BP神经网络在洪涝灾损失快速评估中的应用

Fast evaluation of flood and water-logging losses by BP network

在洪涝灾损失主要影响因素及洪涝灾害特点分析的基础上,采用BP神经网络结构,构造了洪涝灾损失计算的人工神经网络.该网络的输入层反映洪涝灾害的雨情、水情及社会经济和财产状况,中间层反映洪涝灾害的类型,输出层用于表达因洪涝灾害造成的经济损失.应用该网络结构建立了典型流域的洪涝灾损失快速评估模型,并对太湖流域洪灾损失进行评估.评估结果的合理性验证了BP神经网应用于洪灾损失快速评估的科学性.

Based on an analysis of the main affecting factors of flood losses and waterlogging damages and the characteristics of the disasters, an artificial neural network is constructed for fast evaluation of flood and waterlogging losses by use of the BP network. The input part of the network represents rainfall, water regime, social economy, and property situation, the middle part indicates the type of disasters, and the output part expresses the economic losses due to flood and waterlogging damages. By use of the network structure, a fast evaluation model is developed and applied to evaluation of the flood and waterlogging losses of the Taihu Lake Basin, and the results show that the network is reasonable and effective.