Краткое описание

Conventional expert systems consist of rules connected according to different mechanismes depending on boolean logic. It implies strict definitions of the rules and drastic choices at the time of their utilization. Fuzzy logic is an approach to artificial intelligence, grounded on fuzzy sets theory. It manipulates qualitative concepts and uncertain values.
This paper deals with fuzzy control of thermal systems. In a first part, we investigate by simulating the control of a hot water installation. We treat of self-learning, and of the robustness of fuzzy control compared with a standard PID control, in the case of a variable water flow rate (variable time delay). In a second part, we present a simulated regulating of the inside air temperature of a building. We deal with self-learning, closed loop control, and with the improvements brought by fuzzyness with respect to usual open loop control.