Study On Heating Pipe Leakage Diagnosis Based On Deep Learning And IPv6 Internet

Leakage of heating pipes often causes huge property losses and environmental effects.In recent years,the popularity of heating in the country,the aging of heating lines and the damage to the pipelines caused by construction made the leakage of heating pipes occurs frequently.So the research of the detection algorithm and the construction of the monitoring platform play a pivotal role in the timely alarm of the leakage problem and the safety of heating system,which is related to the economic loss of the enterprise and the safety of the staff.Aiming at the defects of the accuracy of traditional machine learning algorithms in the process of heating pipeline leakage,a negative pressure wave leak detection method based on deep belief network and sparse auto encoder is proposed.The method extracts characteristic data such as mean,root mean square,skewness,and kurtosis of negative pressure waves,uses the characteristic data of the negative pressure waveform as the input variable of the model.DBN determines the hidden layer of the model by Gibbs sampling,uses the deep learning to train and extract the features,and obtains the heating leakage detection model.The test result compares with the results of BP,SVM and LS-SVM algorithms shows the improvement of detection accuracy by DBN and SAE.Based on the WinCC software,a monitoring platform based on the heating pipeline system was built.The upper computer interface of the heating system was drawn,and the functions of variables and user management,historical data recording and leak detection alarm were realized.Aiming at the inconvenient problem of different workstations,this paper proposes a webpage publishing scheme based on IPv6 network environment.The platform publishes the software monitoring interface to the browser for client access.The test result shows that the software interface on the server side can be stably accessed in the IE browser.A dynamic data processing method based on OPC communication technology is proposed for the dynamic monitoring process in heating system.The method is based on OPC technology.By establishing an OPC server,creating the OPC Group variable in the WinCC configuration,using the SQL Server database as the storage medium,the real-time data is calculated and transmitted to the model established by the SAE algorithm for testing,and the test result is written in the WinCC software.The test result shows that the method can perform real-time dynamic data testing and perform system leakage alarm.