FIRE AND EXPLOSION
Ventilation controlled fires in tunnels
This Health and Safety Laboratory (HSL) project is jointly funded by the European Coal and Steel Community (ECSC) and the Health and Safety Executive (HSE) mines and railways inspectorates. The notion of ventilation control of fires in naturally-ventilated compartments is well known and their characteristics have been investigated over many years. The investigation of the influence of changes in ventilation on forced-ventilated fires in ducts and tunnels has received less attention. However, it may have important implications for the tactics deployed in the combating of such fires and the control of the resulting smoke hazards. This work addresses this problem largely through a programme of small-scale experiments.
During 2002, work has concentrated on the final design and construction of an experimental facility and experimental programme to study the influence of changes in ventilation on the behaviour and characteristics of tunnel fires and initial experiments with this rig.
A small experimental duct was used for the experiments with heat output and ventilation velocity scaled according to the Froude number with a scale factor of 1:20. The operational principle of this rig is to use two seats of fire. The duct is ventilated using a constantly variable fan in the exhaust. Flow, O2, CO, CO2, smoke and temperature are all monitored. The intention is to use this rig to examine the effect of ventilation on the characteristics of a tunnel fire and particularly to investigate the occurrence of vitiated/fuel-rich tunnel fires and their response to ventilation. Some initial results have been obtained.
A series of experiments have been carried out using a variety of fuels. These indicate that the flame length is approximately linearly dependent on heat release rate and weakly inversely dependent on the ventilation velocity. This indicates that as the flame becomes more ventilated, then there is greater mixing efficiency and the visible flame envelope is smaller. Clearly, for a fire of given size, the increasing of the ventilation velocity for over-ventilated flames results in shorter flames.
In combination with the ventilation rate the oxygen concentration in the second fire zone can be controlled at levels between about 15% and ambient. Tests have been carried out with both liquid fuels as pools and soaked into mineral wool wicks and solid fuels. For over-ventilated fires it is observed that flame spread rate increases with the ventilation rate. However, for vitiated fires the rate of flame spread increases significantly. Further increases of ventilation suggest then that the rate of flame spread begins to decrease.
Thus for over-ventilated fires modest increases in flame spread rates are observed with increases in ventilation rate. However, for the vitiated fire the flame spread initially responds very rapidly to changes in ventilation velocity and there are suggestions that at higher velocities the flame spread rate begins to decrease.
Where oxygen supply to the flame is limited it appears that traditional views of flame spread are not generally valid. Further experiments are required to investigate these effects more fully since the occurrence of rapid flame spread in fuel rich conditions may occur in accident situations. This work will be carried out during 2003.
Prediction of Fire Effects in mines
The University of Nottingham (UoN) is also a partner in the ECSC underground fires project. Its contribution focuses on using computational fluid dynamics (CFD) to replicate a series of experimental conveyor belt fire tests, which were previously conducted in the former fire test gallery at Swadlincote test facility by International Mining Consultants Ltd.
The FLUENT CFD code does not possess a standard combustion model to simulate the combustion of, and flame spread along, a solid surface. Following a critical review of the results of previous belt fire experimental and modelling studies, it was decided to investigate the modelling of the belt combustion and flame spread by modifying an existing model developed for pulverised coal combustion. A model to represent both the aerodynamic and combustion characteristics of a section of belting has been constructed. A solid conducting slab has initially been used to represent the woven polypropylene core of the Type 10 belt. The PVC surface cover of the belt is represented by a series of linear arrays of combustible particles. Once initiated by the heat supplied by a propane gas burner, the particle combustion model simulates at a specified temperature, an initial release of volatile chemical species, which subsequently react with the available oxygen to release heat into the domain. The model simulations conducted to date have demonstrated that the adapted particle model may be used to represent the combustion and flame spread of a solid surface. Further modelling studies will be conducted in the remaining months to validate both the flame spread model and to investigate the incorporation of a belt burn through model.
To quantify the parameters required by the particle combustion model, initial experimental studies were carried out with a number of material samples of both the PVC covering and polypropylene core from a length of Type 10 belt, supplied by Fenner Dunlop. The total calorific value of a total core section taken through the belt was evaluated. In addition, a Thermo Gravimetric Analysis (TGA) was performed on samples of both the PVC and Polypropylene core. From an analysis of the results obtained from these tests, the temperature of initiation, the rate of emission and type of the principal devolatilisation species may be identified. A further analysis of the TGA results will provide a proximate analysis of the belt material.
Development of an underground fire safety risk assessment
The aim of this HSL project, to develop a fire risk assessment strategy for use in coal and other mines, has been achieved and the final report distributed within HSE and collaborating mines.
The draft methodology, described in last year's review, has been tested at a number of mines and a number of comments have been incorporated. The final system is capable of assessing both underground fire and explosion risks, rather than fire alone as originally requested. The system follows standard fire engineering practices and involves the user completing a number of checksheets covering eight different aspects, some of which are summarised below.
What is the probability of a fire occurring - possible ignition sources and fuels?
How do we detect/control it - identification of fire detection/suppression equipment?
How many people are at risk - notification of fire/explosion, means of escape?
Identification of emergency actions and incorporation in mines emergency plan.
Identification of ongoing safety requirements, e.g. training, testing fire fighting equipment etc.
