基于LS-SVM苯乳酸发酵过程的建模

为了解决苯乳酸发酵过程中关键生物参数难以直接在线检测的问题，提出了基于最小二乘支持向量机（LS-SVM）的软测量建模方法. 通过使用径向基核函数来对菌体浓度、苯乳酸浓度建立模型，对建模的理论进行了分析和并进行了仿真研究，同时还采用支持向量机对过程进行了建模，对两种方法的优缺点进行了比较. 结果表明，基于LS-SVM的建模方法预测精度高、跟踪性能好，能提高在线预估的效率，非常适合于苯乳酸发酵过程的在线预估.

Modeling of Phenyllactic Acid Fermentation Process Based on Least Square Support Vector Machine

To solve the difficulties of online measurement for crucial biological variables in the phenyllactic acid fermentation process, a soft sensor modeling method was proposed based on the least squares support vector machine （LS-SVM）, and the model for concentration of mycelium and phenyllactic acid was built by kernel of Radial Basis Function. Theoretical analysis and simulation?study of the modeling was investigated, and a second modeling process was constructed by the support vector machine. Finally, the effects of the two methods modeling were compared. The results show that the modeling method based on the LS-SVM has the advantages of accuracy predition， good tracking performance， improving efficiency of on-line predition, which is very suitable for the on-line estimation of the phenyllactic acid fermentation process.