D5.8: Outcomes of HYPERGRYD demonstration, lessons learnt and guidelines for replication

HYPERGRYD. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 101036656 WP5 – TRL5 demonstration in living labs and virtual labs in LEC Task 5.3 Demonstration and validation actions of RES-based Enabling technologies D5.8 Outcomes of HYPERGRYD demonstration, lessons learnt and guidelines for replication Ref. Ares(2025)2636601 - 02/04/2025 DX.X Xxxxxxxx 2 DISCLAIMER The opinion stated in this report reflects the opinion of the authors and not the opinion of the European Commission. All intellectual property rights are owned by HYPERGRYD consortium members and are protected by the applicable laws. Reproduction is not authorised without prior written agreement. The commercial use of any information contained in this document may require a license from the owner of that information. ACKNOWLEDGEMENT This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement Nº 101036656. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 3 Project Project Acronym HYPERGRYD Project Title Hybrid coupled networks for thermal-electric integrated Smart Energy Districts Grant Agreement number 101036656 Call identifier H2020-LC-GD-2020 Topic identifier LC-GD-2-1-2020 Innovative land-based and offshore renewable energy technologies and their integration into the energy system Funding Scheme Research and Innovation Action Project duration 42 months (From 1 October 2021) Coordinator ARCbcn Website http://hypergryd.eu Deliverable Deliverable No. 5.8 Deliverable title Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication Description The present report is intended to present main outcomes and lessons learnt of the demonstration of Hypergryd solutions implemented in four Live-in-Labs: SONNE, ENVI, KEZO and EURAC. Outcome of the demonstration will be a set of lessons learnt to be used as guidelines for the installation and operation of RES technologies in different DHC and Smart Hybrid Grids. WP No. WP5 Related task T5.3 Demonstration and validation actions of RES-based Enabling technologies T5.4 Model demonstration and validation of the multi-carrier energy dynamic model at ENVI LIL T5.5 Model demonstration and validation of the HYPERGRYD ICT services at SONNE LIL Lead Beneficiary 17, IMP PAN - Instytut Maszyn Przeplywowych im Roberta Szewalskiego Polskiej Akademii Nauk Author(s) Michał Gliński (IMP-PAN), Paweł Zawadzki (IMP-PAN), Weronika Radziszewka (IMPPAN), Jörg Verstraete (IMP-PAN), Mauro Pipiciello (EURAC), Fouladfar Mohammad Hossein (EURAC), Manuela Binder and Martin Bruckner (SONNE) Contributor(s) Sebastian Bykuć (IMP-PAN), David Verez (ARC) Type R Dissemination PU Public Language English – GB Due 31/03/2025 Submission date 31/03/2025 D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 4 Version Date Authors Description V.0.1 7/03/2025 Michał Gliński (IMP PAN) Template for deliverable documents and task planning V.0.2 20/03/2025 Manuela Binder and Martin Bruckner (SONNE) Section regarding SONNE V0.3 25/03/2025 Michał Gliński (IMP PAN) Section regarding KEZO V.1.0 28/03/2025 Michał Gliński (IMP-PAN), W. Radziszewska (IMP-PAN), J. Verstraete (IMP-PAN) Version for internal review (missing input from EURAC) V.1.1 31/03/2025 Mauro Pipiciello (EURAC) Section regarding EURAC V.2.0 31/03/2025 Michał Gliński (IMP-PAN), Final deliverable D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 5 Table of Contents 1 Executive Summary ................................................................................................... 9 2 Introduction ............................................................................................................ 10 2.1 Scope ........................................................................................................................... 10 2.2 Audience ...................................................................................................................... 10 2.3 Abbreviations .............................................................................................................. 10 2.4 Contributions of partners ............................................................................................ 12 2.5 Structure ...................................................................................................................... 12 3 Outcomes and lessons learnt ................................................................................... 13 3.1 SONNE ......................................................................................................................... 13 3.2 KEZO ............................................................................................................................ 19 3.3 EURAC .......................................................................................................................... 26 3.3.1 Description of the LiL and objectives of demonstration .................................................... 26 3.3.2 Substation description and controls .................................................................................. 27 3.3.3 Assessment of the control performance ........................................................................... 29 3.3.4 Lessons learned .................................................................................................................. 