Intelligent compensation prediction of leaching process based on GRU neural network

The intelligent modeling technology is applied to build a predictive model of leaching system, analyze the process characteristics, and provide scientific foundation for the engineering development and application of the new process. Based on the reaction principle and the theory of mass conservation, a mechanism model is established, which mainly includes gold conservation equations, cyanide ion conservation equations and related kinetic reaction velocity equations. The gated recurrent units (GRU) neural network is set up to compensate the prediction errors caused by the unmolded dynamics of mechanism model, and a parallel hybrid model is established. The hybrid model has the dual benefits of high accuracy of data model and excellent generalizability of mechanism model. The superiority and practicability of the hybrid model are also verified by practical test.