Models for control of thermal energy in buildings

Abstract

A large fraction of the world’s energy production is used for HVAC in buildings. It is therefore important to develop improved strategies for the efficient use of energy in buildings. Storage of intermittent energy production is important; storage as hot water in water tanks is the most common way to store energy in private homes/ smaller apartment complexes. Finding good models for building thermal behavior is an important part of developing building energy management systems (BEMS) that are capable of reducing energy consumption for space heating through model predictive control (MPC). In this paper, previous models of temperature dynamics in hot water tanks are considered, and a simple well mixed tank model is compared with a model describing a more realistic stratified temperature distribution. Two models are fitted to experimental data from a hot water tank. Description of temperature stratification requires a distributed model, but a relatively low order discretized model suffices to describe the important effect while simultaneously being useful for BEMS. A suitable hot water tank model in combination with weather forecast enables temperature estimation and prediction in MPC, and allows for finding a suitable water temperature at minimal energy consumption.