Air conditioner control using neural network and PID controller

Abstract

The air conditioning system usually controlled by a simple, proportional integral and derivative (PID) controller with single loop, or by using on-off controller, these controller are widely used because it's simple, but the performance of such a system using these controlling techniques are not accurate, with high power consumption and short compressor life. This paper aims to control air conditioning system using neural network method with a PID controller to increase the efficiency of the system, enhance the performance, and reducing power consumption through a smooth control of the compressor motor. The simulation result shows that using Neural network method either for tuning the PID parameters or for performing the speed control of the compressor is more superior in terms of performance compared to traditional PID method. An optimized control model were obtained resulting in an excellent performance, zero over shoot with fast response time for a multi-input data with a mean square error of 0.0692, a correction values of 0.9998 between input and output, a performance 0.1497 and with an error of 0.0907 .