An improved extreme learning machine model for predicting the mechanical property of AZ80 magnesium alloy

In order to improve the mechanical property prediction accuracy of AZ80 magnesium alloy, sparrow search algorithm (SSA) is optimized by the tent chaotic mapping (TCM) algorithm, yielding TCMSSA, and TCMSSA is employed to improve extreme learning machine (ELM) model for obtaining TCMSSA-ELM model. TCMSSA-ELM model, along with the traditional ELM model, are both utilized to predict the stress of AZ80 magnesium alloy in a temperature range of 523 K to 673 K and a strain rate range of 0.001s^− 1 to 1s^− 1. Comparative analysis is conducted to evaluate the performance differences between the two models. The results indicate that compared with ELM model, the predicted values of TCMSSA-ELM model are closer to the experimental data and the ideal 45° line, proving the higher prediction accuracy of TCMSSA-ELM model. In addition, through algorithmic improvements, the maximum reduction of MAPE is 85.422% and the maximum determination coefficient reaches 0.99956, implying that TCMSSA greatly improves the prediction precision of ELM model. The related results offer a new optimization strategy and model to predict the mechanical property of AZ80 magnesium alloy in each molding stage, and provide a technical reference for applying intelligent algorithms to alloy design.