A novel policy to reduce the electrical cost of a real time control process using an on-line Adaptive Fuzzy Logic Controller

With a very limited knowledge of the plant (delay and monotonicity sign) and without any kind of off-line pre-training, a Self-Organizing and Adaptive Fuzzy Logic Controller is developed to reduce the energy consumption during a control process. Our on-line adaptive controller is able to modify and to adapt the rule base of the system by learning from the real behavior of the plant. A comparison results - based on the real electrical consumption - between several conventional control algorithms and the adaptive one is presented, this latter has been able to provide a quite interesting output signal that optimizes the energy consumption without losing the control performance.