Increased Efficiency Of Hybrid Heat Pump Systems Using Optimal Control In The Redesign Of A Historic Neighborhood’s Heating Supply

Karl Walther, Louis Hermans, Lone Meertens, Lieve Helsen, Belgium

This article illustrates the increased efficiency of optimally controlled hybrid heat pump systems, including air-source and ground-source heat pumps, compared to the respective single-source scenarios. The redesign of the historic neighborhood “Stijn Streuvelstraat” in Bruges / Belgium, is used as an example. Physics-based models of the neighborhood and different supply options are used for a virtual comparison of conventional rule-based controls and optimal control. The efficiency gains of hybrid heat pump systems stem from the ability of optimal control to take into account different source temperatures and the predicted building behavior.

Introduction

Decarbonizing the residential heating sector is a major challenge to achieve the energy and emission targets. For future supply systems based on Renewable and Residual Energy Sources (RES), two topologies are typically distinguished: (1) individual heat pump systems on building level, typically using ambient air or geothermal heat, and (2) collective district heating networks. For modern district heating networks, two sub-topologies are generally distinguished: (a) centrally generated heat on a high temperature level, so called 4^th generation district heating networks (4GDHN), typically through renewable energies (heat pumps, biomass, etc.) or residual heat (waste incineration, industrial processes, etc.), or (b) low temperature networks, so-called 5^th generation district heating and cooling networks (5GDHCN) with low-temperature heat sources (solar thermal, data centers, etc.) and decentral heat pumps on building level.

This article evaluates the increased efficiency of hybrid heat pump systems integrating air-source heat pumps (ASHPs) and ground-source heat pumps (GSHPs) for the collective heat supply in a 4GDHN compared to alternative ASHP-only or GSHP-only scenarios. The redesign of the historic neighborhood “Stijn Streuvelstraat” in Bruges / Belgium, is used as a practical demonstration case. The analyses use physics-based models of the neighborhood and the heating supply to illustrate the potential of an optimally controlled system compared to conventional rule-based controls. The article is based on experiences during the design phase of the “Stijn Streuvelstraat” project. The article uses extended content of [1].

The “Stijn Streuvelstraat” pilot in Bruges, Belgium

“Stijn Streuvelstraat” is a historical neighborhood in Bruges (Belgium) with 15 dwelling units for assisted living (see Figure 1 and Figure 2). Each dwelling unit has a living room, a sleeping room, and a bathroom on the ground floor. The attic spaces are used for technical installations. The total net heated floor area is about 1,000 m². Internal insulation is used for the external walls (U-value 0.27 W/m²K) because the heritage character does not allow external insulation. The target of the renovation is, next to the building retrofit, to design and implement a collective heat supply system that is

• 100 % R²ES-based
• Future-proof facing the effects of climate change
• Cost-efficient
• Capable of providing good indoor thermal comfort

Against the background of these design goals and considering the site conditions, ASHPs or GSHPs are viable heating supply technologies. A solely ASHP-based option was not possible due to acoustic constraints. A solely GSHP-based system was excluded due to the costs of a borefield that is sized robustly enough to take into account uncertainties on the demand side, for example, through changing and extreme climatic conditions. Therefore, a hybrid heat pump supply system including ASHPs and GSHPs was designed (see Figure 2 and Figure 3). The maximum thermal capacities are 28 kW_th for the ASHP and 44 kW_th for the GSHP. Both ASHP and GSHP are modulating. Due to acoustic constraints, ASHPs are limited to a maximum of 80 % operation during night hours. The borefield is split into two parts with four boreholes with 125 m depth each. A buffer tank of 0.5 m³ is located between the supply and demand side. A photovoltaic thermal (PVT) system enables borefield regeneration in summer. Passive cooling is possible via a heat exchanger and provides additional borefield regeneration potential in summer. Domestic hot water (DHW) is provided decentrally in each dwelling unit by an electric heater. Floor heating/cooling was selected as a room-side emission system. Temperature setpoints are defined at 21 °C (heating) and 26 °C (cooling). Each dwelling unit has a mechanical ventilation system with heat recovery.

Simulated efficiency gains of an optimally controlled hybrid system

As described previously, single-source heat pump systems were excluded at the design stage because of project-specific considerations. In the following, we illustrate that the designed hybrid heat pump system also reaches the best efficiency compared to single-source options. The following three scenarios are compared (For the sake of simplicity, the PVT system (see Figure 3) is not included in these scenarios):

• ASHP-only
• GSHP-only
• Hybrid system including ASHP and GSHP

For the comparison, a simulation study of these three different supply scenarios using a physics-based modeling approach is carried out. The 15 dwelling units with in total of 61 thermal zones (see Figure 2) and the supply system (see Figure 3) are modeled using the Modelica language. The observation period is one year. Measured hourly weather data from Leuven, Belgium, covering outdoor temperatures from −10 °C to +40 °C, is used.

