Study On Fire Characteristics And Optimization Of The Distribution Of Urban Fire Station Of Urban Area

City construction is changing with each passing day,and the integration of diversified street spaces and old and new buildings has prompted a variety of fire influencing factors in the urban block environment.For the optimization of the layout of the fire station,it needs to be carried out based on the characteristics of the city's fire.The existing analysis methods have subjective judgments and fail to make full use of the city's geographic data,especially the old streets and small areas that are easy to ignore in the urban area.Property rights houses,etc.,cannot fully and accurately reflect the true fire risk distribution,and the work efficiency of fire stations cannot meet actual needs.Therefore,in this paper,by investigating a place with a high fire risk in a certain urban area of Xi'an,and analyzing the fire time data and spatial data,the machine learning algorithm is introduced to comprehensively analyze the spatial correlation of the fire distribution,and on this basis,the city fire station is proposed.Layout optimization method.main tasks as follows:Based on an in-depth investigation of the fire safety situation in a certain urban area of Xi'an,the historical fire data from 2017 to 2019 is analyzed,and the spatial and temporal distribution of fire is analyzed by using wavelet analysis and nuclear density analysis.Fire time distribution law:the fire frequency fluctuates throughout the entire time domain for 4 cycles.The fire frequency shows seasonal fluctuations,and the fire frequency is higher in winter and summer;the fire spatial distribution law:the fire is mainly concentrated on Jianzhang Road and Sanqiao Street The office is in a state of high concentration,extending outwards and showing a general downward trend,but there are still sporadic high fever points.Among them,old communities and high-rise small property houses are fire-prone areas.Random forest and BP neural network are used to model and analyze the occurrence of urban fires.Through the comparison of evaluation indexes such as deterministic coefficient and mean square error,the random forest prediction model is superior to the BP neural network prediction model.The test results show that the model has good predictive ability;through the analysis of the data outside the bag,the importance of each fire characteristic factor is obtained,and the results show that the density of high-rise small property houses and the density of old communities are the two most important fire characteristic factors.The fire risk point distribution map is established by the distribution of the fire spatial nuclear density and the distribution of the nuclear density of the fire consequence variables,combined with the spatial correlation of the fire distribution,using the positioning rationing model,the layout of the urban fire station is optimized in stages,and is considered in the optimization of the layout of the fire station The corresponding relationship between the risk level and the response time,that is,the risk corresponds to the corresponding level of fire rescue measures,the number and location of the optimized first-level general fire station,small fire station and micro fire station are determined respectively.The research results of this paper can provide decision support for urban area fire prevention and control and fire station layout optimization,and have important social value.