< Back to previous page

Project

Modeling, Optimal Control and HVAC Design of Large Buildings using Ground Source Heat Pump Systems

Since May 2010, the directive 2010/31/EU of the European Parliament compels its Member States to drastically decrease the energy use of buildings, to increase their energy efficiency and to increase the relative amount of renewable energy they use. One of the technologies recommended by the directive is the heat pump which efficiently uses electricity to extract thermal energy from a heat or cold source. In this work, buildings equipped with the particularly efficient hybrid GEOTABS system are considered, consisting of a ground source heat pump (GSHP) coupled to a thermally activated building structure (TABS) system and optionally extended with a gas boiler, radiators or other auxiliary systems. The main objective of this research is to improve the thermal comfort and the energy efficiency of large hybrid GEOTABS buildings by applying model predictive control (MPC) and to improve their economic viability by optimizing the size of the GSHP and of the auxiliary systems as well as the type of the auxiliary systems to install. To this end, Building Energy Simulation (BES) models of an existing office building, a school, a retirement home and a block of flats were created using and extending the open-source Modelica library IDEAS to represent a wide set of hybrid GEOTABS buildings. The models include the building envelope, the heating, ventilation and air conditioning (HVAC) system, the occupancy and a default rule-based building climate controller (RBC). Additionally, a method to linearise the building envelope of the developed Modelica models was developed in order to obtain highly accurate controller models for MPC. The method automatically precomputes the non-linear equations which do not depend on the model states and linearises the other equations. The obtained controller models are then used by a toolchain which semi-automatically generates a linear MPC and tests its control performance on a full year simulation of the developed BES models. Finally, a python tool was created to optimize the economic viability and CO2 emissions of hybrid GEOTABS systems. 
As main results, it was found that hybrid GEOTABS systems were capable of providing very high thermal comfort in all investigated buildings when controlled by MPC, showing that hybrid GEOTABS systems are suitable for a wide range of buildings when appropriate control is used. The developed MPCs could further save between 30 to 50% of the energy cost compared to standard RBC controllers while significantly improving thermal comfort when both the thermal powers to the TABS and the auxiliary emission system and the ventilation supply temperature were optimized simultaneously. However, current practice RBC were found not to always be able to provide the required comfort when, for example, the building was not equipped with a fast reacting system such as an air handling unit with heating and cooling coils to complement the TABS or when parts of the building with significantly different thermal needs were coupled to the same production system. Furthermore, the lack of correlation between the optimal control actions computed by MPC and the past or future ambient temperature indicates that the optimal behaviour achieved by MPC cannot be mimicked by RBC based on heating/cooling curves. Finally, the economic optimization and CO2 emissions analysis of HVAC designs showed that GEOTABS systems without auxiliary production and emission systems are in general 1.0 to 1.8 times more expensive than conventional systems composed of a condensing gas boiler, a compression cooling machine and fan coil units but that properly sized hybrid GEOTABS systems (thus including an auxiliary production and emission system) have in general a lower present cost (PC) over the 20 years life time of the building while they emit between 20% to more than 50% less CO2 than conventional systems. Hybrid GEOTABS systems are thus very advantageous both economically and for the environment and it is advised to always consider their installation for large buildings. 
This work significantly contributed to the fields of building simulation and optimal control by developing new models and tools such as a novel borefield model, contributing to the development of the open-source libraries IDEAS and Annex60, and creating a highly automated method to generate accurate linear building models for MPC. Furthermore, the MPCs developed in this work achieved between 30 to 50% energy cost savings and significant thermal comfort increase compared to current practice rule-based-controllers which goes beyond the typical 15 to 25% cost savings found in the literature. This work also helped the industry forward by designing a tool to help design engineers to choose and optimally size the most appropriate HVAC system for a given building.

Date:4 Jun 2012 →  26 Sep 2017
Keywords:Ground coupled heat pump, Geothermie, Model predictive control
Disciplines:Electrical power engineering, Energy generation, conversion and storage engineering, Thermodynamics, Mechanics, Mechatronics and robotics, Manufacturing engineering, Safety engineering
Project type:PhD project