The final report has been published on HSE's website at http://www.hse.gov.uk/research/hsl_pdf/2002/hsl02-24.pdf
Ignition hazard from conveyor idler rollers
This project was set up with HSL to investigate the potential for ignition of flammable dust or gas mixtures with air by the heat produced during failure of idler roller bearings. To achieve this, an apparatus was developed for failing bearings under controlled laboratory conditions. This apparatus now provides a reproducible means of failing idler roller bearings in a reasonable time. The procedure involves operating the bearings under excessive load, having first thoroughly cleaned the bearings to remove all lubrication.
Using this approach, failure of the bearings has been observed to occur in two stages:
The first stage occurs as a result of failure of the bearing cage, leading to gathering and contact of the ball bearings in the bottom of the bearing race. The ball bearings then cease to roll freely.
The second stage usually occurs shortly after the first stage and involves complete collapse of the bearing and loss of the ball bearings. The axle/ bearing inner race then rubs against the outer race of the bearing and strong friction occurs between the two surfaces.
The temperatures generated have been measured in real time using infrared thermal imaging and contact thermometry and by post-failure metallurgical analysis of bearing parts. From these analyses, temperatures during the first stage of failure have been measured in the range of 200 to 300 oC. During the second stage of failure temperatures as high as 800 oC have occurred.
The programme of work is continuing to examine the ignition of flammable atmospheres during bearing failure, and has been expanded to look at aspects of bearing design on bearing performance leading up to and during failure.
Check tests on Wincoal 'A'
Since the transfer of production of the type P5 explosive, Wincoal 'A', from the UK to France in 1999, the explosive has been manufactured using raw materials sourced in France. During a periodic check test at HSL in January 2002, a sample of Wincoal 'A' failed the P5 series (i) test in methane-air causing 5/18 ignitions/shots, when the normal requirement stipulated in HSE Testing Memorandum TM2 is 0/20.
At the request of HSE's mines inspectorate, a repeat check test was carried out on a different production batch of the explosive. The second batch also failed the test, causing 5/11 ignitions/shots. HMIM instructed the manufacturer to withdraw all existing stocks of Wincoal 'A' from the mines, but agreed to its use as a P3 explosive provided that the explosive was appropriately re-labelled, suitable controls were introduced for its use as a P3, and shotfirers were given appropriate training.
Further tests, mainly funded by the manufacturer, were also conducted on three experimental batches of the explosive. One batch had been manufactured using ammonium nitrate from a source of supply in the UK and two had been made with alternative grades of ammonium nitrate supplied by a French company. Only the batch manufactured with the British ammonium nitrate passed the test satisfactorily. Production of the explosive was resumed using this material.
CFD modelling of gas flows in coal mine goafs
This project was started in April 1999 and completed in June 2002. It was co-funded by HSE's mines inspectorate and the ECSC. The main objectives of HSL's contribution were to develop methodologies for the use of computational fluid dynamics (CFD) to improve understanding and control of methane emissions from the goaf and longwall faces. In addition CFD was used to model inert gas injection into the goaf.
An approach to simulation of methane drainage combining geotechnical modelling and CFD was demonstrated. Geotechnical modelling, based on specified strata, was used to predict the development of stress, strain and hence fracture, due to mining. The results from this modelling were then combined with empirical expressions for the variation in the permeability of coal measure rocks with stress. Combined with information on fracture patterns from the geotechnical modelling this allowed the specification of permeability values, as affected by mining, in CFD simulations. The flow of methane through strata, with and without methane drainage, was simulated, allowing the effects of methane drainage to be examined.
Simulations of flow in the workings were also performed. Flow behaviour and methane build-up at the back-return were simulated, including the effect of leaks in the brattice in this region of the return. Injection of inert gas in the goaf was also simulated.
The CFD simulations performed during the project were three-dimensional, demonstrating that computers now have sufficient power for such simulations to be undertaken. The combined use of CFD and geotechnical modelling allowed simulations to be performed with an improved basis than previous studies in which specification of rock properties was purely empirical. Geotechnical modelling alone can provide information on where to drill to improve drainage performance. CFD additionally offers a way of optimising the number and positioning of boreholes.
A final project partners' meeting was attended in Germany in May 2002. A one day conference to review applications of CFD to mining for interested parties in the UK was to be held early in 2003.
Visual assessment of flammable dust clouds
This project was set up to investigate the feasibility of using visual judgment to establish if a flammable coal dust cloud exists within a mine. The intention was that this could then be used by mine staff caught in emergency situations where they are in, or close to a coal dust cloud to decide if items of potentially incendive equipment should be switched off or not.
The work involved developing an enclosure with an internal mixing system to allow coal dust clouds of controlled concentrations to be maintained so that a visual assessment could be made. To produce such clouds a system of fans and airlines was developed to keep the cloud in suspension. Weighed quantities of coal dust were then introduced to produce the concentration required. Illumination was provided using a miner's cap lamp in isolation and with background lighting.
Working with HSL photographers, a series of tests were carried out to try and assess changes in visibility for dust concentrations. The results of these tests showed that at relatively low concentrations , around 10 gm-3, objects at a distance of approx 1.5 m were visible. At 33 gm-3, objects at 0.75 m were totally obscured. It was also possible to observe varying levels of obscuration at intermediate concentrations, however, the level of analysis required was far in excess of what a worker could rapidly achieve in an emergency situation.
Cataloguing explosion references
HSL's Explosion Control Section holds a large number of papers and reports that have been gathered together from research projects over many years. The early projects were related to coal mining and this is reflected in much of the material. Other material covers dust explosions, gas explosions and sources of ignition. Most of the material on sources of ignition has been catalogued into three databases. The remainder is sorted into subject areas, but is not catalogued. The aim of the project is to organise the material and catalogue it on a single database so the relevant papers can be retrieved from the filing system after a database search.