Analysis of model predictive control system for building automation and control

Abstract

Model Predictive Control (MPC) has been extensively researched on in recent years as computational power has increased to meet its computational needs. Compared to the less sophisticated Proportional-Integral-Derivative (PID) controller, MPC promises to control processes in a more effective and efficient way. Unlike PID controllers which are limited to controlling processes towards a setpoint, MPC has objective functions, constraints, prediction models and optimization to help it control processes. The MPC system implemented was implemented in an office-school hybrid building. This MPC system was multi-zonal and multi-objective, handling ACMV, lighting and shading systems all at once. Energy savings of up to 16.7% was achieved by the MPC system in one of the zones, and projection of the energy consumption towards the entire test site would show that a potential energy saving of 5.35% can be achieved.