NARMA-L2 neural control of a bioreactor

摘要

This paper presents a bioreactor control using a nonlinear autoregressive moving average (NARMA-L2) neural network based feedback-linearization. The NARMA-L2 neural network is trained off-line for forward dynamics of the bioreactor model with redefined output and is then inverted to force the real output to approximately track a command input. The controller has been able to take care of nonlinearly aspect of the system. Simulation results show that the NARMA-L2 neural control strategy has a better trajectory tracking ability than the use of the inverse neural model control strategy, where the control scheme is not very fruitful since the inverse model developed by the neural network is not accurate enough to exercise effective control.