基于粗糙-模糊推理系统的化工过程建模研究

摘要

In the paper we construct a fuzzy neural network based on the rule derived from rough sets method.The initial value of the network parameter is estimated by rule parameter and discretisation results, this makes the trained network be able to converge faster and achieve optimum.It has been applied to modelling the solvent dehydrating tower in PTA complex process, the performance of the model outperforms the common feed-forward neural network.The fuzzy neural network can eliminate redundant information of decision-making system and reduce model' s complexity.%根据粗糙集方法所导出的规则构造模糊-神经网络,由规则的参数和离散化结果估计网络参数的初始值,使网络经训练能较快收敛并达到最优值.将其应用于PTA装置溶剂脱水塔精馏过程建模,所建模型的性能优于普通前馈神经网络,粗糙-模糊神经网络可以消除决策系统的冗余信息,降低模型复杂度.