35 3.4 ENVI ............................................................................................................................. 36 4 Guidelines for replication ........................................................................................ 40 4.1 Modular HP with PCM storage (installed in KEZO LiL) ................................................ 40 4.1.1 Purpose 40 4.1.2 Observed potential technical barriers for replication ....................................................... 41 4.1.3 Replication requirements and recommendations ............................................................. 43 4.2 Sorption Storage (installed in KEZO LiL) ...................................................................... 44 4.2.1 Purpose .............................................................................................................................. 44 4.2.2 Observed potential technical barriers for replication ....................................................... 45 4.2.3 Replication requirements and recommendations ............................................................. 46 5 Conclusions ............................................................................................................. 47 6 References .............................................................................................................. 52 D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 6 D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 7 List of Figures Figure 1 Live-in lab Sonnenplatz (Austria) ........................................................................................... 13 Figure 2 Key performance indicators, Sonnenplatz LEM study simulation scenarios and objectives, GSY simulation study for HYPERGRYD 2023-2024 ...................................................................................... 14 Figure 3 Proposed open-source software solution for centralized demand-side management ........ 15 Figure 4 KEZO Reseach Centre, view on L5 building. ........................................................................... 19 Figure 5 Simplify scheme of H&C systems in KEZO LiL ........................................................................ 20 Figure 6 The modular heat pump connected to KEZO system. ........................................................... 21 Figure 7 Sorption storage connected with Mayekawa CO2 Heat Pump in KEZO Research Centre .... 22 Figure 8 Two-day ahead prediction of the radiators and fan coils heating power in building L5 using GRU approach ...................................................................................................................................... 23 Figure 9 Temporal variation of COP heating: measurement and NN .................................................. 24 Figure 10 Regression curve of the COP heating prediction using NN ................................................. 24 Figure 11 Model-based control of the adsorption storage using the KTH edge panel ....................... 25 Figure 1. Energy Exchange Lab of Eurac Research schematic ............................................................. 26 Figure 2. Overview of the substation's operation within the context of 4th and 5th GDHN. ............... 27 Figure 3. Overview of the four different configuration schemes considered in this study. The DH network is on the left and the building on the right side of each scheme .......................................... 28 Figure 4. Schematic of MPC framework .............................................................................................. 28 Figure 5. Result of comparison the performance of MPC vs. RBC with TOU electricity price (test 1) 31 Figure 6. Distribution of HP activation durations during cheap and expensive periods under RBC and MPC with TOU electricity pricing. ........................................................................................................ 32 Figure 7. Comparison of energy stored per charging cycle and number of charging under RBC (blue) and MPC (red) with TOU electricity pricing ......................................................................................... 32 Figure 8. Result of comparison the performance of MPC vs. RBC with realistic electricity price (test 2) .............................................................................................................................................................. 33 Figure 9. Distribution of HP activation durations during cheap and expensive periods under RBC and MPC with realistic electricity pricing.................................................................................................... 34 Figure 10. Comparison of energy stored per charging cycle and number of charging under RBC (blue) and MPC (red) with realistic electricity pricing ................................................................................... 34 Figure 12 Live-in lab Environment Park (Turin-Italy) ........................................................................... 36 Figure 13 Main buildings within Environment Park ............................................................................. 