relied on as a guarantee that an explosion will not occur at a particular time in a particular mine. One must rely on a high and unremitting level of vigilance, a high index of suspicion and an unhesitating initiation of appropriate action at the slightest indication that an explosion might be imminent. A temporary halt in production is a relatively small premium to pay for assurance that an explosion will not occur.


This article has summarized the detection of gases that might be involved in fires and explosions in underground mines. The other health and safety implications of the gaseous environment in mines (e.g., dust diseases, asphyxia, toxic effects, etc.) are discussed in other articles in this chapter and elsewhere in this Encyclopaedia.



Read 10407 times Last modified on Wednesday, 03 August 2011 18:21

" DISCLAIMER: The ILO does not take responsibility for content presented on this web portal that is presented in any language other than English, which is the language used for the initial production and peer-review of original content. Certain statistics have not been updated since the production of the 4th edition of the Encyclopaedia (1998)."


Mining and Quarrying References

Agricola, G. 1950. De Re Metallica, translated by HC Hoover and LH Hoover. New York: Dover Publications.

Bickel, KL. 1987. Analysis of diesel-powered mine equipment. In Proceedings of the Bureau of Mines Technology Transfer Seminar: Diesels in Underground Mines. Information Circular 9141. Washington, DC: Bureau of Mines.

Bureau of Mines. 1978. Coal Mine Fire and Explosion Prevention. Information Circular 8768. Washington, DC: Bureau of Mines.

—. 1988. Recent Developments in Metal and Nonmetal Fire Protection. Information Circular 9206. Washington, DC: Bureau of Mines.

Chamberlain, EAC. 1970. The ambient temperature oxidisation of coal in relation to the early detection of spontaneous heating. Mining Engineer (October) 130(121):1-6.

Ellicott, CW. 1981. Assessment of the explosibility of gas mixtures and monitoring of sample-time trends. Proceeding of the Symposium on Ignitions, Explosions and FIres. Illawara: Australian Institute of Mining and Metallurgy.

Environmental Protection Agency (Australia). 1996. Best Practice Environmental Management in Mining. Canberra: Environmental Protection Agency.

Funkemeyer, M and FJ Kock. 1989. Fire prevention in working rider seams prone to spontaneous combustion. Gluckauf 9-12.

Graham, JI. 1921. The normal production of carbon monoxide in coal mines. Transactions of the Institute of Mining Engineers 60:222-234.

Grannes, SG, MA Ackerson, and GR Green. 1990. Preventing Automatic Fire Suppression Systems Failure on Underground Mining Belt Conveyers. Information Circular 9264. Washington, DC: Bureau of Mines.

Greuer, RE. 1974. Study of Mine Fire Fighting Using Inert Gases. USBM Contract Report No. S0231075. Washington, DC: Bureau of Mines.

Griffin, RE. 1979. In-mine Evaluation of Smoke Detectors. Information Circular 8808. Washington, DC: Bureau of Mines.

Hartman, HL (ed.). 1992. SME Mining Engineering Handbook, 2nd edition. Baltimore, MD: Society for Mining, Metallurgy, and Exploration.

Hertzberg, M. 1982. Inhibition and Extinction of Coal Dust and Methane Explosions. Report of Investigations 8708. Washington, DC: Bureau of Mines.

Hoek, E, PK Kaiser, and WF Bawden. 1995. Design of Suppoert for Underground Hard Rock Mines. Rotterdam: AA Balkema.

Hughes, AJ and WE Raybold. 1960. The rapid determination of the explosibility of mine fire gases. Mining Engineer 29:37-53.

International Council on Metals and the Environment (ICME). 1996. Case Studies Illustrating Environmental Practices in Mining and Metallurgical Processes. Ottawa: ICME.

International Labour Organization (ILO). 1994. Recent Developments in the Coalmining Industry. Geneva: ILO.

Jones, JE and JC Trickett. 1955. Some observations on the examination of gases resulting from explosions in collieries. Transactions of the Institute of Mining Engineers 114: 768-790.

Mackenzie-Wood P and J Strang. 1990. Fire gases and their interpretation. Mining Engineer 149(345):470-478.

Mines Accident Prevention Association Ontario. n.d. Emergency Preparedness Guidelines. Technical Standing Committee Report. North Bay: Mines Accident Prevention Association Ontario.

