Future-Proofing Buildings for Heat Pump-Integrated 5th Generation District Heating and Cooling: How Building Renovation and Optimal Control Shape Network Performance 

Anna Dell’Isola, Lieve Helsen, Belgium

5^th Generation District Heating and Cooling (5GDHC) networks offer a promising solution to increase the share of residual and renewable energy sources (RES). However, their efficiency depends on both building renovation levels and smart control strategies. This article presents insights from a scenario-based simulation study that examines how varying renovation levels affect system performance. The study demonstrates how Optimal Control (OC) can enhance heat pump operation, reduce peak demand, and support the viability of low-temperature networks for future-proof building clusters.

Introduction

The transition to climate neutrality by 2050 requires a fundamental rethinking of how we heat and cool our urban environments, considering both buildings and entire neighbourhoods. 5^th Generation District Heating and Cooling (5GDHC) networks are gaining increased interest for their ability to provide simultaneous heating and cooling through a bidirectional network. By operating near ambient temperatures (T_s < 30°C), these systems can integrate large shares of renewable and residual energy sources (R²ES), reducing primary energy use and carbon dioxide emissions.

However, the move to low supply temperatures and 5GDHC networks requires the connected buildings to be either highly energy-efficient or supported by advanced, integrated control strategies that can compensate for varying building performance. Renovation of existing buildings with poor energy performance is essential, though deep renovation is expensive and time-consuming, and perhaps not always needed. Meanwhile, managing and controlling multiple distributed heat and cold sources in a 5GDHC network presents significant challenges, with pumping power becoming a critical factor at low temperatures.

Given the importance of lowering building heat supply temperatures, the need for integrated optimal control strategies, and the cost barrier of large-scale renovation, this article presents insights from a simulation study showing how Optimal Control (OC) can act as a system integrator for a small virtual 5GDHC network with varying building renovation scenarios. The content of this article is based on the research work presented in [1].

Thermal Network and Buildings Description

A small virtual 5GDHC system serves as a use case with a Bidirectional energy – Directional medium flow (BiD) configuration, schematically shown in Figure 1. The system consists of three terraced houses and three office buildings, as representative of a mixed small Belgian district, and a central balancing unit. A centralized pump controls the direction of the flow in the network, while a three-way valve at the primary side of the substation modulates the medium flow rate in the substations towards a modulating water-source heat pump (WSHP) or a heat exchanger (HEx) for direct cooling, depending on whether the heating or cooling mode is selected. To allow heating/cooling in each zone of the building, an embedded floor emission system, controlled by valves, is used. A central balancing unit manages the heat imbalances within the network and keeps the network temperature within a specified range (5-16°C) by employing a large modulating air-source heat pump (ASHP) and a large modulating air-source chiller (ASCH). A water buffer tank is added to dampen the peaks in the network.

Ten renovation scenarios, from highly insulated buildings with mechanical ventilation with heat recovery (MVHR) to poorly insulated buildings with simple mechanical exhaust ventilation (MEV), and mixed-renovation scenarios, are analysed. Three different insulation levels (labels A, B, and C) are considered, according to the Belgian classification (Table 1). A summary of the considered renovation scenarios is presented in Table 2, highlighting the main parameters varying among the buildings connected to the district.

Table 1: U-values and air tightness values for insulation label A, B, C

Table 2: Summary of considered renovation scenarios (Label, Ventilation, Q design for Houses and Offices)

Optimal Control

Optimal Control (OC) is used as a system integrator to minimise energy use, while providing thermal comfort. The employed optimal controller is an idealised model predictive controller that uses a detailed white-box system model (without model mismatch) and perfect knowledge of the weather conditions. Physics-based models of two-zones buildings, thermal systems (e.g., heat pumps, chillers), and hydraulic components are developed in Modelica and integrated into a single nonlinear programming (NLP) optimal control problem (OCP) in TACO (Toolchain for Automated Control and Optimisation), an in-house developed Modelica-based toolchain for nonlinear white-box Model Predictive Control (MPC) [2].

The main objective of the optimisation problem is to minimise (1) the total primary energy use, including all heat pumps, chillers, circulation pumps, and centralized pumps, (2) thermal discomfort inside the buildings, under specific constraints (e.g., minimum and maximum indoor temperature, floor surface temperature, network temperature). A more extensive description of the approach used to solve the NLP optimal control problem for a BiD configuration is presented by Dell’Isola et al. [3].

Optimal control problems are carried out for three representative months in winter, mid-season, and summer (January, April, and July).  

To assess the renovation level impact, the different scenarios are compared, focusing on the electrical energy use, peak power, and seasonal performance factor (SPF), evaluated as the total useful heat provided in heating regime or extracted in cooling regime in all buildings, versus the total heat pump electric energy use over the simulation period.

5GDHC Network Performance Influenced by Renovation Levels

Figures 2 – 3 – 4 present the main results for different renovation scenarios for the three representative months. In all monthly simulations, the thermal discomfort remains very low, i.e., below 2.5 Kh/zone. Therefore, the different renovation scenarios are compared under equal thermal comfort conditions.

The study confirms that the renovation level strongly affects the network performance, especially in winter. Well-renovated buildings, with a higher insulation level and a better-performing ventilation system, require a lower heat demand, leading to lower primary energy use as expected (Figure 2). However, it is interesting to notice that even with medium or mixed renovation levels, the optimal control strategy helps maximising system efficiency, offsetting the higher demand with higher condenser heat pump power and comparable coefficient of performance (COP). This results in a relative SPF variation among the different scenarios of only 1.7%, except Scenario 4, with all label C buildings, which results in higher energy use, higher supply temperatures, and limited thermal interaction between the buildings.

