| Carbon Fiber Reinforced Polymer(CFRP)has gradually replaced the traditional method of repairing and strengthening steel structures and has been widely used in many fields due to its good strength-to-density ratio,corrosion resistance,fatigue resistance,and ease of field implementation.However,CFRP strengthened steel structures may produce special defects such as overlapped defects and near neighbor defects during retrofitting or the following service process.The existence of these defects will reduce structural performance and threaten structural safety.Therefore,choosing a convenient and effective Nondestructive Testing(NDT)for defect detection of CFRP strengthened steel structures is of great significance to ensure structural safety.Eddy current pulsed thermography(ECPT)has gained widely attention in research and engineering applications due to its excellent detection performance.This dissertation aims to investigate the feasibility of inspecting this hybrid structures using ECPT and form an ECPT detection system.Specifically,it combines numerical simulation,experimental research,automatic defect recognition with classification and enhancement of defect characterization.The main research work of this dissertation is as follows:(1)Numerical simulation research on ECPT for special defects in CFRP strengthened steel structuresTaking the CFRP strengthened steel structures with special defects as the object,ANSYS finite element simulation software is used to establish a multi-physics coupling finite element model to simulate the detection process.The model is used to simulate the ECPT inspection process of CFRP strengthened steel structures.By analyzing the distribution of electromagnetic field and temperature field in the inspection process under different conditions,the influence mechanism of different types of defects on electromagnetic field and temperature field can be clarified.In addition,the calculation results are compared with the finite element simulation results of ECPT testing for CFRP strengthened steel structures containing only a single type of defect to explore the mutual influence of multiple types of defects at the same time.The finite element simulation study verifies the feasibility of ECPT to detect special defects in CFRP strengthened steel structures.(2)Experimental research on ECPT for special defects in CFRP strengthened steel structuresPrefabricated CFRP strengthened steel specimens containing debonding,delamination,crack,overlapped defects and near neighbor defects,and conduct ECPT inspection experiments.Then the temperature thermal image and temperature data obtained in the experiment are analyzed,and the test results of the test pieces containing single types of defects are compared with the test results of test pieces containing special defects.Through experimental research,the validity of the simulation results is verified and the feasibility of ECPT to detect special defects in CFRP strengthened steel structures is further proved.The applicability of ECPT to detect special defects in CFRP strengthened steel structures is further clarified through experimental exploration.(3)Research on intelligent detection method on ECPT for special defects in CFRP strengthened steel structuresBased on YOLOv3,the YOLOv3-ECPT deep learning algorithm for automatic recognition of CFRP strengthened steel structure defects is established.Infrared thermal imaging images containing only a single type of defects are used as training samples to realize the identification and classification of different type defects in structures containing special defects.In addition,in order to enhance defect characterization,this dissertation proposes a method for processing ECPT detection data based on Convolutional Auto-Encode—CAE-ECPT.Data processing is performed on the acquired infrared thermal image sequence to obtain images with more obvious defect characteristics.This dissertation carried out the finite element simulation and experimental research on ECPT inspection of CFRP strengthened steel structure containing special defects to explore the detection mechanism,which provides a theoretical basis for engineering practice.In addition,this dissertation proposes a Python-based YOLOv3-ECPT automatic defect recognition method and a CAE-ECPT image enhancement method,which provides a basis for the establishment of an intelligent system for ECPT to automatically recognize and detect defects in CFRP strengthened steel structures.In a word,the research content of this dissertation is of great significance to guarantee the safety performance of CFRP strengthened steel structure during service and improve the intelligent level of ECPT inspection for structures. |
