Application of Bayesian Optimization GRU in Dynamic Modeling of Cement Rotary Kiln Firing Process

Aiming at the complex dynamic characteristics of nonlinear, large time delay and multivariable strong coupling in the firing process of cement rotary kiln and the insufficient accuracy of traditional mechanism modeling, this paper proposes a data-driven modeling method based on Bayesian optimization gated cyclic unit. The Bayesian optimization algorithm is used to perform global adaptive optimization on hyperparameters such as hidden layer dimension and learning rate of GRU network, so as to solve the bottleneck problem of low efficiency of manual parameter adjustment and poor generalization of model. This model effectively overcomes the limitation of the mechanism model's simplification assumption for complex heat and mass transfer processes, and it provides a reliable theoretical basis for the model construction of cement firing process simulation system.