Research On The Health Assessment Model Of Sewer System Based On BP Neural Network

Current urban diseases such as water environment pollution,black and smelly water body and urban waterlogging are closely related to the healthy operation of the drainage pipeline system.As an extremely important part of the municipal infrastructure system,drainage pipeline system is directly related to the toughness of the city,public safety and ecological environment protection.In this study,AHP and expert group decision-making were used to construct the health status evaluation system and grading standard of urban drainage system,and the BP neural network model optimized by genetic algorithm(GA)was used to evaluate and demonstrate.Solve the problems that existing method of the single evaluation model,the subjectivity,through neural network powerful ability of self learning and adaptive,a vast amount of historical data model of urban pipeline system and effective use of deep mining,both as a city water management department to provide scientific,effective and feasible decision-making basis,is to promote urban renewal,Guarantee the high quality and sustainable and healthy development of the city to lay the essential premise and foundation.The research results and contents are as follows:(1)This study urban rain,sewage drainage system as the research object,based on the domestic and foreign research present situation and the actual project survey data,build include environmental factors respectively,drainage ability,adaptation and repair ability three aspects of health assessment index system,and index system by using analytic hierarchy process(AHP),the empowerment of sorting and filtering is simple.Single factor analysis was used to analyze the correlation of 14 indexes,such as surface type,pipe age,pipe diameter,pipe material and construction quality,and then the scoring standards of health degree corresponding to the two evaluation systems were obtained respectively.(2)The GA-BP neural network was used to construct the health status assessment model of rainwater and sewage pipeline system by MATLAB software.The topological structure of the constructed rainwater system assessment model was 14-5-4,and the model accuracy was91.58%.The topological structure of the sewage system evaluation model was 14-8-4,and the model accuracy was 89.25%.For the four output indexes of GA-BP neural network model,the trapezoidal distribution membership function is used to carry out fuzzy comprehensive evaluation,and then the comprehensive health status evaluation of drainage pipeline system is realized.The evaluation results of health degree were mainly divided into four grades: Ⅰ(healthy),normal(sub-healthy),normal(unhealthy)and normal(sick).(3)The health status assessment model constructed in this study was validated on the 300 sections of rainwater and sewage drainage pipeline system in the research area,and the results showed that: The prediction accuracy of GA-BP rainwater pipeline evaluation model was92.33%,and the health degree of gradeⅠ,gradeⅡ,gradeⅢand gradeⅣpipelines in the region were30.30%,62.03%,6.00% and 1.67%,respectively.The prediction accuracy of the sewage pipeline evaluation model was 91.63%,and the proportion of the gradeⅠ,gradeⅡ,gradeⅢand gradeⅣpipelines in the region were 28.28%,62.39%,7.00% and 2.33%,respectively.By comparing the emergency repair record and maintenance plan of the pipeline in this area,it is proved that the verification results are consistent with the reality and have good feasibility and applicability.