The performance of each of the three supply options is compared for two control approaches: a conventional rule-based control and an optimal control. The conventional rule-based controls use a proportional integral (PI) controller to modulate the heat pumps according to the tank target temperature determined by a heating curve. In the hybrid case, a PI sequence controller is used to prioritize either the ASHP (at mild ambient conditions) or the GSHP (at cold ambient conditions). The priority switches depending on the outdoor air temperature using a hysteresis controller. The floor heating valves are regulated by PI controllers connected to the zone temperature. The optimal control uses the previously described physics-based model to predict the future building and supply system behavior. A mathematical optimizer uses these predictions to adapt the heat pump modulation degree and floor heating valves to minimize (1) the total electricity consumption (including heat pumps and circulation pumps), (2) thermal discomfort in each dwelling unit, and (3) nighttime modulation of the ASHP to reduce noise levels.

Figure 4 illustrates the heat pump behavior of the three supply options (ASHP-only, GSHP-only, Hybrid) for conventional rule-based control (left column) and optimal control (right column). The selected 2-day period has a high ambient temperature swing between 0 °C and 20 °C (see Figure 4-a). The ground temperature is in the range of 8 °C to 10 °C.

With conventional rule-based control (Figure 4, left column), the heat pumps are activated at night (see Figure 4-b-i) when the zone temperatures drop below the heating setpoint (see Figure 4-c-i).[1] In the hybrid scenario, only the GSHP is used. Due to the inertia of the floor heating system, the zone temperature drops about 0.5 °C below the heating setpoint.

With optimal control, the heat pump behavior differs significantly between the three scenarios: in the ASHP-only case, the optimal controller, ‘aware’ of the higher heat pump COP at higher outdoor temperatures, solely operates the heat pump during the day (‘ASHP-only’ in Figure 4-b-ii). This ‘preheating’ behavior avoids heat pump operation at night, leading to the highest indoor temperatures among the three scenarios, up to 23 °C (see Figure 4-c-ii). In the GSHP-only case, the heat pump operation is much smoother with the highest compressor power around midnight (‘GSHP-only’ in Figure 4-b-ii). In the hybrid case (‘Hybrid’ in Figure 4-b-ii), the optimal controller selects the two heat pumps according to the source temperatures: the ASHP is operated during the day and the GSHP at night.

The maximum compressor power with a conventional rule-based control is between 7 kW and 9 kW. In all optimally controlled scenarios, the maximum compressor power used is significantly lower (between 3 kW and 5 kW). This underlines the ability of optimal control to shift operation to periods with higher source temperatures (see ASHP-only case), and to avoid peak operation with high supply temperatures and lower COP (see GSHP-only scenario). Moreover, in all optimally controlled scenarios, the heating setpoint of 21 °C is significantly better maintained than with conventional rule-based control due to the ability of optimal control to anticipate the floor heating inertia and avoid excessive temperature drops.

[1] The heat pump is activated when the storage tank temperature drops below a threshold (not depicted in Figure 4).

Figure 5 compares the total annual performance, including the annual heat generation, the electricity consumption, and the Seasonal Coefficient of Performance (SCOP, annual condenser heat / annual electricity used). The comparison highlights that the optimally controlled system yields lower heating demand (−8% to −10%) and lower electricity demand of the heat pump compressor(s) (−24% to −32%). The savings are achieved through the integrated optimal control of the floor heating valves and the operational management of the heat pumps by taking into account the Coefficient of Performance (COP) dependent on source temperatures (air and ground).

An important aspect of the comparison is how the different supply options are ranked.

With the conventional rule-based control, the ASHP-only option has the lowest SCOP of 3.9, significantly below the optimally controlled system, because the rule-based control activates the heat pump at night (see Figure 4-b). The difference in the GSHP-only option is smaller and achieved by avoiding peak operation with high supply temperatures. With optimal control, the hybrid case reaches the highest SCOP of 5.6, compared to the ASHP-only option (SCOP=5.2) and the GSHP-only option (SCOP=5.4), because the operation of both heat pumps is optimally distributed according to the source temperatures and the corresponding COP.

Conclusions

This article uses physics-based models to analyze the increased efficiency of optimally controlled hybrid heat pump systems compared to the respective single-source scenarios. For the “Stijn Streuvelstraat” case study, the hybrid system reaches the highest SCOP of 5.6 compared to the GSHP-only option (SCOP=5.4) and the ASHP-only option (SCOP=5.2). The efficiency gains stem from the ability of the optimal controller to take into account the different source temperatures and the future system behavior. This highlights the advantages of optimal control as system integrator in increasingly complex supply systems. A standard conventional rule-based control can not fully exploit this potential and achieves a lower SCOP of 4.3 for the hybrid system compared to an SCOP of 4.5 for a GSHP-only system.

Beyond the increased efficiency through optimal control, using a hybrid system enables the use of smaller individual systems. A smaller ASHP has the advantage of lower noise emissions, and smaller GSHPs, particularly a smaller borefield, reduce investment costs. Finally, having two heat sources managed by an optimal controller provides resilience facing the effects of climate change. The related study in [1] also highlights the competitive life cycle costs of hybrid systems.

Author contact information

Acknowledgements

The authors acknowledge the funding by the European Union through the SEEDS project under the Horizon Europe Programme (grant Agreement: 101138211). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Executive Agency (REA). Neither the European Union nor the granting authority can be held responsible for them.

References

[1] Walther, K., Hermans, L., Meertens, L., Helsen, L., 2025. Simplifying decision-making in the model-based co-design of building energy systems through automatically generated optimal controllers. Building Simulation 2025, Brisbane, Australia.