Similar observations can be made for the transitional season in April in Figure 3, showing a similar ranking for the impact of the renovation levels.

Figure 4 shows interesting results for the summer period. The inclusion of buildings with a lower renovation level influences the system’s optimal operation, leading to heat pump activation even in the cooling season. Optimal control exploits the buildings’ thermal inertia and the diversity in thermal loads between houses and offices to optimise thermal interaction.

Figure 5 shows how lower-performing residential buildings are used as heat sinks (still fulfilling thermal comfort constraints) during office cooling peaks, reaching the lowest total electric energy use for scenario 2. The decentralized heat pumps absorb heat from the network, limiting the need to operate the energy-intensive central chiller. On the other hand, the unnecessary activation of the heat pump for the building’s owner is penalised in the SPF calculation.

Heat pumps and chiller capacity

Figure 6 illustrates how central/decentralized heat pumps (HP) and central chiller (CH) capacities vary across different district renovation scenarios.[1] The results highlight how building performance shapes both peak demand and the optimal operation strategy for the district network by minimising total electrical energy use and thermal discomfort.

In worse-performing districts, buildings with lower insulation and thermal inertia require more continuous heating to ensure thermal comfort. This leads to larger decentralized HP capacities at the building level. The central HP capacities may either increase (e.g., scenario one vs. 1_2) or remain about the same (e.g., scenario two vs. 2_3), with the central HP operating more continuously but at a similar capacity. However, this comes at the cost of higher total energy use, as the heat pumps typically run for longer periods in worse-performing districts. On the other hand, the central chiller capacity always rises as district performance decreases (higher cooling demands), in order to provide direct cooling without an additional decentralized chiller.

[1] For Scenario 1, 2 and 2_1 the second highest peak power was considered. The highest peak power of the central HP has been filtered out, since it is a residual occurring just once at the end of the simulation period.mp Systems Under Dynamic Operating Conditions“, RWTH Aachen University, Aachen, Germany, 2022. doi: 10.18154/RWTH-2022-06640.

In better-performing districts, high insulation and good air tightness give buildings larger thermal inertia. Optimal control of these buildings results in less constant heating demand, with more distinct peaks occurring when heat pumps operate at their highest COP. The decentralized HP capacities are lower due to the lower demand (lower buildings heat losses), while the central HP capacity varies depending on the mix of building types (similarly to the considerations above). Regardless, total energy use is always lower in better-performing districts. Chiller capacity also consistently decreases with better renovation levels, thanks to reduced cooling needs.

The differences between scenarios with Label B and Label C are smaller than the jump to scenarios with Label A buildings. This is because both B and C use the same MEV system, while label A benefits from mechanical ventilation with heat recovery, significantly improving energy performance and reducing system loads.

In scenario 4, a lower renovation level for both the residential and office buildings makes the heat pumps and chiller sizes clearly higher compared to the other scenarios.

Conclusions

This study shows that 5GDHC networks can achieve strong performance even in districts with mixed building renovation levels, if equipped with an integrated optimal control strategy. The summer heat pump activation strategy illustrates how advanced control can transform what appears to be a disadvantage, such as poorly performing buildings, into a system-level asset, thereby helping to balance the network and to reduce overall energy use.

For the extreme renovation scenarios, the results align with existing literature: well-renovated buildings (Scenario 1) deliver high efficiency and good thermal comfort, while poorly-renovated districts (Scenario 4) are unsuitable for 5GDHC due to higher loads and reduced flexibility. In such cases, where the system is no longer self-balanced (as may also occur in offices-only or dwellings-only dominated districts), a 4^th Generation District Heating (4GDH) network may be more appropriate. Overall, the findings demonstrate that optimal control is essential to unlock the full potential of low-temperature district heating and cooling systems.

Future work will include full-year simulations, a sensitivity analysis on geographical location using different weather datasets, and an investigation of smart ventilation control. In addition, the observed trends in peak power requirements for heat pumps and chillers across scenarios highlight the need for further research into integrated optimal sizing and control strategies in 5GDHC networks.

Acknowledgement

The authors acknowledge the funding by the European Union through HeriTACE project (101138672) under the Horizon Europe Programme. Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or HADEA. Neither the European Union nor the granting authority can be held responsible for them.

Author contact information

 References

[1]. A. Dell’Isola, K. Walther, L. Helsen, K. Leuven, and T. Park, “Building renovation level requirements for a 5th Generation District Heating and Cooling Network,” in Proceedings of Building Simulation 2025, Brisbane, Australia, 2025.

[2]. F. Jorissen, W. Boydens, and L. Helsen, “TACO, an automated toolchain for model predictive control of building systems: implementation and verification,” Journal of Building Performance Simulation, vol. 12, no. 2, pp. 180–192, Mar. 2019, doi: 10.1080/19401493.2018.1498537.

[3]. A. Dell’Isola, L. Hermans, and L. Helsen, “Optimisation of 5th Generation District Heating and Cooling Networks for different Flow Configurations,” Journal of Sustainable Development of Energy, Water and Environment Systems, vol. 13, no. 2, pp. 1–19, Jun. 2025, doi: https://doi.org/10.13044/j.sdewes.d13.0577.