Generalized multivariable dynamic artificial neurol network modeling for chemical processes

This work presents a navel systematic approach to acquire good-quality plant data that can be efficiently used to build a complete dynamic empirical model along with the use of partial plant knowledge. A generalized Delaunay triangulation scheme is then implemented to find feasible operating boundaries that may be nonconvex an the basis of the existing plant data. The Akaike information index is adopted to assess partial plant knowledge as well as noisy plant data. The information free energy is calculated for acquisition of good-quality new plant data that will improve the dynamic model. The new experimental data suggested by the information analysis, together with the previous data sand prior plant knowledge, are used to train a new dynamic empirical model, Multivariable model predictive control for a high-purity distillation column using the acquired. model based on the proposed approach is also studied. Comparing with PRBS and RAS schemes, the proposed approach outperforms the rest.