Dynamic data driven application system can improve simulation performance and accuracy by collecting and incorporating dynamic data from fire area. Based on BP artificial neural network, a frame construction of forest fire model selection of suitability was designed. Forest fire model selection knowledge was produced through BP artificial neural network. The system implemented automatic and intelligent selection of forest fire models. BP artificial neural network model of forest fire model selection was build by treating forest fire environment data as input variables and treating appropriate forest fire model as output variables. Additionally, we studied the methods acquiring and calculating data of input and output. The system implemented a mechanism of automatic model selection driven by dynamic data technology. We selected 72 items experimental data from historical forest fire records in Beijing to test and confirm the validity of model selection. It was found that the reliability of model selection was more than 80％.%基于BP人工神经网络方法设计林火模型适宜性选择技术框架结构,通过神经网络形成林火模型选择知识,实现林火模型的自动化和智能化选择;以火场环境因子为输入变量,以适宜火场环境模拟的林火蔓延模型作为输出变量,构建林火模型选择神经网络模型;研究输入、输出因子数据的获取与计算方式,实现动态数据驱动的林火模型自动选择机制.以北京市为例,选择有详细火场情况记录的72场林火作为试验样本,其中60条记录作为学习样本集,12条记录作为验证样本,对神经网络进行学习和验证,结果表明:模型选择精度可达到80%以上.
展开▼
