Keywords: model quality, model evaluation, REDIPHEM project, safety studies, hazard assessment, dense gas dispersion, European Community, quality assessment, industrial hazards, atmospheric dispersion models, air pollution, environmental pollution, modelling
Model quality and safety studies
The paper describes the EC initiative on model quality assessment and emphasises some of the problems encountered in the selection of data from field tests used in the evaluation process. Further, it discusses the impact of model uncertainties in safety studies of industrial plants. The model evaluation group (MEG) is a European initiative on evaluation of technical models used within the major industrial hazard area and is supported by the EC DGXII, Directorate-General for Science, Research and Development. Closely related to MEG is REDIPHEM, which is the acronym of a research project supported by the EC Environment Research Programme focusing on assessment of the quality of experimental data and models for dense gas dispersion created within the EC funded research programmes. Technical models are used in a number of areas of industrial hazards assessment. It is becoming more and more apparent that most of these have never been through a procedure of evaluation, but nonetheless are used to assist in making decisions that may directly affect the safety of the public and the environment. As a major funder of European research on major industrial hazards, DGXII is conscious of the importance of ensuring that model development is of a standard that is commensurate with the importance of model use. The aim of MEG is to improve the culture in which models are developed and used and so ensure that technical models used in all aspects of major hazard evaluation are up to date with technical developments and used by personnel well versed in their applicability and functioning. In 1994, MEG finalised a generic protocol for model evaluation and guidelines for model developers. The two documents prepared by MEG have been through a preliminary review by 10-20 experts and will be further developed based on the experience obtained in the working groups. MEG has so far initiated three working groups on specific subject areas: dense gas dispersion, liquefied pool fires and vapour cloud explosions. This paper describes the model evaluation protocol intended for users who wish to assess if a certain model is appropriate for use in solving a given problem. Further, the findings from the REDIPHEM project related to dense gas dispersion will be highlighted. Finally, the paper will discuss the need for model quality assessment in safety studies.