Development of predictive control strategies for building climate control

Abstract

The rapid growth in energy usage and CO2 emissions has become a critical issue for the whole world. It is noteworthy that buildings are a major contributor to global primary energy consumption. Among building services, use of energy in heating-ventilation-air-conditioning (HVAC) system is particularly significant (about 50\% of the total building energy consumption). Therefore, the development and implementation of effective control strategies to optimize the operation of HVAC systems in the context of energy usage is essential. One such class of advanced control approaches is model predictive control (MPC). The fundamental idea behind MPC is to use the dynamical model of the process (thermal environment of building in this case) to predict its evolution and optimize the control input signal based on those predictions.

The first part of this thesis provides a detailed introduction to MPC and is devoted to studying the application of MPC in the context of building energy system and climate control.

In the second part of the thesis, a novel concept of cooperative optimization of building energy systems is presented. In this framework, climate control of a group of buildings connected to shared heat pumps is investigated and the benefits of cooperative optimization are analyzed. The proposed approach is capable of achieving a significant reduction in total peak power consumption and can play a major role in advancing the concept of building climate control on a larger scale like smart cities.

The remaining part of the thesis is dedicated to the design of advanced robust MPC controllers to control indoor building environment. These controllers provide robustness against uncertainties present in the system which may arise due to various reasons like plant model mismatch, uncertain system gain and/or additive external disturbances. The robust MPC controller presented in this thesis is intended to handle uncertainty in the thermal model of the building and also additive bounded disturbances to the systems mainly from solar irradiance and ambient temperature.

In the final chapter, a home energy management system (HEMS) is presented which uses MPC and mixed integer linear programming (MILP) to schedule the home appliances in a manner that the peak power consumption and total electricity cost is reduced.