Functional State Modelling of Biotechnological Processes and Their High Quality Control in the Class of Variable Structure Systems

点击翻译按钮获取中文摘要

In biotechnology there are still high requirements of efficient automation systems and algorithms. Conventional control algorithms, like PI or PID, cannot provide optimum performance over the whole operating range of a fermentation process due to its inherent nonlinear and time- variant properties. The complexity of the fed-batch processes and lack of adequate models limit the application of conventional controllers and require several tuning adjustments. Moreover, a flow injection analysis (FIA) system has been developed, which can measure glucose without any probing system and ensures a fast measurment of glucose. The common modelling approach is to synthesise one global process model but there is a big disadvantage, namely complex model structure and big number of model parameters, which complicate the model simulation and parameter estimation. One alternative, but not so wellstudied, way is the implementation of functional state modelling approach. It is a new concept which helps in monitoring and control of complex processes such as bioprocesses. The main advantage of functional state modelling is that the parameters of each local model can be separately estimated from other local models. The implementation of functional state modelling approach allows the elaboration and tuning of simpel control algorithms for each state, which will be consecutively integrated into variable structure control system. So, high control algorithms, namely variable structure systems, variable structure systems with boundary layer and fuzzy variable structure systems are elaborated to take into account the peculiarities of FIA and to overcome some of disadvantages of conventional control algorithms. For fed-batch cultivations of Escherichia coli and Saccharomyces cerevisiae these high control algorithms allow to control the substrate (glucose) concentration at a low set point so that undesirable byproducts like acetate or ethanol will not be produced during the cultivation and the simulations made presents the efficiency ot these algorithms. One of the goals of the project will be implementation of elaborated high control algorithms for real fermentations of E. coli and S. cerevisiae. The application of presented high control algorithms for real process should be made to establish their applicability and to compare with re sults achieved with working extended Kalman filter. The further required improvements, both in modelling and control, will be done to receive better system performance.