Research On Public Building Fire Risk Assessment Control Model Based On Full Life Cycle

With the speeding up of urbanization process, large space, multi-function, crowded public buildings are increasing. In case of fire, it will cause enormous casualties and property losses, so control the public building fire effectively plays an important role in the fine management of city. Both in the design phase and operation phase, the building fire risk assessment is an effective means of fire management, can optimize the design of the building and raise the capacity to prevent and control the fire. Fire insurance in the future calls for accurate building fire risk assessment.The current methods of the public building fire risk assessment almost without considering different characteristics the building has in different phases, the assessment result was not associated with phases. There is a big workload in assessment process, assessment result is not effective for the fire risk control in different phases. At the same time, the establishment and grading standard of now used index system of fire risk assessment are not clear, result in there are too many human factors in the process of grading. The grading results of different experts are quite different and scattered, reducing the objectivity and accuracy of the final assessment result. In addition, there are a lot of fire risk factors in the public building, various factors are connected with each other, it is also a difficult problem to build accurate fire risk assessment model.To solve above problems, this thesis includes three aspects of work, they are as follows: Firstly, put forward the concept of full life cycle of fire risk assessment, according to different fire risk factors in different phases, the fire risk evaluation is made respectively in the design phase and use phase. Secondly, the objective factors of building fire risk are divided into passive and active factors according to the characteristics. Passive factors mainly refers to the relatively stable factors and are not easy to change. Active factors are easy to change because of the change of equipment and function. In addition, quantitative classification method of index is put forward and implemented by using the simulation method. Thirdly, the training samples and test samples are established in this thesis, support vector machine algorithm is adopted to establish the fire risk assessment model, this method is used for fire risk assessment, get a precise result in the case of less samples.Building has different fire risk in different phases of the full life cycle, the assessment index system and assessment model established in this thesis, can make fire risk evaluation in the different phases objectively and scientifically. The assessment result is certain and convergent, less influenced by human factors, more suitable for building fire risk assessment in the case of small samples. This thesis provides a effective method for public building fire risk assessment, the method has practical value in engineering.