Study On HDPE Membrane Chromatography Signal And Leakage Edge Recognition System Under Landfill Condition

With the improvement of people’s living standards,the amount of domestic waste generated continues to reach a new high,and sanitary landfill has become one of the main methods of domestic waste disposal.Affected by many factors,the landfill will have different degrees of leakage during operation.The landfill leachate entering the surrounding environment of the landfill will cause great damage to the water and soil,and even endanger human safety in severe cases.The realization of intelligent welding repair at the leakage of buried HDPE membrane has become one of the difficulties faced by researchers.At present,the research on the problem of landfill mainly focuses on the leakage location.However,due to the coating of multilayer media on the landfill membrane in operation,the HDPE membrane is invisible,and the leakage area cannot be identified,which causes technical obstacles to the subsequent welding repair.Therefore,this thesis studies the leakage edge identification of HDPE membrane under landfill conditions.Firstly,the longitudinal positioning of HDPE membrane under landfill conditions is studied by using the signal distortion characteristics,the traditional long and short time window method is improved,a four-time window method model is proposed,and the vibration signal of each structural layer of the landfill is extracted by combining seismic while drilling technology;According to the different vibration characteristics of the signal excited by the collision between the rigid drill bit and different media,the calibration signal theory is introduced,and the cross-correlation processing and complex wavelet domain transformation analysis of the calibration signal and the working signal are proposed.After the feasibility verification,when the main frequency of the vibration signal is concentrated in 1000 ～ 2000 Hz and the cross-correlation value is greater than 0.8,the longitudinal positioning of the target layer can be realized.Then,the leakage edge recognition theory is studied.After the medium on the membrane in the leakage area is removed by drilling technology,the image information of the leakage area is collected by the image acquisition device,the data set is constructed,and the image processing model is established for preprocessing;Extract the image target information features,build a classification and recognition system based on binary tree theory and support vector machine,improve the particle swarm optimization algorithm,and propose an adaptive particle swarm optimization algorithm to optimize the parameters of the recognition system to optimize the system performance.Finally,combined with the theoretical research content of machine vision technology,the software and hardware parts of a series of related modules such as leakage information acquisition module,image processing module and leakage edge recognition module are designed.The experimental results show that the recall rate of surface leakage information of the original image is high after image processing,which shows that the image processing model can effectively retain the leakage information and meet the application requirements.When the parameter combination is penalty factor C = 8 and kernel parameter g = 4.74,the model is optimal.The overall recognition accuracy of the optimized classification model is 98.89%,including the recognition accuracy of complete image is 99.58%,block leakage is 98.33%,and slit leakage is 98.75%,which is significantly higher than that before optimization.