A predictive fuzzy logic controller with an adaptive loop for the manufacture of resin adhesives

Control of exothermic reaction in manufacturing of resin adhesives has been performed, for long, through human reasoning and expertise. Human involvement has been a source of control errors, which affect the quality of resin product. An earlier paper on this topic has incorporated fuzzy logic control (FLC) in the control loop of the manufacturing process to support and accommodate human expertise and to avoid human errors in maintaining a sustained product quality. A one-step predictor has also been included in the FLC loop to compensate for the inherent time delay in the process system. The current research substantiates the FLC methodology and proposes further modifications in the fuzzy logic controller in order to overcome some shortfalls in the earlier system such as reduced speed of control and varying steady state errors in temperature responses. The main reason for these shortfalls is some unmeasurable parameters of the system that vary slowly and unpredictably. An approach of adaptive control is found appropriate in compensating such parameter variations and is applied as an outer loop to the earlier FLC system. Several experiments performed on this system confirm that the new proposal is effective in overcoming the shortfalls of the earlier FLC system. (C) 2001 Elsevier Science Ltd. All rights reserved.