Control of nonlinear time-varying systems using fuzzy relational models

The majority of chemical processes are non-linear in nature. Increasingly, nonlinear models are being used as key parts of chemical plant control schemes. A problem with any model is that it can become inaccurate over time and consequently for good control of time-varying processes some sort of model adaptation is required. In this paper a fuzzy relational model (FRM) is incorporated into a multivariable fuzzy internal model controller (FIMC). A mechanism for online adaptation of the model is described and implemented. Results are presented for a simulated MIMO system.