基于模糊分类变系数的铅锌烧结过程综合透气性状态预测

针对铅锌烧结过程的强非线性、时变等特点,运用智能集成建模的思想,提出一种模糊分类变系数透气性状态预测方法。首先深入机理分析和工况参数相关性分析研究,采用神经网络方法建立工艺参数和时间序列透气性预测模型;然后借助于模糊组合器实现两个子模型的有机组合,设计了模糊分类变系数综合透气性集成预测模型结构,其中加权系数由工况波动程度确定。运行结果表明:提出的集成模型具有较高的预测精度和较强的自学习能力,并且在工况波动严重的情况下,仍然具有好的预测效果,该模型具有一定的灵敏度和鲁棒性。

Synthetical Permeability State Prediction Based on Fuzzy Coefficient-Variable for Lead-Zinc Sintering Process

Considering the characteristics such as strong nonlinear,time-varying in the lead-zinc sintering process,the predictive method of fuzzy coefficient-variable of permeability is presented by applying the idea of intelligent integrated modeling.Based on the analysis of mechanism and the relativity of technology parameters,the neural network predictive models of technology parameters and time sequence are(established).The structure of fuzzy coefficient-variable of synthetical permeability integrated predictive model of lead-zinc sintering process is designed through a fuzzy classifier.The weight coefficient of these two sub-models is decided by status fluctuation.The results of simulation show that the proposed model possesses high precision and strong self-study ability.The precision of the integrated model is higher than that of either sub-model,even on the condition of big fluctuation,and this model possesses definite sensitivity and robustness.