Abstract
An artificial intelligence evaluation system using neural network was developed to upgrade the creep and creep-fatigue damage assessment methodology for heat resistant steels of fossil power plants through image analyses of EBSD(Electron BackScateer Diffraction pattern) maps. KAM(Kernel Average Misorientation) maps were obtained for creep and creep-fatigue damaged austenitic stainless steel SUS304HTB and the stratified data were manipulated to evaluate damage degree. The system consisted of an input layer, intermediate layers and an output layer. As the activation function, ReLU(Rectified Linear Unit) function was used for the intermediate layers and Softmax function was used for the output layer. The evaluation results of the proposed system were compared with the results of the conventional quantitative damage evaluation method(the master curve method). As a result, the estimated damage accuracy of the artificial intelligence evaluation system developed in this research was proved to make some improvement compared with the estimated damage accuracy using the conventional evaluation method. The introduction of the neural network is considered to be effective for evaluating even for insufficient number of experimental data. Thus machine learning methodology utilizing neural network is proved to have the potential of versatile data analysis method applicable to various sorts of metallographic investigation.
