Fatigue Damage Assessment Of Q235 Based On Data Fusion Of Magnetic-thermal Characteristic Data

In the development of modern industry,the metal structure as an important part of production activities,its working conditions are increasingly demanding,for the metal components of the wear resistance,heat resistance,tensile strength,strength,toughness and other properties are more and more harsh.Especially in the working parts under alternating load or under the effect of alternating hot and cold fatigue is prone to fatigue,with the passage of time,structural parts in the stress concentration area and defective areas of failure or even serious consequences,in order to solve these problems,A variety of non-destructive testing technology is introduced into the damage assessment of metal structures.In order to accurately evaluate the status of metal structures,the use of magnetic and thermal detection means for the metal structure of fatigue damage identification and measurement to provide important methods and technical means.In the manufacturing industry,for the metal structures that have served one or more life cycles,the damage is often the result of a combination of multiple induction factors and structural functional characteristics in a complex service environment.It is difficult to rely solely on a single detection method To achieve the precise positioning and quantification of damage,this paper uses multi-feature to compensate for the existence of a single feature of the lack of metal components for the assessment of the degree of fatigue to provide technical support.In order to clarify the correlation between magnetic memory signal and infrared thermal image characteristics during the fatigue process of ferromagnetic components.the quantitative assessment of fatigue damage was carried out.In this paper,the central hole Q235 steel specimens were selected for axial tensile fatigue test,the establishment of thermal infrared detection and magnetic flux leakage detection system,the samples were collected in different thermal cycles under the infrared image and magnetic field under the magnetic field signal.Through the magnetic memory signal and thermal imaging characterization of the Q235 steel specimen during fatigue,the variation rules of the magnetic-thermal characteristics were studied,and the energy ratio,temperature average,correlation coefficient and peak gradient of magnetic memory signals were extracted to extract Four characteristics as data layer to construct the non-linear mapping of damage magnetic and thermal characteristics,and based on the training function LM BP neural network algorithm to establish a fatigue damage magnetic-thermal fusion evaluation model for network training and analysis of error results.Taking the Q235 steel as the verificationsample,the low-frequency energy ratio,the temperature average,the correlation coefficient and the peak-to-peak ratio of the magnetic memory signal are used as the evaluation indicators.The experimental results are in good agreement with the predicted results,which shows that the model is feasible.In the practical application,the fatigue life prediction of Q235 support rod is taken as an example.The fatigue life(cycle)of the metal structure under working condition is obtained as(777.83 1776.44 2689.65 3765.664881.42 units: times).In this paper,the experimental platform of magnetic-thermal fusion is completed.The wavelet transform and binary data analysis method are used as multi-scale analysis tools to extract the magnetic and thermal characteristics.The data fusion is carried out based on BP neural network algorithm.The quantitative assessment of fatigue damage can be achieved.Through the above conclusions,the quantitative assessment of metal fatigue damage can be provided.However,because the BP neural network algorithm is still in the basic stage,it is necessary to optimize the algorithm in the later stage.Locate and quantify the degree of damage to metal fatigue.