| Summary: | There are, on average, over twenty-five accidental gas explosions every year in the UK, each requiring an investigation to determine the origin and cause, in order to satisfy regulatory requirements, or for the purposes of criminal or civil litigation. The most important conclusion to be drawn from the investigation is the identification of the source of the gas release. This is most often determined through interpretation of the severity of pressure and thermal damage to the building and its contents. Guidance in reference books states that gas explosions exhibit characteristic pressure and thermal damage dependent upon the concentration of the fuel/air mixture prior to ignition. It is believed that a number of investigations have resulted in the incorrect apportion of blame as a consequence of the misinterpretation of forensic evidence. The key objectives of the study were to answer three questions. Firstly, can ignition of a fuel lean or fuel rich mixture cause significant structural damage to a building? Secondly, is it possible to determine the gas concentration in a building, prior to ignition, from the severity of the thermal damage? Thirdly, do materials exposed to a transient flame front always exhibit thermal damage? The results of four experimental programmes, and over one hundred and fifty explosion tests, are reported in this thesis. Explosion tests were conducted in explosion chambers ranging from 1 m3 to 180 m3. Experiments were conducted in single and interconnected enclosures, with and without obstacles (including furniture) and with a number of ‘marker’ boards to assess the severity of thermal damage. A number of parameters were varied; including, fuel type, concentration, distribution of gas, congestion, ignition position and vent size and failure pressure. The results demonstrate that under the right conditions, fuel lean and fuel rich explosions can cause overpressures that have the potential to structurally damage buildings (> 200 mbar). A number of mechanisms have been proposed, detailing the manner in which gas explosions propagate from one room to another. This knowledge provides a valuable new insight into how complex a vented explosion in a typical building can be, and how the design and construction of a building can affect the magnitude of the explosion. Several causes of the development of high overpressures have been identified; the ignition of a flammable cloud outside the vent opening(s), the sudden increase in mass combustion as a turbulent mixture in a secondary compartment is ignited by a propagating flame front passing through an interconnecting doorway, and the highly turbulent ‘jetting’ expanding flame, driven by the venting process, propagating through a doorway and towards a vent opening in a secondary enclosure. Evidence is presented that shows it is possible to generate pressures capable of causing structural damage to buildings with volume blockages of as little as 0.57%, if vent openings do not allow sufficient outflow. However, the obstacle geometry, and its location to other obstacles and the enclosure, were found to be critical in the development, or otherwise, of damaging overpressures. The experiments have demonstrated that it is possible to use the severity and extent of thermal damage to wall coverings and wood surfaces, sustained during a gas explosion, to provide useful information on the gas concentration, its distribution throughout the building prior to ignition and the depth of any flammable layer. It is demonstrated that it is possible to assess the severity of the thermal damage to various materials in order to estimate the natural gas concentration prior to the explosion to the nearest 2%. The most suitable materials, in terms of forensic indicators, appear to be softwood covered with either gloss varnish or white oil based gloss paint. Such surfaces are common in buildings as door frames, window frames, etc. However, it is shown that quick drying paints are less susceptible to thermal damage and may cause the misinterpretation of evidence which could lead to an incorrect diagnosis of the origin and cause of an explosion. In tests where a flammable gas/air layer was present, thermal damage may be observed above the nominal layer boundary (for natural gas), down to the lowest level where the concentration was originally above 8% ± 1%. It is shown to be possible to estimate the layer depth from the damage to an accuracy of approximately 15 cm. The results of the experimental programmes presented in this thesis, provide new knowledge and understanding of the development of gas explosions in buildings and how this knowledge may be used to better interpret forensic evidence found in a gas explosion.
|
|---|