37 Figure 14 Model of the medium and low voltage electric distribution system of ENVI in ENCO. Red points represent nodes with electric demands ................................................................................... 38 Figure 15 Mod Model of the thermal distribution network of Envi Park in SAInt. Nodes describe the supply or demand maximum hydraulic pressure ................................................................................ 38 D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 8 List of Tables Table 1 Technology specific KPIs panel ................................................................................................ 22 D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 9 1 Executive Summary The goal of HYPERGRYD project is the development of a set of replicable and scalable cost effective technical solutions to allow the integration of Renewable Energy Sources (RES) with different dispatchability and intrinsic variability inside Thermal Grids as well as their link with the Electrical Grids, including the development of innovative key components, in parallel with innovative and integrated ICT services formed by a scalable suite of tools for the proper handling of the increased complexity of the systems from building to Local Energy Community (LEC) levels and beyond, and accelerate the sustainable transformation, planning and modernization of District Heating and Cooling (DHC) towards 4th and 5th generation. HYPERGRYD also aims at developing real time management of both electrical and thermal energy flows in the coupled energy network complex, including the synergies between them. Therefore, HYPERGRYD aims at three over-arching General Objectives: • To prove Smart Energy Networks as the future of Efficient Energy Management in DHC in synergy with the Electrical Grids in LEC/Smart Cities of the future; • To define the roadmap to design and plan future DHC as well as the modernization of the existing ones in different climates and RES penetration levels toward 4th-5th generation, • To demonstrate HYPERGRYD RES-based Enabling Technologies, Smart Energy Grid Solutions empowered by new ICT tools and services as the key for this evolution. During the project, the HYPERGRYD’s solutions will be implemented across four Live-In-Labs cases in three representative climates, with special consideration to their cost effectiveness and potential replicability to finally achieve these three main objectives. The purpose of this deliverable in short is: i) to presents main outcomes of a demonstration of HYPERGRYD solutions for stakeholders, ii) set of lessons learnt, iii) guidelines for replication for three enabling technologies. Target audience of this report is the wide group of stakeholders from district heating sectors and H&C industry. On behalf of Authors Michał Gliński, IMP PAN D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 16 flows in order to optimize or initially plan them, focusing primarily on physical simulations. As a result, the software calculates: • the return temperature from generation assets; • the supply temperature that reaches the users; • the thermal losses across the pipes; • the pressure and temperature drops for each pipe (both on the supply and return side); • the pressure at the nodes (supply and return side); • the flow rates in the pipes and network-side heat exchangers; • the outlet temperature at the nodes on the network side. The flow rate and temperature profiles on the secondary side of the heat exchangers connected to the nodes are not taken into account. GET provides the optimization of existing or planned DHC networks as a service based on the « Exergoeconomic optimization tool for 4th and 5th generation of DHC » to optimize the operation of the DHC grid from an exergoeconomic point of view towards the 4th generation of DHC, considering investments, fuel and operation costs, and the price index, as well as flexible heat and electricity tariffs and costs for the reduction of GHG emissions. Potential use cases for this service can be: • The simulation and calculation of the integration of decentralized production systems in order to integrate RES and P2H, which is particularly interesting in the case of variable tariffs for producers. For example, low tariffs at special time slots can be used optimally for feeding heat from a heat pump into the DHC system. Variable tariffs for heat consumers are not common yet, but this could be an interesting aspect for the future. • The investigation of potential reductions of heat losses and pumping costs. • An exergy-based analysis to reduce the exergy level, the temperature levels, and the heat losses. • The comparison of different scenarios, e.g. the current situation and a scenario considering a grid extension or the integration of new consumers or of decentralized feed-in systems. • The optimization of the DHC grid based on a pressure drop in order to pinpoint bottlenecks or locate suitable positions for booster pumps. • The calculation of the effects of a grid extension to the whole DHC grid. Some results from use case 1, mainly those related to the integration of flexible heating systems into the electrical grid and dynamic pricing effects on flexible assets, were used as input. More information about the tools can be found in D3.2, D3.4, D3.6 and D3.8 of this