A stochastic grey system model for fire safety indexing evaluation on buildings

評估樓宇消防安全指數的隨機灰色系統建模

Student thesis: Doctoral Thesis

View graph of relations

Author(s)

  • Tin Cheung CHEUNG

Detail(s)

Awarding Institution
Supervisors/Advisors
Award date2 Oct 2008

Abstract

Fire safety is a complex system encompassing a large number of factors, particularly for special or large and complex buildings. Primarily, it encompasses a number of objectives, which varies with different sectors of the community. To facilitate evaluation on the fire safety level of buildings, a fire safety evaluation hierarchy is developed. For this hierarchy structure on fire safety, the highest level is the goal, then the objectives, at next level the components (tactics) and finally the attributes. The attributes represent positive and negative fire safety features that account for an acceptable large portion of the total fire safety. Whilst the objectives under the goal of fire safety are “life safety”, “property protection” and “fire prevention”, the components supporting these objectives are “means of escape” (MOE), “fire resisting construction” (FRC), “means of access for fire fighting and rescue” (MOA), “fire service installations” (FSI) and “fire safety management” (FSM). With the systematic decomposition of attributes for the MOE, FRC, MOA, FSI and FSM components, a multi-attribute model consisted of 37 attributes is built up. By developing quantitative criteria for assessing the performance of the attributes in terms of linguistic scores, the fire safety level of a building can be evaluated. Due to the presence of various uncertainties, the fire safety indexing evaluation has to be in the form of a non-deterministic system. The proposed fire safety indexing evaluation is in fact a semi-quantitative method, involving selection of important parameters based on professional judgment and experience as well as the allocation of weights to each parameter. Expert judgment is used to compensate for uncertainties in the analytical model. A stochastic grey system model is finally presented for reflecting such uncertainties in the output results. Apart from adopting the reliability interval method to facilitate experts’ fuzzy assessment on the importance of variables, grey system theory and Monte Carlo method are used in the weighting formulation at the attribute and component levels. This novel fire safety assessment model is then employed to evaluate the fire safety level of a residential building to demonstrate its applicability. The case study also illustrates that the model is an efficient, cost-effective and simple evaluation tool on the fire safety level of buildings.

    Research areas

  • Stochastic systems, Fire risk assessment, Mathematical models, Fire protection engineering