Research On Microwave Nondestructive Testing And Imaging Recognition Approach For Corrosion Cracks In Metal Pipelines

The harsh working environment causes the surface of metal pipes(such as petroleum pipelines,natural gas pipelines,etc.)to be prone to various corrosion cracks,which will greatly reduce the residual strength of the pipeline and seriously affect the health of the pipeline.At the same time,for the requirements of pipeline inspection in severe working environments(such as high temperature,heavy dust,etc.),conventional non-destructive testing methods lacks of good anti-interference ability,while microwave non-destructive testing is less affected under such conditions.Therefore,the microwave non-destructive testing technology can be applied in in the detection of pipeline cracks.Furthermore,with the development of microwave detection imaging technology,the microwave detection result has better readability than the traditional detection.Therefore,how to obtain the information such as the position,size,area and boundary of the crack through microwave imaging information is the key issue.Therefore,onto figure out the problems existing in pipeline inspection,the research work on pipeline microwave non-destructive inspection imaging,crack boundary identification methods and crack classification methods are carried out as follows:(1)The quantitative relationship between corrosion cracks and microwave reflection coefficient are studied by numerical simulation.In this thesis,the simulation model of pipeline time-domain microwave nondestructive testing is established by CST simulation software.The finite integral method is used to calculate the model in the time domain.The influence of lift off and working frequency on the phase of reflection coefficient is studied.The selection rule of the optimal working frequency is determined and the relationship between the optimal frequency and the lift-off when the rectangular waveguide probe is used.At the same time,the width of different types of corrosion cracks is studied under the optimal frequency and lift-off.For the penetrating corrosion cracks and pitting corrosion cracks,the quantitative relationship between depth and reflectance phase are presented,respectively.(2)Pipeline microwave non-destructive testing imaging system is constructed and some experiments are carried out.Firstly,based on the research of simulation results,the microwave non-destructive detection compensation algorithm for surface with the complex shape is proposed,which eliminates the influence of the lift-off on the phase of the reflection coefficient in the surface detection.Then,the microwave non-destructive inspection imaging system is built.On the experimental platform,the microwave nondestructive testing experiments on different types of cracks were carried out.First to study the quantitative relationship between the depth of the corrosion defect and the phase of the reflection coefficient,secondly the boundary position and reflection coefficient gradient value of the through defect and the circular pitting defect were studied.The simulation result is verified by the experimental study.(3)The boundary recognition algorithm and classification algorithm for microwave non-destructive testing imaging are investigated.Based on the traditional Canny edge detection operator,an improved Canny edge detection algorithm is proposed.In the improved Canny edge detection algorithm,an edge recognition algorithm for pipeline microwave nondestructive detection with automatic selection of double threshold and variable Gaussian filter are designed.Then,the improved algorithm is applied to identify the defect boundary according to the experimental data of the different type cracks.Finally,the convolutional neural network,is adopted to classify the types of cracks.Finally,the capability of the proposed method is verified by a case study.