project. The HYPERGRYD Project successfully demonstrated the potential of these advanced ICT tools to optimize and integrate renewable energy sources within thermal and electrical grids. The validation D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 17 of these tools across multiple stakeholders—including private households, municipalities, DHC operators, and electric grid operators— provided valuable insights into both technical and nontechnical aspects of the implementation of the tools and services presented. The key findings of the validation process can be found in detail in D5.6 of this project and can be summarised as follows: • Private Households/Consumers/Prosumers: Their primary interest lies in being reliably supplied with heat, hot water, and electricity at reasonable prices rather than engaging with technical details. Although the residents in Großschönau demonstrate openness to contributing to renewable energy initiatives, bolstered by a long-standing culture of sustainability, barriers such as administrative complexity and high initial investments hinder full participation in advanced energy management systems. Granting third-party access to heat pumps, PV systems, or battery inverters for coordinated energy management faced significant skepticism from local stakeholders. Concerns include potential energy unavailability and unexpected costs due to technical errors. Many residents prefer retaining full control, though voluntary recommendations instead of direct remote control could be a more acceptable alternative. Trust and transparency are crucial, as willingness to grant access depends on clear benefits. While skepticism remains, data sharing is already practiced in existing local and regional energy communities, where participants benefit from lower energy costs, reduced grid fees, and market independence. These insights highlight the importance of clear communication and user-centric implementation of energy management solutions. Also, on the subject of flexible tariffs for electricity and heat, local stakeholders remained cautious, while most international participants at the workshop found flexible tariffs relevant. Many local consumers may struggle to adapt to hourly price changes, preferring simpler day/night tariffs. However, some prosumers have already begun adjusting their behavior, using PV electricity for EV charging or appliances when available. Successful adoption of dynamic heating models would require gradual learning and targeted education efforts to increase acceptance. Thus, simplified implementation, financial incentives, and transparent communication of benefits are prerequisites for engagement. These insights emphasize the importance of aligning technical advancements with user needs and preferences for successful integration into local energy systems. • Municipalities: Municipalities such as Großschönau recognize the benefits of optimizing electricity and heat supply, integrating RES, and advancing climate goals. However, financial constraints pose significant challenges. Investments in infrastructure must balance economic returns with public priorities. Municipalities require robust financial models and technical support to scale renewable energy initiatives while addressing competing local needs. Clear policy frameworks and funding mechanisms tailored to municipal budgets could enable greater adoption of innovations. Regarding DHC, municipalities rarely act as planners or D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 18 operators of DHC systems directly in Austria. It is more common for a cooperative to be established for these purposes. • DHC Operators: DHC operators recognize the value of innovative tools for optimizing existing and planned networks or enhancing them, whereby they prefer using a service rather than using a tool and rely on personal contact with the service-provider. Key priorities include the reduction of temperature losses and the optimization of summer operations, both of which contribute to greater operational efficiency and lower energy costs. However, broader adoption of the tools requires addressing technical and data-related challenges, ensuring the cost-efficiency of solutions, and designing user-friendly interfaces. Cost-efficiency is emphasized as a critical consideration, not only for optimizing operations, minimizing infrastructure investments, and justifying network expansions, but also for incorporating decentralized feeders and technologies into existing systems and implementing dynamic pricing strategies. • Electric Grid Operators: Although the HYPERGRYD tools align with the future needs of electric grids, direct feedback from regional electric grid operators was limited due to collaboration challenges and regulatory complexities. Regional stakeholders reported grid overloads and restrictions on feed-in capacities because of the rapid expansion of PV systems over the last year. While tools like those developed by KTH require real-time electricity pricing for effective peer-to-peer utility trading, dynamic tariffs are still in their infancy in the region. Technical requirements like the widespread installation of smart meters are almost completed. Regulatory frameworks that accelerate the adoption of flexible tariffs and incentivize