Application of fuzzy control in naturally ventilated buildings for summer conditions

The objective of this work is to develop a fuzzy controller for naturally ventilated buildings. Approximate reasoning has proven to be in many cases more successful control strategy than classically designed controlled scheme. In this paper the process of designing a supervisory control to provide thermal comfort and adequate air distribution inside a single-sided naturally ventilated test room is described. The controller is based on fuzzy logic reasoning and sets of linguistic rules in forms of IF-THEN rules are used. The inputs to the controller are the outside wind velocity, direction, outside and inside temperatures. The output is the position of the opening. A selection of membership functions for input and output variables are described and analysed. The control strategy consisting of the expert rules is then validated using experimental data from a naturally ventilated test room. The test room is located in a sheltered area and air flow inside the room, the air pressures and velocities across the openings together with indoor air temperature and velocity at four locations and six different levels were measured. Validation of the controller is performed in the test room by measuring the air distribution and thermal comfort inside the room with no control action. These data are then compared to the air temperature and velocity with the controller in action. The initial results are presented here, which shows that the controller is capable of providing better thermal comfort inside the room.