Behrouz Memarzadeh - Electrical engineering Department Bahonar University Kerman, Iran
Mohammad Ali Mohammadi - Electrical engineering Department, Bahonar University Kerman, Iran

The air-conditioning systems now are usually used in the buildings. They are designed to make a good atmospheric environment for people to live or to produce a special environment for some special industrial places. Temperature control is important for both human comfort and the need in industry. In this paper, first a model for building is built based on some assumptions: no influence for heat; no influence from the people activities in the room. Then a Neuro-Fuzzy method is used to keep the temperat-ure indoor around 22°C with the control from fuzzy logic and the learning process in neural network. A multi-layer neural network is used to establish the fuzzy logic system. The advantage of fuzzy logic is that it can build the contro-ller without the model, so in this paper, it is easy to apply in the air-conditioning system. Then a self-tuning fuzzy logic is designed. A neural network is added in the system controller. The controller puts the advantages of fuzzy logic and ANNs together. In the method, the neural network is to build the structure for the fuzzy controller and the weights in neural network can represent the parameters of fuzzy logic system. The air-conditioning system in the paper is the heat pump system in one room which is used to supply different temperature to the room. And there is one fan working with constant air flow rate in the system. Matlab is used as the tool to make models and process calculations.

Modelling, temperature control, Fuzzy Logic, Neural Network, HVAC system