Experimental testing and modelling of variable capacity air-to-air heat pumps

Details

Supervisors

Kummert, Michaël

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] : ingénieur civil électromécanicien

Abstract

Since decarbonizing the electricity grid and reducing society's energy consumption is becoming paramount, it is essential to improve the energy efficiency of the building sector, which accounts for an important part of the global energy consumption. Air-to-air heat pumps (which exchange energy between indoor and ambient air) allow to heat or cool a space with greater efficiency than traditional heating systems, which makes them a key component in improving energy efficiency. They are often preferred to ground source systems because of their lower investment cost; however, they have limitations. Their operating temperature range is limited and their capacity and efficiency decrease at low ambient temperatures, which can be a problem in cold climates. Moreover, most existing heat pumps operate by alternating between on and off modes (i.e. they cycle), which has the side effect of lowering their performance. Variable capacity heat pumps tackle these problems by using variable speed compressors to adjust their capacity as needed, thereby removing the need for cycling. Thus, they can reach higher energy efficiencies, but also have higher investment costs. Variable capacity air-to-air heat pumps (VCAAHPs) are gaining in popularity, especially in cold climates where they can sustain wider temperature ranges. As opposed to fixed capacity heat pumps, VCAAHPs lack simulation models, making it difficult to integrate them into simulations of HVAC systems. Manufacturers provide performance data, but they are severely incomplete, and not always accurate. Therefore, there is a clear need for a VCAAHP model and a more complete performance map. The main goal of the thesis is therefore to develop a VCAAHPs model to use in simulations of HVAC systems. Developed as a component for the TRNSYS software, the model uses the "black box" approach and relies on performance maps. To build a complete performance map that could actually be used by the model, experimental tests were conducted at the CanmetENERGY laboratory in Varennes, Québec. Apart from the performance data, these tests are also useful in establishing the control strategies of the heat pump, which play an important role in the model as well. Test results show that heating capacity stated by manufacturer is slightly overestimated at low temperatures, but it matches well the measurements above -10 °C. Capacity and outdoor temperature values appear to be less correlated in cooling mode, yet similar trends are still observed. The COP values given by the manufacturer seem to match the measurements quite well, which gives confidence to the data used in the model. Regressions are developed from experimental results to obtain performance values for conditions that are not covered by the initial performance map. They allow users to build a complete performance map even without performing tests. Finally, simulations results reproducing the testing environment show a similar behaviour to what is observed in experiments, with some discrepancies. Some behaviours specific to the tested machine were not considered in the model. This makes the model more generic and parametrizable, using data typically provided by manufacturers. Model usage is demonstrated through a case study modelling a residential unit in the Montréal climate. Various implemented behaviours are highlighted by showing results of specific days.