An expert adaptive fuzzy logic control system
The objective of this study was to develop a self-tuning rule-based process control system which can use both numerical and symbolic operations. An expert adaptive fuzzy control system was developed and then applied to a thermal process unit configured to possess nonlinear characteristics and uncertain disturbances. The control system included a fuzzy logic controller (FLC), a fuzzy process model and an adaptive algorithm. The multi-layered control system employed linguistic variables and the rule-based fuzzy logic controller on the lowest control level. The next controller level included a fine tuning scheme useful for adapting the scaling factors of the fuzzy logic controller to changing process conditions.
Simulation and real-time results demonstrated that the fuzzy rule-based controller obtained better performance than the PI controller under uncertain conditions. The PI controller was more sensitive to disturbances and dynamics change than were the fuzzy logic controllers. The adaptive algorithms improved the performance of the non-adaptive fuzzy logic controller. These adaptive algorithms decreased response time and reduced the error and offset under uncertain conditions.
In the expert adaptive fuzzy logic control system, the adaptive algorithm used the fuzzy model to speed or correct the control action by adjusting the scaling factors of the fuzzy logic controller. A high degree of accuracy for the fuzzy model was not critical, because there existed a feedback controller (FLC) to compensate for model inaccuracies. Using linguistic variables and an expert system, this control system is easy to read, modify and extend. The expert adaptive fuzzy logic control system is recommended for complicated processes which possess unknown characteristics.