您的位置:首页  > 论文页面

基于BP神经网络的城市区域火灾过火面积预测研究

发表时间:2016-04-30  浏览量:2069  下载量:858
全部作者: 周德闯,丁超,陈明毅
作者单位: 中国科学技术大学火灾科学国家重点实验室
摘 要: 选取某城市区域为研究对象,使用基于元胞自动机的火蔓延概率模型计算了不同起火点、不同环境风条件下的火蔓延过程。以元胞自动机的模拟结果为依据,对计算数据进行处理构建训练样本,并借助Matlab神经网络工具建立了城市区域火灾过火面积的BP人工神经网络(artificial neural network,ANN)预测模型。研究结果表明,该模型能快速预测城市不同区域火灾在不同环境风条件下过火面积的动态变化,可为城市规划设计和消防扑救提供有益的参考。
关 键 词: 安全工程;过火面积;神经网络;城市区域;火蔓延
Title: Burned area prediction of urban district fire based on BP artificial neural network
Author: ZHOU Dechuang, DING Chao, CHEN Mingyi
Organization: State Key Laboratory of Fire Science, University of Science and Technology of China
Abstract: The fire spreading processes under different parameters in a urban district are computed by fire spreading probability model of cellular automata. The investigated parameters include the location of ire source and environmental wind. The BP artificial neural network (ANN) model of burned area is built by Matlab neural network tool on the basis of cellular automata simulation results. The model can simulate the changing process of burned area under different environmental winds, providing the certain reference for city planning and fire control.
Key words: safety engineering; burned area; neural network; urban district; fire spread
发表期数: 2016年4月第8期
引用格式: 周德闯,丁超,陈明毅. 基于BP神经网络的城市区域火灾过火面积预测研究[J]. 中国科技论文在线精品论文,2016,9(8):856-862.
 
0 评论数 0
暂无评论
友情链接