collaborative efforts with energy communities will be critical for unlocking the full potential of such tools. Concerning use case 2, the grid operator may not be the primary target for such tools, because it often delegates grid planning to internationally renowned companies with established expertise. These companies typically rely on their years of experience rather than using tools such as SAInt. • Researchers and engineering offices on the other hand, prefer the usage of tools over services. The most notable engineering office in the Großschönau area, which also planned the DHC network of Großschönu, is “innoVAT GmbH”. This office has many years of experience and started planning grids before specialized software was available. During a site visit, the engineer of this company could describe numerous details about the network design based purely on experience. This is why they are still refusing to use a software and continue planning grids or grid extensions only, relying on the 2D and 3D computer-aided design software application « AutoCAD » and Microsoft Excel for calculations. The validation of the SAInt tool by engineering offices would require more detailed training on the tool. Language barriers might also pose a problem, as German-language tools tend to be preferred. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 19 In summary, the validated tools can guide sustainable energy transition strategies and support modernization towards the 4th and 5th-generation of DHC systems. The integration of flexible assets, such as heat pumps and decentralized RES, combined with advanced ICT solutions, has demonstrated clear potential to improve energy efficiency, reduce costs, and enhance grid stability. Future efforts should address regulatory barriers, user education, and cost-sharing mechanisms to ensure wider adoption and long-term success. 3.2 KEZO The KEZO Live-in lab is located in KEZO Research Centre in Jabłonna near Warsaw. KEZO is one of the most modern research complexes dealing with the use of renewable energy in Poland. Constructed in 2015, it was meant to combine the function of research centre, conference centre and living laboratory. KEZO has a number of heat sources, renewable energy sources, energy storage units for both electricity and heat. Stakeholders are IMP PAN and companies working in the field of heat pump, energy storage, EV, PV, and other institutes conducting experimental research in Reseach Center. Figure 4 KEZO Reseach Centre, view on L5 building. The KEZO Research Center consists of three buildings: 2 laboratory buildings (B1, B2) and one main building L5 that houses offices, additional labs, a conference room and guest rooms. The cooling and heating system consists of four main lines that connect all buildings with heat and cold sinks, and sources placed along the way of the main lines. Each building has its own ventilation, fan coils, radiators and floor heating unit in which heat exchangers are supplied from the main heating and D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 20 cooling lines. There is no connection with the DHN of the city. KEZO possesses over 180 kWp of PV in various systems including on carports, trackers, building integrated photovoltaics, as well as a 12 kW wind turbine and over 60 m2 of solar collectors. Smart electricity meters have been installed in each laboratory to measure and log energy consumption. Within the HYPERGRYD project, many modifications were made to the KEZO energy and H&C systems for preparation to connection the new Enabling Technologies. For determination H&C characteristics of KEZO building and performance maps of existing and new technologies, additional ultrasonic heat meters and energy meters with MODBUS communication have been installed there. Two of the new Enabling Technologies developed within the framework of Task 2.2, i.e. the modular Heat Pump with PCM storage and sorption storage, have been delivered to KEZO LiL for testing. The first step was to integrate the technologies and software developed in WP2 and WP3 respectively with the existing energy system in KEZO Living Lab and with the already installed selected technologies for heat and cold generation. The works were carried out based on the detailed strategy for tests and validation developed in Task 5.1. Figure 5 Simplify scheme of H&C systems in KEZO LiL This demonstration activities have been carried out in Task 5.3, with focus on the determination of KPIs for benchmarking of the HYPERGRYD solutions and to validate developed models and control strategies. