Hierarchical integrated system optimisation and parameter estimation using a microcomputer based system

The advent of inexpensive and more powerful microcomputers has facilitated the implementation of on-line optimization schemes to industrial processes. Most optimization algorithms require a process model and their performance clearly depends on how accurate the model is to reality. Unfortunately, for many industrial processes an accurate model may be very difficult, time consuming or impractical to obtain. In this paper a hierarchical integrated system optimisation and parameter estimation technique developed by combining the interaction balance method[6] and the modified two-step approach[1][8] has been investigated using a microcomputer based system. Real-time simulation results obtained show that two versions of this technique (a single iterative algorithm and a double iterative algorithm) are capable of achieving the real process steady state optimal operating condition in spite of the model used to solve the optimisation problem not being a faithful representation of reality. The results also demonstrate that the double iterative version has an important advantage under real time operation. It significantly reduces the time for determining the real optimum operating condition.