基于神经网络的蓄滞洪区洪灾风险模糊综合评价.pdfVIP

6O 中国农村水利水电·2008年第6期 文章编号：1007—2284(2008)06—0060—05 基于神经网络的蓄滞洪区洪灾风险模糊综合评价 李绍飞，余 萍，孙书洪 (天津农学院水利工程系，天津300384) 摘 要：洪灾风险评价属于多准则、多层次的模糊综合评价问题，在深入分析洪灾风险成因的基础上，以致灾因子、 孕灾环境、承灾体属性为准则层，构建了蓄滞洪区洪灾风险评价指标体系，制定了洪灾风险等级和评价标准。基于BP 神经网络，构建了蓄滞洪区洪灾风险评价的BP模型，将其应用到5O年一遇洪水情况下，大黄铺洼I、Ⅱ、Ⅲ区的洪灾风 险评价中，结果表明，I区洪水危险性高，但人口稀少、经济不发达，洪灾风险程度最低；III区洪灾风险程度稍高于I区， Ⅱ区洪灾风险程度较高。以上结论与实际情况基本一致，验证了BP评价模型的合理性。 关键词：神经网络；蓄滞洪区；洪灾风险；评价；指标体系 中图分类号：TV87 文献标识码：A Fuzzy Risk Assessment of Flood Hazard Based on Artificial Neural Network for Detention Basin LI Shawfei，YU Ping，SUN Shu—hong (Tianjin Agriculture College，Tianjin 300384 China) Abstract：Risk assessment of flood hazard is a problem of multi-principle fuzzy synthetical evaluation．In this paper，based on an anal— ysis of flood hazard factors，a multi—principle index system was suggested，taking disaster—inducing factors，disaster environment gestations，properties of hazard bearing body as principle indices．Evaluation grade and standard of flood hazard were established．A method of assessing flood hazards in detention basin was formulated，based on BP neural networks．Then it was applied to the risk e— valuation in Dahuangpuwa in Haihe River basin，with the flood—control standard of once in fifty years．The results show that flood hazard risk in Subarea I is low，and it iS high in Subarea 1I．The above—mentioned conclutions are accordant with the actual situation， thus verifying the rationality of this method． Key words：artificial neural network；detention basin；flood hazard risk；assessment；multi—principle index system 洪水灾害突发性强、发生频率高，无论受灾人数还是经济 用Arc／Info等地理信息系统软件提供的空间分析功能进行洪灾 损失在各灾种中都占据首位或次首位，我国是洪水灾害最频繁 风险区划与评价，该方法属于经验性的统计方法，评价模型为线 的国家之一，每年洪灾所造成的损失呈上升趋势。洪灾威胁严