Study On Defect Factors Analysis And Prediction Model Of Drainage Network

Drainage pipe network is an important part of urban infrastructure,and it is also an important vanguard of urban flood control and sewage.With the continuous improvement of the level of urbanization,there are numerous safety accidents caused by the deterioration of pipelines,and the problem of urban drainage has become increasingly prominent.In recent years,the drainage pipe network system has been continuously expanded,but how to manage the drainage pipe network system is also a key issue.Nowadays,one of the main management tasks of urban underground pipeline network is to conduct regular defect detection of drainage pipelines in order to grasp the current situation of pipelines.Local detection results in a lot of waste of resources and hinders the development of urban intelligence.Therefore,being able to identify the key factors of the deterioration of the drainage network and establish a prediction model to predict the deterioration of the drainage pipeline is called an effective measure for efficient management of drainage management.In this paper,a binary logistic regression model and a multivariate logistic regression model are established to identify the key structural factors of drainage pipe defects and carry out quantitative analysis.Based on Python,the Markov Monte Carlo method and Bayesian inference methods are combined to construct a The deterioration prediction model of the drainage pipe network realizes the identification and prediction of deterioration factors of the drainage pipe network,and provides the initial state of pipeline defect classification detection.This paper mainly studies the following three parts:(1)Based on the detection results of drainage pipeline defects,extensive literature review and expert experience,the main factors for the deterioration of drainage pipelines of different types of pipeline drainage are classified and sorted;according to the specification standards,the drainage pipeline defects are reorganized into seven categories of defects,and they are classified according to the defect level.Assess the severity of pipeline defects.Finally,the analysis of the influence law of the defects of the drainage pipe network is carried out,which can provide a reference for the managers and maintenance personnel of the urban municipal drainage pipe network.By establishing a logistic regression model,the quantitative analysis of the causes of the defects of the drainage pipe network is completed.First,a binary logistic regression model is established to identify the significant influencing factors of the drainage pipe defects.Provide reference for optimization of drainage pipe network design.(2)By constructing the MCMC pipeline deterioration prediction model,a series of operations such as calling the drainage pipeline deterioration model and running the model are realized.Obtain the Markov estimated transition probability histogram and the deterioration curve of the defect grade of the drainage pipe.Finally,it is shown that the method of building a drainage pipe deterioration prediction model based on the MCMC method constructed by calling the Jupyter Notebook encoder is efficient and feasible.Judging the progress of the deterioration of the pipeline,and providing guidance for the detection and drainage of the drainage network.(3)Taking the drainage pipe network in Longgang District,Shenzhen as an example,relying on the logistic regression model and the MCMC pipeline deterioration prediction model,the key factors of the drainage pipe network deterioration in this area were identified and quantitatively analyzed.The study found that "pipe depth" and "pipe diameter" are significant factors that affect the occurrence of defects in drainage pipes.The Markov transition probability histogram is obtained by sampling through the M-H algorithm,and the deterioration curve of the drainage pipe is obtained by running the code.It is concluded that the model deduces the real deterioration of the basic composite pipeline,which proves that the Bayesian technique and the MCMC method are better prediction models for the deterioration of the drainage network.This study demonstrates the effectiveness of the logistic regression model and the MCMC pipeline deterioration prediction model,and the constructed two models can be well applied to the study of large-scale drainage network.Provide strong technical support for managers and decision makers of related municipal drainage network.The technical method and system platform established in this paper are universal and have the potential to be popularized and used in other regions.