【摘要】 近年来在全球变暖的气候变化环境下,极端降雨事件的频次和强度不断增加,城市内涝灾害频发,内涝灾害风险评价能够有效减小城市内涝灾害的损失。以汕头市濠江区为例,采用神经网络改进权重的层次分析法,拓展了传统的九分法标度,构建了汕头市濠江区内涝灾害风险评价体系。结果表明,濠江区内涝灾害风险整体呈现中南部高,西北部低的特点;其中玉新街道、滨海街道、马滘街道和达濠街道风险值较高,需要加强防范。利用历史数据验证评估结果,验证结果表明,约80%的历史洪涝灾害点分布在高风险区域,与历史灾害点的验证信息一致。运用神经网络改进了层次分析法的权重确定方式,将传统九分法的中间变量拓展到小数点后三位,并由机器打分确定权重已一定程度上减小主观性,可为相关部门的洪涝灾害的预警发布和防洪指挥调度提供科学指导。更多还原

【Abstract】 Under the background of global climate change, the frequency of extreme rainfall events increases, and the urban water-logging disaster occurs frequently.The risk assessment of water-logging disaster can effectively reduce the loss.Taking Haojiang District of Shantou City as case study, this paper improves the weights of the analytic hierarchy process(AHP) by neural network, expands the traditional nine-point scaling, and constructs the water-logging disaster risk evaluation model.The results show that the water-logging disaster risk in Haojiang District is generally high in the south-central region and low in the northwest region.Especially, Yuxin Street, Binhai Street, Majiao Street and Dahao Street have higher risk values, which need to take preventive measures.After verification of evaluation results with historical data, it is shown that about 80% of historical flood disaster points are distributed in high-risk areas, which is consistent with the verification results of historical disaster points.The weight determination method of AHP is improved by neural network, the intermediate variables of the traditional nine-point method is expanded to three decimal places, and the weight is determined by the computer to reduce subjectivity to a certain extent.The results of risk assessment can provide technical support for flood warning and flood control scheduling within a certain time limit.更多还原