高斯过程集成算法的发酵过程软测量建模

Soft-sensor modeling of fermentation process based on Gaussian process ensemble model

针对谷氨酸发酵过程一些关键参数不能在线测量而导致的建模精度不高的问题,Bagging和高斯过程回归算法相结合,提出一种基于Bagging算法集成高斯过程的软测量建模方法。该算法使用Bagging技术从训练样本集中选取若干子训练样本集,利用该若干子集形成许多高斯过程模型,并通过平均组合方式进行集成,得到最终的模型输出。将该集成算法应用到谷氨酸发酵过程的软测量建模中,实现了对谷氨酸浓度的准确预测,相对于单一高斯过程模型,具有更高的预测精度和鲁棒性。

To solve the problem of low modeling precision due to some key parameters unable to be measured on line in the glutamic acid process,a Gaussian process ensemble model for soft-sensor is proposed based on Bagging and Gaussian process algorithms.This algorithm selects a number of sub-training sets from the whole training sample using Bag-ging technique,and then trains many Gaussian process sub-models using the sub-training sets respectively.The output of the model is achieved by averaging all the outputs of the sub-models.The ensemble algorithm is applied to the soft sensor model of the glutamic acid fermentation process and achieves more accurate prediction and better robustness compared with single Gaussian process model.