基于LS-SVM的水泥熟料煅烧过程f-CaO软测量研究

针对水泥厂熟料质量指标f-CaO含量难以在线测量的问题,提出一种基于最小二乘支持向量机(LS-SVM)的软测量建模方法。仿真实验研究表明,最小二乘支持向量机建模具有良好的学习能力和泛化性能,且对数据样本的依赖程度低,是一种有效的软测量建模方法。给出的以DCS系统和SQL数据库为基础,通过Visual C 6.0编程的软测量实现方案,符合国内水泥厂实际情况,易于实现。建立的软测量模型和给出的软仪表实现方案对于保证水泥熟料质量和实现回转窑烧成系统的优化控制具有重要意义。

Soft-Sensor for f-CaO in Cement Clinker Calcination Process Based on LS-SVM

Aiming at the difficulty in measuring the content of f-CaO in cement clinker,a soft-sensor modeling is presented based on least squares support vector machine(LS-SVM),and a realization scheme is given.Study of simulation experiment shows that the LS-SVM has good learning and generalization capability,and less dependence on sample set.The soft-sensor scheme is based on DCS and SQL database programmed by Visual C 6.0.This scheme accords with the actual situation of the domestic cement plant,and it is easy to realize.The soft-sensor modeling and realization scheme are of important significances for ensuring the quality of cement clinker and realizing optimal control of the whole kiln systems.