A new prediction model based on the leaching rate kinetics in the alumina digestion process

The leaching rate of alumina in the alumina digestion process is usually obtained via off-line analysis with a long time delay, leading to delayed control of the process and creating ongoing problems, such as a low leaching rate and wasted energy. Therefore, prediction of the online leaching rate is highly important. Based on mechanistic analysis of the double stream digestion process and the digestion kinetics of diaspore, a kinetics model established at the laboratory scale was scaled up to an industrial process. The unknown model parameters were estimated from the industrial data using a state transition algorithm (STA), which is a new and effective optimization algorithm. An error compensation model based on the kernel extreme learning machine (KELM) was subsequently built, and a prediction model for the leaching rate of alumina was established by parallel connection of the kinetics model with the compensation model. The validation results show that the model can predict the leaching rate of alumina for 90% of the samples with relative errors within ± 2% compared with the actual industrial data. The developed model will be further evaluated for control in the corresponding industrial process.