Skip to Main Content
Complex industrial processes such as chemical reactors, steelmaking, and dyeing processes are difficult to control automatically, due to the complexity and uncertainty of the processes. Parametric control methods require some a priori knowledge of the process model structure as well as possibly the range of the process model parameters. Such methods as nonlinear robust control and adaptive control belong to this category. In the case when model identification is not feasible, nonparametric methods offer an alternative solution to control. One such method, fuzzy logic control (FLC) is used to simulate the decision-making activities of an experienced expert. Usually, the control decisions of an expert can be expressed linguistically as a set of heuristic decision rules. These rules are used to build rulebases for FLC; further algorithms are used to convert the results of these rules to quantitative outputs. Several fuzzy logic control schemes have been developed, including some adaptive methods. However, many of these methods can only handle single input, single output (SISO) systems. In this paper, a multi-input, multi-output scheme is developed. The method is based upon a SISO technique previously developed for a self-learning FLC with on-line scaling factor tuning. The method employs optimization techniques to provide the FLC outputs and an application to dyeing processes is discussed. Simulation results employing the MIMO FLC on a three dyeing process are presented. Finally the experimental dyeing process testbed which will be used for hardware verification is discussed.