Design and Implementation of Sintered NdFeB Performance Prediction System Based on Machine Learning

In order to improve the intelligent level of sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}$</tex> Fe <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">${}_{14}\mathrm{B}$</tex> process, realize the quality control forward and process optimization of sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}\text{Fe}_{14}\mathrm{B}$</tex> process, this paper constructs a machine learning sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}\text{Fe}_{14}\mathrm{B}$</tex> material performance prediction system. Based on the adaptive weighting method of gradient descent, the system integrates ElasticN et model and XGBoost model to construct a combined prediction model. The results show that the RMSE of the combined prediction model for <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{B}_{\mathrm{r}}$</tex> is 0.031, and the coefficient of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> is 0.628; the RMSE of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{H}_{\text{cj}}$</tex> is 0.0.36, and the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> score is 0.897; the RMSE of <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{BH}_{\max}$</tex> is 0.036, and the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathrm{R}^{2}$</tex> score is 0.851, which realize the accurate prediction of sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}\text{Fe}_{14}\mathrm{B}$</tex> magnet performance. At the same time, the front-end interface of the system is friendly and intuitive, even users without deep technical background can quickly get started. With the help of the system, users can easily adjust the input variables, view the expected performance of sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}\text{Fe}_{14}\mathrm{B}$</tex> materials under different conditions, and then make more intelligent production decisions, which has great potential application value for improving the quality and efficiency of sintered <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\text{Nd}_{2}\text{Fe}_{14}\mathrm{B}$</tex>.