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 21 The first tested device was the modular heat pump with PCM storage for heating and DHW purpose, it was developed by Ochsner and AIT. The second device is the sorption storage developed by SORTEC in cooperation with CNR. Both devices were very advanced, but also very complicated in both technical dimension and management. The modular heat pump with PCM storage was supplied from BTES storage as the lower heat source in 5th GDHC system (Fig.6). The device was tested in two modes: DHW production and heating mode, for which the heating efficiency and COP was calculated. The characteristics of charging and discharging the PCM storage were determined. Based on the obtained results, the technologic specific KPIs were calculated, which were then compared with those values for traditional heat pumps. The calculated KPIs are worse in comparison, which is due to too a high temperature on supply side of PCM storage. For a full phase transition of RT57 in the PCM storage, there should be an inflow temperature over 60°C; such temperatures are on the border of what modular heat pump can provide with minimal efficiency. For the whole charging process of the PCM, the total COP is below 1.5. In the construction of the modular heat pump, 2 additional plate heat exchangers were used. These were meant to protect the water installation of KEZO against leaks and against seeping of the RT57 into the circuit. Unfortunately, those plate heat exchangers were causing additional resistance in transferring heat which caused worsening of the operational parameters of the modular heat pumps. For the heating mode the maximum COP was around 2. The PCM storage has an average degree of compactness of 56% and a energy storage capacity of 4.17 kWh. The energy storage efficiency was ca. 85%. Figure 6 The modular heat pump connected to KEZO system. The sorption storage was supplied with water of 80-85°C from the high temperature CO2 heat pump (Fig.7). For proper operation of this device, there has to be an additional circuit for consuming heat, i.e. the average temperature circuit MT, with a supply temperature of 30-35°C from e.g. heating system, and low temperature circuit which would take cold water at 12-16°C for cooling the buildings. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 22 The storage was tested in two modes: 1) heat storage, 2) production of heat and cold simultaneously. The energy storage capacity is 8 kWh and the maximum energy storage efficiency was about 67.3%. The time of charging the tested storage is 1 h 46 min with a charge power of 4.7 kW. A full discharge takes 1h 29 min with 5.4 kW of discharge power. The storage is based on sorption and desorption processes, so it can be used as a mid-term storage. Figure 7 Sorption storage connected with Mayekawa CO2 Heat Pump in KEZO Research Centre Technology specific KPIs for the modular heat pump with PCM storage and sorption storage were determined based on tests carried out in KEZO LiL within Task 5.3. These KPIs are listed in Table 6, the last column contains the results of the calculation based experimental data. Table 1 Technology specific KPIs panel Technology/ Tool KPI Index Symbol Results Modular heat pump COP heating KET 4.1 COPheat 1.92 COP cooling KET 4.2 COPcool NA COP DHW with PCM losses KET 4.3 COPDWH 1.44 COP DHW without PCM losses KET 4.4 COPPCM 1.69 PCM storage Degree of compactness KET 5.1 ΦPCM 56% Energy storage capacity KET 5.2 ESCPCM 4.17 kWh Energy storage density KET 5.3 ESDPCM 50.1 kWh/m3 Energy storage efficiency KET 5.4 ηPCM 85.5% DHW charging cycles per day KET 5.5 DHWcycles NA Charge/discharge time KET 5.6 tchPCM, tdisPCM 1 h 6 min 15 min Sorption storage Operating temperature levels KET 6.1 TchSOR, TdisSOR 85°C 13°C D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 23 Energy storage capacity KET 6.2 ESCSOR 8 kWh Energy storage efficiency KET 6.3 ηSOR 67.3% Charge/discharge power KET 6.4 Qchnom,SOR, Qdisnom,SOR 4.7 kW 5.4 kW Charge/discharge time KET 6.5 tchSOR, tdisSOR 1 h 46 min 1h 29 min The next demonstration activities were performed tests of the existing and the new enabling technologies in variety configuration according to defined Use Cases: • UC1. Simulation of building, hotel and tertiary loads in small districts. Using historical data collected via the KTH edge panel, prediction models were developed to estimate the heating and cooling loads of the KEZO buildings. These models enable the identification of the key parameters influencing heating and cooling demands within the research facility, while also facilitating the energy flow assessment within the local sector coupling system. In this deliverable, there will be a particular focus on the heating demand since most of the data was collected in the winter season. Below, is a two-day ahead prediction of the heating load corresponding to the radiators and fan coils in the Building L5 using Gated Recurrent Unit, which is a type of recurrent neural network (RNN). Figure 8 Two-day ahead prediction of the radiators and fan coils heating power in building L5 using GRU approach Below, the prediction performance of the NN model for heating COP prediction are displayed, along with the prediction error metrics. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 24 Figure 9 Temporal variation of COP heating: measurement and NN Figure 10 Regression curve of the COP heating prediction using NN • UC2. DHC with coordinated operation of heat pumps, sorption cooling, storages, and PV production. For this use case, the energy flexibility available within the KEZO emulated DHC system, enhanced by the newly installed hardware, will be leveraged to optimize energy flow management. In the demonstration phase, the sorption energy storage serves as the primary flexibility mechanism for regulating energy flows related to heating and cooling demands. Since this storage system is powered by a CO₂ heat pump during the testing phase, its sector coupling potential is also considered. To this end, the ability to perform load shifting and shaving through the charging and discharging cycles of the adsorption storage will be analyzed and simulated using load prediction data from UC1. