Modeling And Control For Recombinant E. Coli High-Density Fermentation Process

Recently, with the increasing demand for bio-fermentation products in various industries, the producing process of bio-fermentation has been regarded widely. For the fermentation process, the focus research is to optimize the control process and finally increase production and save production cost. However, fermentation is a time-varying, nonlinear and high coupling process, the internal mechanism is very complicated, and some important variables are difficult to be measured on line. These problems make it not easy to model and control the fermentation process and can't meet the production needs nowadays.For the above problems, this paper gives the modeling method and optimal control scheme for high-density E. coli fermentation process by analyzing the process and combining with the related algorithms and theories of intelligent control system.In this paper first uses the least squares support vector machines (LS-SVM) method, which is good for non-linear functional approximation, to develop a model for high-density E. coli fermentation process with limited sample data, and verify the accuracy of the model; Secondly, we put forward the LS-SVM inverse system method by combing the LS-SVM algorithm with inverse system method. This method has realized to decouple the complicated nonlinear fermentation system and control the fermentation process by using LS-SVM to identify the inverse model of the process and constructing a pseudo-linear system though connecting the obtained LS-SVM inverse system in front of the original system; Finally, in order to improve the capacity of anti-interference and robustness, a decoupling internal model control scheme based on LS-SVM inverse system method is proposed. Through a lot of simulation experiments, the accuracy of modeling method and the feasibility of control scheme in different situations is proved to achieve satisfactory results.The methods proposed in this paper can also be used for other fermentation processes and complex non-linear system control areas.