The multi-model predictive control method research on the outlet temperature control of ethylene cracking furnace

Multi-model predictive control has become an effective method for handling the process of nonlinear system. But the traditional multi-model has large model tracking error and slow convergence speed when it is used to the MIMO nonlinear system solving the condition with large scale transition of operating condition. To solve these problems, a new structure of multi-model called multi-hierarchical model has been presented. This structure consists of many layers that each layer is comprised multiple models. The number of sub-models in each layer is different. In the condition of the same global operation space, upper layer has smaller number of sub-models, and lower layer has lager number of sub-models. Because of this structure, the models chosen from different layers can deal with the operating condition changed with large scale. But, the model switching method presented by the author is not flexible enough. In this paper, a new model switching method between different layers is presented. This method uses the error of output and the variation of output error as the layer switching rules. In the end of this paper, the simulation results of pH neutralization process which is a MIMO nonlinear system demonstrate that the multi-hierarchical model using the new model switching method is superior to single-hierarchical model with smaller model tracking error, better convergence speed and stability.