A Leakage Detection Method Of Petroleum Tank Based On Infrared Imaging

The preliminary phase of petroleum tanks leakage is the critical time to avoid the expansion of leakage,to prevent explosion accidents,and also to conduct emergency rescue.The traditional manual detection and combustibles concentration detection tank leak detection methods can only be effective when the leak lasts for a period of time and the combustible concentration reaches a certain degree,which the timeliness is poor.A detection method of typical leakage type of petroleum tank based on infrared imaging is proposed to realize the early leakage defect location and leakage type identification.The method is verified by simulation analysis and simulation experiment to be correct and effective.The research work and achievements are as follows:1^st.Based on the principle of infrared imaging and the basic theory of heat transfer,the heat transfer process of petroleum tanks under normal operating conditions is analyzed.The finite element analysis method is used to solve the differential equation of thermal conductivity,and the temperature field theoretical model of petroleum tanks under normal operating conditions is established.2^nd.Considering the difficulty of obtaining the geometric parameter information of the leakage defect of the storage tank,the defect model of the storage tank under the Cartesian coordinate system is established.Based on the principle of equivalent thermal resistance homogenization,the correction method of the heat transfer coefficient is proposed,and the expression of leakage boundary function is derived.Then the theoretical model of the temperature field of the petroleum tank under the leakage condition is established.3^rd.The simulation environment is set up to study the time series temperature field simulation model of petroleum tank under two different typical leakage types of point leakage and slot leakage,and analyze the general regularity of the temperature field distribution of storage tanks under leak conditions.Based on the Tensor Flow deep learning framework and the Matlab platform,the detection method for the typical leakage types of petroleum tanks is proposed by using the method of Alex Net migration learning.4^th.Based on the theoretical model and simulation results of the temperature field of the petroleum tank,865 infrared images under two typical leak types of point leak and slot leak were collected,and simulation experiments were carried out.The results indicate that the two methods can achieve the detection of tank leak defects,with accuracy rates of 94 percent and92.73 percent,respectively,verifying the correctness of the theoretical model of the tank temperature field under leak conditions and the feasibility of the detection method for typical tank leak defect types.The research work completed and the results obtained provide guidance for discovering tank leaks in time and taking effective measures within a critical period of time.The proposed infrared image identification method for tank leakage types can provide reference for engineering practice.