Study On The Decision Making Of High-rise Buildings Fire Emergency Based On For On Dynamic Bayesian Network

The land price in cities is expensive,the high-rise buildings can increase the utilization rate of land per unit area and meet the housing needs of people.However,the high-rise buildings structures are complex,and there are many internal combustibles,which are more likely to form fire problems such as three-dimensional fires and evacuation of personnel.High-rise buildings fires cause a large number of casualties and economic losses to society every year.Therefore,it is imperative to study fire emergency decision-making for high-rise buildings.The fires in high-rise buildings are sudden and the fire spreads rapidly,so how to make emergency decisions followed quickly is the focus of scholars' current research.On the one hand,the evolution path and law of high-rise building fire scenarios were analyzed,then the dynamic Bayesian network is used to predict the scenario of high-rise building fires,and calculate the most likely scenario of the fire at the next moment.On the other hand,the similarity retrieval method is used to match the current fire target scenario with the similar scenario,finding the solution ideas and solutions of similar scenarios and providing ideas and methods for the fire workers to make emergency decisions.Firstly,aiming at the characteristics of high-rise building fire accidents,information acquisition time and data dynamics,the Grounding theory is used to determine the elements of high-rise building fire accident scenarios.This dissertation analyzes and complies four primary constituent elements,ten second-level constituent elements and 36 situational information components,which provided a basis for the next step of collecting fire accident site information and determining the weight of the scenario elements.Aiming at the diversity characteristics of high-rise building fire evolution path,Bayesian network integrating graph theory knowledge and probability theory knowledge can effectively solve this kind of uncertainty problem.Therefore,the dynamic Bayesian network was used to construct a dynamic evolution network model of high-rise building fire scenarios,and GeNIe software was used to calculatethe most likely evolution path of the fire at the next moment.Secondly,AHP is used to determine the weight of each element.Through the case retrieval,the scenarios similar to the target scenarios were matched from the scenario library to obtain solutions for similar scenarios.According to the actual situation of the current fire,the similar scenarios' solution are corrected and learned,and an emergency decision-making scheme suitable for the current fire is formulated.Finally,the model was validated,using a scenario library containing 20 high-rise buildings fire accident scenarios.The similar scenarios of the target scenarios are found through the analysis and calculation,then the similar scenarios are corrected and the solution suitable for the current fire accidents is formulated,proving the feasibility of the decision-making model.This study not only enriches the research theory of emergency decision-making for high-rise building fire accidents,but also provides a reference for emergency decision-makers to make quick and accurate fire emergency response strategies,and has important practical significance for effectively reducing fire losses.