Prediction of magnesium reduction extent and optimization of retort structure based on the machine learning (ML)

The intelligence of magnesium smelting equipment requires a fast and accurate digital model as the basis, and providing a digital model through "multi-field coupled numerical simulation, fast-prediction model, and then intelligent optimization design" is an important method. This article is based on the Discrete Element Method (DEM) to construct a natural and disordered stacking model, that is close to reality, of the briquette layer in the retort of magnesium production. Through single-factor numerical simulation and multi-factor orthogonal simulation design, 1894 valid data points were obtained. Then four machine learning models were trained, tested, and validated on this dataset. Finally, the optimal machine learning model was obtained and used to predict the magnesium reduction rate under the new design conditions. The following conclusions are obtained: (1) In response to the limitations of traditional ordered stacking models, a natural disordered stacking model of briquette layer is established based on the DEM. The free-fall motion of briquettes is simulated to obtain the true porosity and contact characteristics. The bridging treatment is used to solve the problem of grid division, and a coupled model of heat transfer coupled the reduction reaction in the briquette layer is constructed. The natural disordered model results show a reaction time of ~10 hours which is closer to the real filed data. The simulation results also can clearly show the advancing process of reduction on the briquette layer in the retort under high temperature and vacuum conditions. (2) The single factor simulation results on the disordered stacking model show that factors such as briquette size, briquette stacking layer thickness, and retort wall temperature have a significant impact on the magnesium reduction rate of the retort. Keeping other factors constant, increasing briquette size, decreasing briquette stacking layer thickness, or increasing retort wall temperature can significantly reduce reaction time. This article comprehensively considers factors including the simulation results, and the magnesium production for one retort, magnesium production rate, and retort-steel’s service life, and provides the optimal range of values for each factor: briquette diameter of 20 mm-30 mm, briquette stacking layer thickness of 100 mm-140 mm, and small fluctuations around 1473 K for the retort wall temperature. (3) The results of four classic machine learning models for predicting the reduction rate of magnesium for retort show that eXtreme Gradient Boosting (XGBoost) has the highest prediction accuracy (R2=0.997) and the smallest prediction deviation for new design cases. The Random Forest (RF) model has a large prediction error (R2=0.8654) due to the low discrete values of features such as briquette stacking layer thickness in the dataset. The SHapley Additive exPlanations (SHAP) analysis of the XGBoost model shows that the thickness of the briquette stacking layer has the greatest impact on the reduction rate, followed by the temperature of the retort wall and the diameter of the briquette, while the diameter of the inner tube has the smallest impact. The application of XGBoost model can provide a reliable prediction tool for the optimization of magnesium-production retort structure and process parameters, while also compensating for the long prediction time of traditional simulation models.