基于LSSVM的氧化铝浓度软测量  

Soft measurement model of alumina density based on LSSVM

作　　者：林景栋[1] 徐春慧[1] 张瑞洋[1] 

机构地区：[1]重庆大学自动化学院,重庆404100

出　　处：《轻金属》2014年第4期31-35,共5页Light Metals

基　　金：重庆市自然科学基金(CSTC2009BB3209):知识与数据驱动的电解槽优化控制

摘　　要：针对铝电解工业中氧化铝浓度直接测量难度大、效率低、实时性差等特点,对铝电解槽数据库中的生产过程参数进行研究,提出一种基于LSSVM的氧化铝浓度软测量方法。首先,通过对铝电解工艺的分析和实际生产的观测确定模型变量,建立模型;然后,运用径向基核函数和粒子群算法对模型的参数进行处理和优化;最后,运用LSSVM基本原理和回归分析实现浓度测量。实践及仿真结果表明,采用该方法提高了氧化铝浓度测量的准确性和铝电解过程的稳定性。According to the difficulty ol direct measurement of the alumina concentration, low efficiency, poor real - time characteristics, aluminum reduction process parameters in the aluminum reduction cell database are studied, and a soft measurement model of alumina density based on least squares sup- port vector machine （LSSVM） is proposed. First through analysis of the aluminum reduction process and observation of the actual production experience, the model is established. Then the radial basis kernel function and particle swarm algorithm are used to process and optimize the model parameters. Finally, the basic principle of LSSVM and regression analysis arc used, the alumina concentration measurement is realized. Practice and simulation results show that this method has good predictive ability to improve the accuracy of alumina density and alumina concentration measurement process stability.

关 键 词：氧化铝浓度 小二乘支持向量机 向基核函数 子群算法 

分 类 号：TF821[冶金工程—有色金属冶金]