PID neural network control of SUT building energy management system

PID neural network controller design for Sharif University of Technology (SUT) building energy management system (BEMS) is addressed in this paper. The most important characteristics of process systems are time delay with model uncertainties. Artificial neural networks can perform adaptive controller properties through learning processes. PID neural network has the advantage of both conventional PID controllers and the neural structure. Simulation results using modified Hooke-Jeeves optimization method show that this controller has short convergence time and quick learning speed and the performance of the closed loop system is very good