铅锌烧结过程烧穿点的集成预测模型

An integrated prediction model for burn-through-point in lead-zinc sintering process

针对影响铅锌烧结过程烧穿点的因素具有不确定性的特点,提出一种基于信息熵技术的烧穿点集成预测模型.首先利用软测量技术获得烧穿点.然后建立基于满意聚类的T-S预测模型以降低不确性因素所带来的影响,并将共轭梯度法和粒子群优化算法有机结合起来进行T-S模型中各个子模型的参数辨识,以提高辨识精度.接着建立基于工艺参数的神经网络预测模型.最后考虑到信息熵技术具有信息融合和降低不确定性的能力,利用其将以上预测模型进行集成.实验结果表明所提出的集成预测模型具有较高的预测精度和较强的适应性.

To deal with the uncertainty in the determination of the burn-through-point（BTP） in a lead-zinc sintering process, we develop an integrated prediction model based on the information entropy technology to predict the BTP. The BTP value is acquired by using the soft-sensor technology; a Takagi-Surgeon（T-S） model of satisfactory clustering is developed to reduce the negative effects brought by the uncertainties. To improve the identification accuracy, a particle swarm optimization algorithm combined organically with the conjugate gradient algorithm is applied to identify the parameters of each sub-model of the T-S prediction model. Next, a technological-parameter-based model for predicting the BTP is established using neural networks. To make use of the capabilities of information fusion and uncertainties reduction in the information entropy technology, we integrate the two prediction models by using the information entropy technology. The experiment results show that the proposed integrated prediction method features high precision and strong adaptability.