摘要
为准确、实时地测量海洋生物酶制剂发酵过程中的菌体活性浓度,提出了基于SUKF(scale unscented transformation kalman filter)算法的非线性状态估计软测量.算法采用Kalman滤波框架,嵌入非敏变换,通过变尺度变换有效解决发酵高维非线性模型采样的聚集劣化效应.在σ点集对称采样策略中,根据发酵各维均值的先验信息增加了均值点.通过采用交叉检验办法选择模型参数,并将算法与支持向量机、径向基神经网络算法进行了试验比对.结果表明,SUKF软测量的训练和测试最小方均根统计误差减少2%左右.该软测量方法不需要建立精确的发酵模型和观测模型.对于非线性系统辨识,SUKF具有更好的泛化性能,且方法精度高.
In order to accurately and in real-time measure lysozyme biomass activity concentration in the fermentation process of marine biological enzyme preparation,the soft-sensing algorithm with nonlinear state-estimation was proposed based on scale unscented transformation Kalman filter （SUKF） algorithm.The Kalman filter framework was adopted and embedded with scaled unscented transformation（SUT）.The aggregation degradation effects of high-dimensional and nonlinear fermentation model were effectively solved in sample.By σ-point set with symmetric sampling strategy,the mean points were increased according to the fermentation of priori information of each dimension mean.Using cross-validation method to select model parameters,the method was compared with the support vector machine （SVM） and the radial basis function neural network （RBFNN） algorithm.The results show that the smallest root mean square statistical error between training and testing in soft sensing with SUKF is reduced by about 2％.The establishment of accurate fermentation model or observation model is not required in the proposed method.For nonlinear system identification,SUKF shows better generalization performance with high accuracy.
出处
《江苏大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2013年第6期699-703,共5页
Journal of Jiangsu University：Natural Science Edition
基金
国家"863"计划项目(2011AA09070301)
江苏高校优势学科建设工程项目(PAPD
苏政办发〔2011〕6号)
关键词
溶菌酶
菌体浓度
发酵过程
软测量
卡尔曼滤波
变尺度非敏变换
lysozyme
biomass concentration
fermentation process
soft sensor
Kalman filter
scale unscented transformation
