In British underground coalmines, fire alarms detect fewer fires than mine personnel, particularly during the early combustion stages of smouldering and smoke emission. However, as collieries become more automated, there are fewer personnel to detect fires. A recent survey by UK Coal showed that current fire detectors, based on semiconductor sensors that detect gaseous products of combustion, are under-utilised. This is because they are susceptible to interferents, resulting in false alarms and instrumental drift and are not user-friendly. They are also virtually obsolete. Carbon monoxide sensors, however, are better for detecting spontaneous combustion and more developed fires. More recently, a withdrawal of men from one coalmine occurred when equipment failed to detect a significant conveyor failure, which produced large amounts of smoke. This joint project, involving HSE, the Health and Safety Laboratory (HSL), UK Coal Ltd and TES Bretby Ltd, was therefore initiated to investigate the performance of proposed solutions to the sensor shortage. The solution must have the sensitivity of the existing system but not be susceptible to typical interferents, e.g. diesel exhaust, methane, coal dust.
Previous work had tested a number of possible sensors before developing a system based upon a combination of an optical high sensitivity smoke detector (HSSD) fitted with a cyclone to remove coal dust; and nitric oxide or nitrogen dioxide electrochemical sensors to distinguish smoke from diesel exhaust. Initial work on preparing a suitable HSSD had already been carried out and this project was designed to test the system underground in a colliery. The HSSD was tested without a cyclone and with a 2 micron dust-size, cut-off cyclone. When there was no cyclone fitted the coal dust concentrations in the colliery were too high and variable to be able to distinguish smoke. As expected the addition of a cyclone greatly reduced the coal dust reaching the HSSD making it much easier to distinguish smoke. However, the cyclone still allowed a significant fraction of coal dust to reach the HSSD resulting in a delay to alarm of at least two minutes.
A 0.5 micron cut-off cyclone was designed and shown to distinguish smoke from coal dust in laboratory tests. Underground testing has shown that the 0.5 micron cut-off cyclone allowed less coal dust through than 2 micron cyclone, but it has not yet been tested with smoke underground. A problem with the smaller cyclone is that it needs a significantly larger and more powerful pump to maintain the airflow through the HSSD, which complicates commercialisation of the system. Future work will also look at the possibility of reducing the pump size and airflow whilst maintaining the HSSD sensitivity. If successful, it is hoped that commercialisation of the multi-sensor system comprising of a HSSD and oxides of nitrogen and carbon monoxide sensors, would ensue.
HSL carried out an examination of failed hoses from two incidents at Rawdon Mines Rescue Station in 2006. Both incidents occurred during the re-filling of portable oxygen cylinders. In both cases the flexible hose connecting the cylinder to the charging panel failed close to its attachment point at the cylinder. An oxygen fire accompanied both failures. The items received for examination comprised the failed hose, metal connectors and oxygen cylinder and t-piece from both incidents. The interior of the connectors and t-pieces of both cylinders were heavily contaminated with combustion products. HSE’s Mines Inspectorate asked HSL to determine the ignition source, the direction of the flame path through the components and the materials used in the construction of the cylinder valve seating and the connector unit. They also asked whether these materials could have caused the ignition in the absence of any contamination.
The combustion debris within the components was comprised of distinct layers that suggested that multiple or sequential combustion events had occurred during the course of the incident. There were indications that the combustion process occurred during both the filling operation and discharging of the cylinder following the failure of the hose. An examination of the possible causes for the ignition led to the conclusion that adiabatic compression heating of the hose was responsible for the failures.
A comparison was made between the operating procedure at Rawdon and that at the Mansfield Mines Rescue Station where there had been no reported hose failures. This concluded that the Rawdon procedure contained a greater opportunity for adiabatic compression heating of residual gases in the hose because the pressure rise during re-filling potentially occurred in a single step rather than the more gradual three-stage rise using the sequence at Mansfield. The recommendation was made that Rawdon adopt the Mansfield filling procedure.
In addition to the operating sequence leading to failure, the examination of the connector unit revealed that within its central bore was a non-return valve. It was considered that this could increase the risk of adiabatic heating, as it was able to prevent the equalisation of pressure between the oxygen cylinder and hose, at an intermediate pressure, prior to the recharging operation. This non-return valve, although present at the Mansfield site apparatus was felt not to possess as great a threat as at the Rawdon site because of the different filling sequence. The report nevertheless recommended that removal of the valve from the apparatus be considered.
In the second incident, the hose that failed was not of the usual PTFE lined braided stainless wire outer sheath construction. Instead it was a Kevlar reinforced polyurethane hose with a nylon inner core. According to the manufacturer’s website this hose is recommended for use with mixed gases. As the hose was being used here in a more highly reactive oxygen environment, and the materials used in its construction had lower melting points than the original hoses, it was considered that the hose was inappropriate for this purpose, and that it was the same mechanism involving adiabatic compression heating that caused the hose to fail.
HSL carried out investigations for HSE’s mines inspectorate following two separate incidents at the end of 2005 that involved foams catching fire in British coal mines. This is a highly charged political situation, as polyurethane (PU) foams have been banned in UK collieries for many years following a multiple fatality fire.
The first incident at Kellingley colliery involved Rocsil, a non-PU foam, which caught alight in a particular and unusual set of circumstances. The investigation, completed in 2006, showed that the fire was difficult to start and easily contained, but the job is being held open pending a conclusion on wider issues around foams.
The second incident occurred at Daw Mill Colliery and involved a PU resin, widely used for filling fissures in the rock to minimise the risk of roof/wall collapse during mining. The polyol and isocyanate components that form the resin were accidentally mixed with water, which led to the production of a foam rather than a solid resin. A block of this foam caught fire as a result of self-heating. HSL's investigation, completed in 2006, showed that mixing water in with the two components that form the resin will lead to foam and that the reaction can lead to high temperatures within the foam. Together, with an imbalance in the component proportions, HSL measured temperatures of nearly 200°C. This heat can be trapped in the insulating foam for some time, and is higher than the temperature that would be of concern for self-heating in Daw Mill's coal. Again the job is held open as the incident is not considered 'finished' due to wider concerns.
A follow on project from the Daw Mill investigation outlined above was to investigate the use of PU resins in water bearing strata. Concern was raised about the potential for PU resin being injected into the strata and passing through fissures into a large 'void' which might contain water. The project addressed the question of whether this would lead to the production of foam which might overheat and catch light within the void, or subsequently be released during coal cutting or collapse. A mock-up fissure / void system was built and PU resin injected. This showed that foam could be produced and the results of this were passed to the Mines Inspectorate for consideration.