Mitchell, D and F Burns. 1979. Interpreting the State of a Mine Fire. Washington, DC: US Department of Labor.

Morris, RM. 1988. A new fire ratio for determining conditions in sealed areas. Mining Engineer 147(317):369-375.

Morrow, GS and CD Litton. 1992. In-mine Evaluation of Smoke Detectors. Information Circular 9311. Washington, DC: Bureau of Mines.

National Fire Protection Association (NFPA). 1992a. Fire Prevention Code. NFPA 1. Quincy, MA: NFPA.

—. 1992b. Standard on Pulverized Fuel Systems. NFPA 8503. Quincy, MA: NFPA.

—. 1994a. Standard for Fire Prevention in Use of Cutting and Welding Processes. NFPA 51B. Quincy, MA: NFPA.

—. 1994b. Standard for Portable Fire Extinguishers. NFPA 10. Quincy, MA: NFPA.

—. 1994c. Standard for Medium and High Expansion Foam Systems. NFPA 11A. Quncy, MA: NFPA.

—. 1994d. Standard for Dry Chemical Extinguishing Systems. NFPA 17. Quincy, MA: NFPA.

—. 1994e. Standard for Coal Preparation Plants. NFPA 120. Quincy, MA: NFPA.

—. 1995a. Standard for Fire Prevention and Control in Underground Metal and Nonmetal Mines. NFPA 122. Quincy, MA: NFPA.

—. 1995b. Standard for Fire Prevention and Control in Underground Bituminious Coal Mines. NFPA 123. Quincy, MA: NFPA.

—. 1996a. Standard on Fire Protection for Self-propelled and Mobile Surface Mining Equipment. NFPA 121. Quincy, MA: NFPA.

—. 1996b. Flammable and Combustible Liquids Code. NFPA 30. Quincy, MA: NFPA.

—. 1996c. National Electrical Code. NFPA 70. Quincy, MA: NFPA.

—. 1996d. National Fire Alarm Code. NFPA 72. Quincy, MA: NFPA.

—. 1996e. Standard for the Installation of Sprinkler Systems. NFPA 13. Quincy, MA: NFPA.

—. 1996f. Standard for the Installation of Water Spray Systems. NFPA 15. Quincy, MA: NFPA.

—. 1996g. Standard on Clean Agent Fire Extinguishing Systems. NFPA 2001. Quincy, MA: NFPA.

—. 1996h. Recommended Practice for Fire Protection in Electric Generating Plants and High Voltage DC Converter Stations. NFPA 850. Quincy, MA: NFPA.

Ng, D and CP Lazzara. 1990. Performance of concrete block and steel panel stoppings in a simulated mine fire. Fire Technology 26(1):51-76.

Ninteman, DJ. 1978. Spontaneous Oxidation and Combustion of Sulfide Ores in Underground Mines. Information Circular 8775. Washington, DC: Bureau of Mines.

Pomroy, WH and TL Muldoon. 1983. A new stench gas fire warning system. In Proceedings of the 1983 MAPAO Annual General Meeting and Technical Sessions. North Bay: Mines Accident Prevention Association Ontario.

Ramaswatny, A and PS Katiyar. 1988. Experiences with liquid nitrogen in combating coal fires underground. Journal of Mines Metals and Fuels 36(9):415-424.

Smith, AC and CN Thompson. 1991. Development and application of a method for predicting the spontaneous combustion potential of bituminous coals. Presented at the 24th International Conference of Safety in Mines Research Institutes, Makeevka State Research Institute for Safety in the Coal Industry, Makeevka, Russian Federation.

Timmons, ED, RP Vinson, and FN Kissel. 1979. Forecasting Methane Hazards in Metal and Nonmetal Mines. Report of Investigations 8392. Washington, DC: Bureau of Mines.

United Nations (UN) Department of Technical Cooperation for Development and the German Foundation for International Development. 1992. Mining and the Environment: The Berlin Guidelines. London: Mining Journal Books.

United Nations Environment Programme (UNEP). 1991. Environmental Aspects of Selected Non-ferrous Metals (Cu, Ni, Pb, Zn, Au) in Ore Mining. Paris: UNEP.