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 25 Figure 11 Model-based control of the adsorption storage using the KTH edge panel • UC3. Real data-driven simulation of multi-carrier DHC. Based on data collected via the edge panel, it was possible to develop data-driven models for the CO2 heat pump and for the sorption storage. Later, those models, along with the heating/cooling demands and PV production, are used for the simulation of electric and thermal power flows within the coupled local grids at KEZO. Based on the findings from UC1 and UC2, UC3 was successfully validated by integrating load prediction, system models, and control strategies into a comprehensive data-driven simulation of the KEZO emulated sector coupling network. The primary objective of this simulation was to assess the accuracy of the developed model by analyzing prediction errors, evaluating simulation computing time— reflecting the efficiency of advanced energy management algorithms—and determining the optimal costs associated with satisfying heating and cooling demands through optimal control strategies. To achieve this, KTH conducted digital simulations of the KEZO LiL's targeted HVAC system under varying outdoor (weather) and indoor (occupant) conditions. The focus was not on simulating the KEZO emulated DHC network as a whole, but rather on evaluating the impact of optimally controlling the innovative hardware technologies developed within HYPERGRYD, particularly the sorption storage, on the energy flow within the HVAC system of Building 5 and the energy flows resulting from that. This approach enables a deeper understanding of the interaction between the optimally driven energy storage solutions and sector-coupled energy distribution. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 32 Figure 17. Distribution of HP activation durations during cheap and expensive periods under RBC and MPC with TOU electricity pricing. Figure 18. Comparison of energy stored per charging cycle and number of charging under RBC (blue) and MPC (red) with TOU electricity pricing The described MPC decisions reduce dependency on expensive periods and enhance cost savings, allowing for the same load to be satisfied as the RBC. Although the overall electricity consumption increases by 2.3 kWh (2.4%) and the system COP decreases by 0.3 (5.8%), mainly due to the HP operating at lower COP during overcharges, the operating cost (electricity) significantly reduced from 18.9 € (RBC) to 10.44 € (MPC), representing savings of about 44.79%. Test 2: Performance assessment with realistic electricity price The test 2 conditions are the same as those in test 1, with one single difference: while in test 1, the electricity price considered is fictitious, here it is a realistic one. It varies between 0.11 and 0.27 €/kWh, according to the variation in the “PUN Index GME” (short for Prezzo Unico Nazionale, or National Single Price) over three days in December 2024 (11-14). The “PUN Index GME” is the D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 33 reference index of the Italian electricity market, i.e. the reference index of electricity traded in the Day-Ahead Market (MGP). Figure 19 presents the results of Test 2, which yield similar outcomes to those of the previous tests. Here, the average electricity price (0.19 €/kWh) serves as the threshold considered in the analysis to differentiate between expensive periods (prices higher than this threshold) and cheap ones (prices lower than this threshold, grey areas of the fourth plot). The MPC strategy activates the HP more frequently during periods of low electricity prices, taking advantage of lower costs. During these times, MPC extensively uses the HP to maximize TES charging. In contrast, RBC shows the highest levels of HP operation during peak electricity price periods, closely following the load pattern. This difference is due to MPC’s ability to predict future trends, including electricity prices and load peaks, enabling it to make more informed and optimal decisions. An example is between 2:00 and 4:00 on day 1, when the electricity price was low, the MPC strategy predicted an upcoming thermal load peak in the early morning. Consequently, the tank setpoint was increased to its maximum value, allowing the system to supercharge the tank in anticipation of the demand. Around midday, the electricity price dropped again, prompting the MPC to charge the tank further. However, it is worth noting that the tank's thermal capacity also played a significant role in the decision-making process. For instance, the MPC activated the HP at 18:00 on the first day despite the high electricity price due to a thermal load peak and insufficient state of charge in the tank to meet the demand. Figure 19. Result of comparison the performance of MPC vs. RBC with realistic electricity price (test 2) D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 34 The total durations of HP activation were similar in the two cases, with 1199 minutes for RBC and 1207 minutes for MPC. Figure 20 shows that under RBC, the HP activations were