Modeling and control of an air handling unit using dynamic fuzzy neural networks.

Abstract

Air handling units are used everywhere in the world to have controlled temperature as well as humidity in closed areas depending on the requirements of the particular environment. The Air Handling Unit (AHU) has been available for many years and classical control methods are used to control the parameters of the Air Handling Units (AHU ‘s). After the introduction of the Artificial Intelligence (Al), their concepts are getting popular in every aspect of engineering, manufacturing, knowledge discovery systems and data mining applications. Here in our work, Artificial Intelligence concepts are being applied to model and control the Air Handling units to have better control and include intelligence to such control systems. The advent of dynamic fuzzy and neural networks has inspired new resources for possible realization of better and more efficient control. The advantages of fuzzy- neural based control systems over traditional control systems arc that they do not need to know the mathematical model of the physical system, non-linear mapping and self learning abilities. This thesis describes the Air Handling Unit’s components, structure and algorithm of dynamic fuzzy neural networks and how the DFNN is used to control the AHU’s operating parameters by dynamically modeling and controlling the plant. The thesis discusses the control of supply air pressure and off coil temperature of the Air Handling Unit (AHU) using Dynamic Fuzzy Neural Networks (DFNN) with the help of Matlab and Simulink and results are compared with conventional PID controller to illustrate the advantages of using DFNN.