Development and Implementation of a Biomass Soft-sensor Model in Penicillin Fermentation Process
The soft-sensor model that depicts the mathematic relation between auxiliary variables and primary variables is the key of the soft-sensor technology. Considering the complicated modeling of the penicillin fermentation process, the biomass soft-sensor model using a hybrid soft-sensor modeling method is proposed. The hybrid model can represent various soft-sensor models in a unified form model. The hybrid model combines the first principle model (FPM) and the data-driven model. For the part of the FPM, the particle swarm optimization (PSO) is used for parameter determination according to the given optimal objective of the quadratic form. For the part of the data-driven model, the support vector machine (SVM), which maps the input data into a high (possible infinite) dimensional space and constructs an optimal hyperplane in this space, is adopted to avoid local minimum value. This modeling method has been used in the modeling of the penicillin fermentation process and the simulation results indicate that the feasibility of the method that can guarantee the reliability of the hybrid model.
biomass soft-sensor hybrid model first principle model particle swarm optimization support vector machine
ZHAO Liqiang WANG Jianlin YU Tao
School of Information Science and Technology, Beijing University of Chemical Technology, Beijing, China, 100029
国际会议
第八届国际测试技术研讨会(8th International Symposium on Test and Measurement)
重庆
英文
3352-3355
2009-08-01（万方平台首次上网日期，不代表论文的发表时间）