Olavi Keski-Rahkonen

Probabilistic simulation of fire scenarios

A risk analysis tool is developed for computation of the distributions of fire model output variables. The tool, called Probabilistic Fire Simulator (PFS), combines Monte Carlo simulation and CFAST, a two-zone fire model. In this work, the tool is used to estimate the failure probability of redundant cables in a cable tunnel fire, and the failure and smoke filling probabilities in an electronics room during an electronics cabinet fire. Sensitivity of the output variables to the input variables is calculated in terms of the rank order correlations. The use of the rank order correlations allows the user to identify both modelling parameters and actual facility properties that have the most influence on the results. Various steps of the simulation process, i.e. data collection, generation of the input distributions, modelling assumptions, definition of the output variables and the actual simulation, are described.

Landfill Fires in Finland

On average, there were 633 sanitary landfills in operation in 1990-92 in Finland. Annually, 380 landfill fires occurred, one-quarter of which were deep fires. The total amount of waste burnt in landfill fires was estimated to be 84,000 tonnes year-1. Unsufficient covering and compacting, ash disposal and deliberate fire starting were reported to be the most typical reasons for waste ignition. The most severe deep fires lasted for 2 months. A 10 m high, 35,000 m3 large experimental bank of waste was built, instrumented and set on fire from a 7 m deep well located at the centre of the area. Wood was burnt in this ignition well to produce glowing embers, which were covered with waste and soil layers. No spread of fire was detected after covering the ignition well. However, the fire was not extinguished by covering but remained smouldering throughout the experiment. To extinguish the fire, it was necessary to dig up the smouldering fire and cool it.

A Proposal for the Goals and New Techniques of Modelling Pedestrian Evacuation in Fires

In this article, we propose the goals for evacuation simulations in the context of the fire safety engineering. It is proposed that the safety of a building design should be measured using F-N plots that are based on the fire statistics. A new evacuation code is developed that allows the modelling of ‘panic’ situations and interaction between evacuation simulation and the state-of-the-art fire simulation. The major features of the new code are described and first preliminary results are shown. The method presented was found to run satisfactorily, and fast enough for practical purposes. When the results were compared against the results obtained using Simulex and buildingExodus codes, a good agreement was found in two of the three cases but for a case with congested corridor considerable differences occurred.

Electrical ignition sources in nuclear power plants: statistical, modelling and experimental studies

Electrical ignitions of fires in nuclear power plants (NPPs) were studied by analysing statistical information from both nuclear and non-nuclear installations, modelling the most common simple physical ignition processes, and finally carrying out experiments on some of the modelled scenarios. Statistical assessment indicated cables to be a significant cause of electrical ignitions in all kinds of installations. Modelling of some relevant scenarios of cable fires and carrying out a number of experiments gave quantitative information on the circumstances and physical processes of the dangerous cable ignition situations.

Two-model Monte Carlo Simulation of Fire Scenarios

A risk analysis tool called Probabilistic Fire Simulator (PFS) is developed for the computation of the distributions of fire model output variables and the sensitivities of the output variables to the inputs. PFS performs a Monte Carlo simulation using different fire models, including CFAST two-zone model and FDS fluid dynamics model. In this work, a new technique is developed for the use of two different fire models in the same Monte Carlo simulation. The two-model Monte Carlo technique provides a computationally effective means to improve the accuracy of the fast but inaccurate models, using the results of the more accurate but computationally more demanding models. The technique is tested in three scenarios: approximation of an analytical function, calculation of a ceiling jet temperature and a simulation of a simple room fire.