Increased Energy Efficiency in Buildings using Model Predictive Control

Abstract

The main objective in this thesis is to explore if a model predictive control scheme can increase the energy efficiency in office buildings with waterborne heating systems. The number of office buildings is constantly increasing, which displays the importance of efficient control systems. Since the aim is to decrease the energy consumption, improving the control systems in existing buildings and developing smart control schemes for new buildings are equally important. That is why the vision in this thesis is to design a control scheme that theoretically can be introduced to all new and existingoffice buildings. The model predictive controller s objective is therefore to minimize the supply water temperature to the heating system, while fulfilling a set of defined indoor temperature demands.The model is the most important tool when comparing different control schemes. It is derived using an electrical analogy, and includes the most important thermodynamical relations in three rooms, a ventilation system, and a district heating system. To achieve optimal control of the supply-water temperature, a constrained optimization problem is introduced through a model predictive control scheme. Two versions of the model predictive control scheme are compared with a conventional control scheme. The first version is an ordinary formulation with some ad-hoc solutions including time-varying output constraints, while the second version is a robust formulation includingslack variables. The energy-consumption analysis imply that a 33 % - 34 % reduction potential is obtainable if a model predictive control scheme is introduced.