A Study Of Evacuation Strategies For High-rise Building Fires Based On ACP Methodology

Buildings are main places where people live and work. Recently, with the development of social economy and the limitation of space, high-rise buildings have become the mainstream of urban construction in many cities of the world. However, high-rise building has the characteristics of huge architecture, many people, and few of exits. Being the sudden and multiple disasters, fires take heavy threats on the occupants in high-rise building. At present, occupant evacuation is the main measure to prevent the disaster around the world. Planning occupant evacuation effectively and improving evacuation efficiency is an important problem of evacuation management. However, an evacuation system is a complex system with interactive occupant, architecture, and disaster scenario, which has the essential difficult to build model exactly and do real experiments. This dissertation is guided by the ACP approach(Artificial Systems, Computational Experiments, Parallel Execution), which is a new methodology on modeling, analyzing, managing and controlling complex system, taking the study of evacuation strategy as the entry point. Key problems of evacuation strategy including strategy assess, influence factor analysis, and strategy optimization are studies.The main work of the thesis is as follows.Firstly, this dissertation provides a comprehensive review of occupant evacuation on building fires. The research method and status of evacuation and evacuation strategy are reviewed in detail. Based on the analysis of difficulties and deficiencies in existing research, we outline the necessity and feasibility of ACP methodology.Secondly, modeling of high-rise building evacuation systems is studied. Guided by the ACP methodology, the artificial evacuation system is built by combining agents, space grids, and fire numerical analog technologies. Not only solves the problem of evacuation model, but also provides a model for strategy studies later.Thirdly, evacuation strategies are assessed using fire scenarios. Based on the artificial evacuation system, coupled with specific scenarios, different evacuation strategies are quantitatively evaluated by computational experiments. One hand, the study provides a concrete way of strategy assessment based on ACP methodology. The other hand, it verifies the idea that making strategy should base on fire scenarios. This part of the study can guide how to evaluate strategies in actual evacuation. Considered the uncertainties, Monte Carlo imitation is introduced in the computational experiments.Fourthly, human influence is quantitatively studied. An evacuation process involves a variety of certain or uncertain factors in which human factors make an important influence on evacuation process and efficiency. Based on the above artificial evacuation system, fire scenario and high-efficiency strategy, three objective and two subjective human factors are studied by orthogonal or uniformity computational experiments method. The objective human factors are crowd density, age composition, and gender ratio. The subjective human factors are pre-evacuation time and occupant’s character. Obtained data by studies can be used as an important reference for reasonable distribution of occupants in actual high-rise buildings.Fifthly, an innovative management strategy based on group evacuation is proposed. In order to implement strategy conveniently and optimize route easily, the shortest route is designed by the graph theory for evacuation groups. Different sizes of evacuation group are studied by computational experiment and assessed by evacuation efficiency. The size of optimization group is observed for guiding actual evacuation.Sixthly, strategy optimization based on parallel execution is proposed. For evacuation is a complex dynamic process, optimization of evacuation strategy is a very difficult problem. Based on the connotation and function of a parallel system, a framework of parallel evacuation system, which is designed by considering actual requirements of evacuation, is given. Then, a parallel and interactive mechanism which is drive by data is constructed on the parallel evacuation system. The optimization of evacuation strategy is carried out by the parallel, dynamic, and real feedback mechanism.Related issues of evacuation strategy for high-rise building fires are studied based ACP methodology. The research results will help people to better understand emergency evacuation processes in high-rise building fires, guide evacuation processes effectively, and improve rational understanding on human influences. And it affords a new foundation for emergency evacuation management.