The development of a relational fuzzy model based controller for an industrial process

A fuzzy logic control system for difficult nonlinear processes has been developed at Strathclyde University. This controller is structured in a very similar manner to the standard internal model control design, differing only from the conventional system in that it uses a relational fuzzy model. The relational fuzzy model-based control system has previously been tested extensively on simulations and lab-scale experiments. The work discussed in this paper describes the preliminary steps for the first industrial application of this control system. The control problem investigated is the control of moisture in the outlet product from a dryer. The dryer is used to process material left over from Whisky distillation into animal feed. The material is dried by contact with a hot `screw´ arrangement which runs through the centre of the dryer. The control is difficult in the dryer both because of the dead-time and also due to the nonlinearity which occurs with changes in the dryer loading