Multi-Phase Modeling Of Nosiheptide Batch Fermentation Process Based On GA

With the quick development of biological technology, microorganism fermentation has become more and more important in national economy. Control and optimization of fermentation process is of great importance to improve product quality, and reduce energy consumption and production costs, while modeling of fermentation process is the basis of fermentation control and optimization.As a kind of microorganism fermentation products, nosiheptide was discovered by French in the1960s and by years of research, it has been developed as an effective and low toxicity antibiotic growth promoter. Nosiheptide has wide prospect of market, so modeling of nosiheptide fermentation process and carrying on optimal control research based on the model has important significance and practical value. The thesis is concerned with the problem of multi-phase modeling of nosiheptide batch fermentation process, and the design of modeling platform. The main contents are as follows:According to fermentation kinetics, the model of nosiheptide batch fermentation is developed, including the biomass growth model, the substrate utilization model, the dissolved oxygen model and the nosiheptide production model. Through analyzing the multi-phase property of nosiheptide fermentation process, then dividing the parameters which need to be identified into two parts:one part of the parameters’values are not affected and remain the same during the whole nosiheptide fermentation process, while the other part is affected by the different phases properties of the nosiheptide fermentation process, then the parameters need to be sub-valued. GA (genetic algorithms) is used to determine the parameters of this model. Using experimental data to verify the model, it is shown that the model fits well to the experimental results of nosiheptide fermentation process. Then the model is put into online application, with OUR-based identification of growth phases to determine the phase conversion time of fermentation process, and nosiheptide fermentation process online prediction is realized.The nosiheptide fermentation process platform is also developed based on the models. This platform combines programming features of C#. numerical analysis and graphical display function of Matlab and powerful data access capacity of Oracle. In the offline modeling interface, users can simply select the appropriate kinetic model, historical data, then the corresponding model parameters can be estimated, and a multi-phase model of the fermentation process is established:in the online application interface, users just need to input initial parameters according to actual situation. Based on online mearments and phase switching time point which can be calculated according to pressure in gas volume fraction, the cell concentration, substrate concentration, dissolved oxygen concentration, product concentration can be predicted and the corresponding data can be saved.