Knowledge-based process identification from relay feedback

We present the development of a knowledge-based approach to dynamics estimation suitable for process control systems. A self-contained module for dynamics estimation is developed which is based on an integration of procedures for frequency response estimation, algorithms for conversion from frequency response to transfer function models and suitable heuristics. Knowledge-based techniques are adopted and this development forms a module within a larger project for the development of an intelligent process control system. The effectiveness of the proposed approach is illustrated by both simulation and real-time experiments for process dynamics estimation on a pilot scale coupled-tanks system.