A multiple model approach to robust control and operation of complex non-linear processes

A multiple model approach to the dynamics and control of chaotic chemical processes is developed in this paper. In the proposed approach, a complex, non-linear model can be reduced to a set of localised, linear sub-models. Each local model is only valid in a particular operational regime, which is determined by employing the gap metric method. Mini-max optimisation algorithms are used for the design of robust controllers based on the multiple linear models. Because of the incorporation of the gap metric method, the number of operational regimes is minimised. Consequently, the numerical optimisations are carried out more effectively than a similar mini-max optimisation problem solved in the literature with a large number of local models. The novel controller design method developed in this work is successfully applied to a continuous stirred tank reactor with chaotic dynamics and multiple steady states. It is interesting to observe that the closed-loop stability can also be achieved by employing simple PI control loops with a special tuning technique. However, there exists a trade-off between the performance and simplicity, which should be accounted for in decision-making