NONLINEAR MODELING OF COMPLEX LARGE-SCALE PLANTS USING NEURAL NETWORKS AND STOCHASTIC-APPROXIMATION

This paper deals with a general methodology for system grey-box Identi fication. As is well-known, the tuning of accurate models of real plan ts (obtained, for instance, by using the physical knowledge of the pla nts and the technicians' expertise), on the basis of the measures prov ided by the available sensors, remains a challenge. In this paper, a t uning methodology for complex large-scale models, is presented. The pr oposed technique is based on the suitable use of neural networks and s pecific stochastic-approximation algorithms. It is therefore possible to design a simulator that can be connected in parallel with a real pl ant, thus providing the plant technician with information about inacce ssible variables that are useful for supervision purposes. The propose d methodology is applied to a section of a real 320 MW power plant. Si mulation results on the tuning algorithm show the effectiveness of the approach.