Designing of a Self-Learning Artificial Neural Network Controller for Critical Heating, Ventilation and Air Conditioning Systems

Artificial neural networks (ANN) has emerged as a powerful learning technique to perform complex tasks in highly non linear dynamic environments. This work addresses the stability and efficiency problems with standard Heating, Ventilation and Air-Conditioning (HVAC) systems by implementing a self-learning ANN controller. Although traditional systems such as Proportional, Integral, and Derivative (PID), On-Off controllers, and so on are used, they fail to give intelligence and induce mathematical complexity in implementation. They take longer to acquire a high degree of stability, use a lot of energy, and produce oscillations and peak overshoots. This work focuses on employing a self-learning ANN based intelligent controller to scheme the air cooling system. It uses the user's inputs to estimate the fan and water flow speed in order to provide comfort with the least amount of energy consumption and settling time. The type of practice examined here is applicable to many other types of non-linear control issues. In neural topology, the Back Propagation (BP) technique has been used. The PID and Self Learning ANN controllers have been compared, and MATLAB Simulink has been used to display the results. Using the Self Learning ANN controller architecture, real-time hardware for the HVAC system has been created and compared with the PID controller.