The nonlinear process control with internal model and multilayer neural

When the method for the calculation of the command law does not exist, it is advantageous to use control structures which are implemented with neural networks. This paper deals with the following aspects: 1) the internal model control with emphasises on the performance of the controlled system regarding the input signal and the perturbation; and 2) the controlling system behaviour at the process parameter variation and at the manifestation of the unmodeled dynamics.