中文摘要：

针对氧化铝焙烧过程具有强非线性、检测滞后等特点,提出一种基于混沌灰狼优化算法（CGWO）参数优化在线贯序极限学习机（OSELM）的氧化铝质量预测模型。在基于机理分析和变量相关性分析的基础上,选择氧化铝质量指标预测模型的输入变量,采用在线序贯极限学习机的方法建立模型,并利用改进的混沌灰狼优化算法得到最优的初始权值和隐含层偏差,实现焙烧过程氧化铝质量预测建模。采用工业过程数据对提出的方法进行实验验证,仿真结果表明：所建立的预测模型具有更好的精度,从而验证了方法的有效性。

英文摘要：

Aiming to the problems in the alumina sintering process,such as strong non-linearity and large time delay,a model of online sequential extreme learning machine based on chaotic grey wolf optimization（ CGWO） is proposed,for predicting the quality of alumina sintering process. The input variables of predicting model for alumina quality indicator are chosen based on the mechanism analysis and the correlation analysis of variables,and then the model is established by using online alumina Sequential Extreme Learning Machine（ OSELM）. The chaotic grey wolf optimization algorithm is applied to optimize the parameters including the initial weights and hidden layer bias in order to obtain the predicting model of quality in alumina sintering process. Based on industrial process data,the proposed method is verified. The prediction results show that the proposed models have better performance in terms of root mean square error and accuracy,thus the effectiveness of the method is verified.