New boundary conditions and adaptive control of fixed-bed bioreactor s

Modelling, simulation, state estimation and feedback control of a distributed parameter fixed-bed bioreactor are studied in reference to two different types of boundary conditions. The bioreactor is described by two nonlinear and coupled partial differential equations including a single space variable. Classical Danckwerts boundary conditions and so-called Salmi and Romanainen boundary conditions are the two alternatives applied. For process model simulation the method of orthogonal collocation is used in the spatial discretisation. The idea of linearising state-feedback control is applied in the controller design. For control law calculation, two estimation procedures are used: one for estimating the state variables and the specific growth rate, and the other for obtaining the substrate consumption rate of the micro-organisms in the reactor. The state estimator is based on asymptotic observer design. Estimation of the substrate consumption rate is based on temporal discretisation and on the recursive exponentially weighted least-squares procedure resulting in adaptive nature of the final control law.