摘要
老旧社区是城市承灾系统的关键短板,遭受雨洪灾害冲击时老旧社区的防灾能力最为薄弱,提高老旧社区雨洪韧性对于增强城市防涝应急管理能力至关重要。基于PSR(压力-状态-响应)模型构建老旧社区雨洪韧性评估指标体系,采用BP神经网络(back propagation neural network,BPNN)建立了老旧社区雨洪韧性评估模型(PSR-BPNN),并进行模型检验和训练测试,最后以江苏省镇江市京口区老旧社区为例进行实证研究。结果表明:经学习训练,评估模型拟合优度R 2接近1,误差(Mu)=1×10^(-7),模型精度良好;20组训练集相对误差均接近于0,5组测试集平均误差为0.0935,相对误差范围为0.0086~0.2865,模型具有较强的适应性;案例社区的雨洪韧性综合得分L^(*)=0.548,在I~V级雨洪韧性等级划分中处于第Ⅱ级,社区雨洪防灾韧性较差,社区承受了较强的外部压力(P)、状态层存在诸多薄弱环节(S)、响应层为弱响应(R)。PSR-BPNN模型有助于揭示老旧社区雨洪等外部灾害韧性形成机理,研究结果对于削减社区外部压力、补齐社区设施短板和增强社区灾害响应能力具有重要参考价值。
Old communities are the key shortcomings of the urban disaster-bearing system,and their disaster prevention capacity is the weakest when they are hit by rainstorm flood disasters,so improving the stormwater resilience of old communities is crucial to enhancing the emergency management capacity of urban flood prevention.Based on the PSR(pressure-state-response)model to construct an old community stormwater resilience assessment index system,a BP neural network(back propagation neural network,BPNN)was used to establish an old community stormwater resilience assessment model(PSR-BPNN),and model testing and training tests were conducted,and finally,an empirical study was conducted with the old community in Jingkou district,Zhenjiang City,Jiangsu Province as an example.The results show as follows.after learning and training,the evaluation model fit goodness of fit R 2 is close to 1,error(Mu)=1×10^(-7),the model accuracy is good.the relative error of 20 training sets are close to 0,the average error of 5 test sets is 0.0935,the relative error range is 0.0086 to 0.2865,the model has strong adaptability.The case community has an overall stormwater resilience score of L^(*)=0.548,which is in level II in the I-V stormwater resilience classification.The community has poor stormwater resilience,and the community is subjected to strong external pressure(P),many weak links in the state layer(S),and weak response(R)in the response layer.The PSR-BPNN model helps to reveal the formation mechanism of external disaster resilience such as stormwater in old communities,and the research findings have important reference values for reducing external pressure on communities,making up for the shortcomings of community facilities and enhancing community disaster response capacity.
作者
尚磊
张友志
张董寅
钟媛
SHANG Lei;ZHANG You-zhi;ZHANG Dong-yin;ZHONG Yuan(School of Civil Engineering and Architecture,Jiangsu University of Science and Technology,Zhenjiang 212100,China;Jiangsu Kexing Engineering Construction Superivsion Co.,Nanjing 210021,China;Jilin University of Architecture Engineering Laboratory of Building Energy Efficiency Technology,Changchun 130118,China)
出处
《科学技术与工程》
北大核心
2023年第26期11338-11348,共11页
Science Technology and Engineering
基金
教育部人文基金(17YJAZH123)
江苏省社会科学基金(2019XZB016)
江苏省研究生科研与实践创新计划(SJCX22_1977)
镇江市重点研发计划(SH2019024)。
关键词
老旧社区
雨洪灾害
PSR模型
社区韧性
BP神经网络
定量评估
old community
rainstorm flood disasters
pressure-state-response(PSR)model
community resilience
BP neural network
quantitative assessment