Leaching rate prediction for multimodal hydrometallurgical leaching process based on data- driven compensatory modeling

In the process of hydrometallurgy, the leaching process plays a vital role as the first process of hydrometallurgy. Leaching rate, as the main production index of leaching process, reflects the quality of leaching process, and it is especially important to build a model that can accurately predict the leaching rate. However, due to the long lag, multimodal and nonlinear characteristics of the leaching process, simple methods cannot accurately predict the change of leaching rate. To tackle this issue, this article proposes a solution that involves identifying the data of the same modality and building a leaching rate prediction model. Using the MVO-DBSCAN algorithm for data partitioning and constructing Mechanisms and BiLSTM-Attention Hybrid Models for predicting Leaching Rate. Simulation results show that the proposed method is effective.