Analysis and Prediction of Hardness Change at High Temperature Based on BP Neural Network

With the development of science and technology, more and more equipment are facing the problem of service under extreme temperature. With the increase of temperature, the properties of materials often change, which leads to the failure state of materials at high temperature. However, the existing material testing methods, such as hardness, are mainly based on the traditional normal temperature testing. Although high-temperature hardness testing has been developed in recent years, the common upper limit temperature of the incubator is about 1200 °C, which is far lower than the denaturation temperature of most materials. Therefore, this paper analyzes the results of the change process of the hardness value measured at high temperature with temperature, and puts forward a high-temperature hardness prediction concept based on BP neural network. Based on the measured high-temperature hardness data, it predicts the subsequent change of the hardness value with the continuous increase of temperature, and looks for the temperature point of hardness failure. Through iterative learning, the fitting degree of the calculation result of the hardness change prediction model is close to one, which can well predict the result.