Dynamic soft sensor modeling based on state detection and impulses response template

Accurate and reliable prediction of melt mass flow rate is crucial in polypropylene production. In order to establish an accurate prediction model, a process state detection method and a novel dynamic modeling method is proposed, and the model parameters are identified by improved swarm optimization algorithm. A polypropylene product melt mass flow rate soft sensor model is established based on process state detection and impulses response template. According to the research on the data from real plant, the experiments demonstrate that even under dynamic state, the proposed approach can improve the prediction accuracy, and the soft sensor model has good tracking ability and meets the requirement of on-line optimal control.