Research On Safety Operation And Maintenance System Dual-Prevention Of Utility Tunnel Driven By Data

With the deepening of China’s urbanization and modernization,urban underground pipeline accidents have occurred from time to time,causing serious harm to the national economy and people’s lives.As a modern and highly intensive municipal infrastructure,the utility tunnel can effectively solve a series of social problems,such as "road zipper" caused by the direct laying of traditional municipal pipelines.However,the utility tunnel involves a wide area,complicated maintenance and management,the existing experience and supporting policy system are not perfect,and there are hidden dangers such as fire and man-made disasters,which pose great challenges to the safety management of the utility tunnel.The utility tunnel accidents did not happen suddenly.Before it happened,it would release a large number of kinds of signs and information from various sources.Therefore,this paper combines the utility tunnel project in Mazhai District of Zhengzhou City with safety management and deep learning related methods to analyse,preprocess and integrate information on the signs of accidents in the utility tunnel to achieve utility tunnel risk prediction and try to avoid pipeline accidents and provide security guidance for operation manager units.The main research contents include:Firstly,establish a double prevention system for utility tunnel based on multi-source data.Through the analysis of the risk characteristics and the characteristics from multi-source monitoring data of utility tunnel,this paper used RBS and Delphi method to analyze the factors that affect the risk status of the utility tunnel from five aspects: utility tunnel main structure and ancillary structure,utility tunnel pipeline,ancillary facilities,utility tunnel environment and management safety,and finally determine 48 index factors.An AHP-based utility tunnel operation and maintenance risk evaluation model was established,and the utility tunnel in Mazhai District,Zhengzhou City was taken as an example.The calculated risk value was 1.08,which belongs to a slight risk state,which is consistent with the facts and verified This paper proposes the reliability of the risk assessment method;At the same time,according to the characteristics of the operation and management of the utility tunnel,the process of hidden dangers investigation and management of the utility tunnel is optimized,and the automatic investigation of hidden dangers of the utility tunnel based on the Apriori model is realized,which improves the efficiency of the hidden danger investigation and management of the utility tunnel.Secondly,establish a utility tunnel operation and maintenance risk prediction model based on long and short-term memory cycle neural network(LSTM).Aiming at the problem of the prediction accuracy of the utility tunnel operation and maintenance risk evaluation,this paper fully considers the internal relationship between the evaluation index and the utility tunnel risk level,and establishes the LSTM prediction model to learn the relationship between each index and the utility tunnel risk level.The model is a five-layer neural network with an average prediction accuracy of 97.55%.It can solve the problem of gradient disappearance and gradient explosion caused by the long-term dependence problem of the utility tunnel operation and maintenance data,and realize the prediction of the safe operation and maintenance risk of the utility tunnel.Finally,a dual-prevention safety operation and maintenance system based on data-driven utility tunnel is established.This article is based on the B / S architecture,based on demand analysis,system structure and database design,using Pycharm + Python + Django + My SQL as a development tool to develop a dual prevention safety management system suitable for i utility tunnel operation and maintenance.The system implements the entire safety operation and maintenance process of the utility tunnel.At the same time,it uses the collected massive floating operation and maintenance data for the verification of the risk prediction model and the intelligent inspection and governance model of hidden dangers.Furthermore,it plays a role in assisting decision-making for modules such as emergency management of utility tunnel.Through the development and deployment of the system,preliminary application shows that the system can effectively improve the safety operation and maintenance management level of utility tunnel.