almost evenly distributed between cheap (48.5%) and expensive (51.5%) periods, indicating no responsiveness to price variations. The system operated solely based on thermal demand without considering electricity prices. In contrast, the MPC with a variable pricing strategy significantly shifted operations toward cheaper periods. Approximately 67.19% of the HP activations occurred during low-cost periods, while 32.81% occurred during high-cost times. Figure 21 gives an insight into the number of charges and the energy stored per charge. As in the previous test, the number of charges increased even more, reaching 37 shorter and more frequent activation periods. It should be noted that no explicit term was included in the objective function to penalize or control the number of HP activations, which likely contributed to this behaviour. This approach enables the heat pump to operate more flexibly, adapting effectively to the variable conditions of the electricity price and load by storing energy in smaller increments, mostly between 2 and 10 kWh. Overall, the MPC strategy resulted in a 6.86% reduction in operating costs, from 18.07 € to 16.83 €. Figure 20. Distribution of HP activation durations during cheap and expensive periods under RBC and MPC with realistic electricity pricing. Figure 21. Comparison of energy stored per charging cycle and number of charging under RBC (blue) and MPC (red) with realistic electricity pricing D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 35 3.3.4 Lessons learnt MPC consistently outperformed RBC by lowering electricity costs and achieving notable savings. The analysis shows that MPC effectively shifted TES charging to periods with lower prices while meeting load demands. However, the increased activation cycles suggest a dynamic response to changing conditions. To improve this, it may be helpful to add a term to the objective function to reduce these cycles. The most critical insights derived from the research activity on the described DH substation and advanced controls are: • MPC effectively utilizes dynamic pricing and load shifting by scheduling heat pump operations during periods when electricity costs are lower and optimizing TES charging strategies. This approach not only leverages dynamic electricity pricing to reduce operating costs but also shifts energy consumption to off-peak times. As a result, it enhances overall system efficiency and improves demand-side management. • Continuous model training ensures systems adapt to changing conditions and remain accurate, maintaining optimal performance over time. • Accurate thermal load forecasting is critical for effective MPC. Employing ANN to develop a ROM provides precise and adaptable forecasts, enhancing system responsiveness. • Efficiency improvements through various substation schemes: By exploring different substation schemes, it is possible to enhance the system's efficiency. Implementing temperature-based transitions and combining hysteresis methods facilitated smoother and more effective changes between operational schemes, ultimately leading to improved substation efficiency. Although this study has demonstrated significant benefits from using MPC in decentralized substations, there is still potential for further improvement. The following are suggestions for enhancing the system and directions for future research based on our findings: • Longevity constraints for equipment: While MPC reduced electricity costs, it may also result in a higher HP switching rate, which can lead to increased wear on equipment over time. This could be incorporated as a constraint in the MPC's decision-making process. • Explore advanced machine learning techniques for model training, including reinforcement learning, to enhance model adaptability and prediction accuracy across various conditions. • Optimize the sizing and operation of TES: Analyse the effects of various TES sizes and configurations on system performance to enhance TES design for specific applications and improve energy storage utilization. • Integrate renewable energy sources: Investigate the integration of renewable energy sources, including solar thermal and photovoltaic (PV) panels, along with battery storage, into the MPC framework to enhance sustainability and reduce dependence on grid electricity. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 36 • Expand the study to incorporate cooling applications: This research primarily concentrated on heating applications. Future investigations could broaden the focus to include cooling operations within the district heating network. Implementing model predictive control (MPC) in cooling systems could enhance performance and efficiency throughout the year, offering comprehensive solutions for both heating and cooling requirements. • Expand the analysis to encompass the entire heating season and consider various climates: The study focused on three typical winter days for a specific climate to experience the MPC operation and its advantages. By extending the analysis to the entire heating season and considering various climates, it is possible to investigate the operation of the MPC under different load conditions and achieve seasonal cost savings. • Expand the study by considering variable DH conditions: the MPC resulted in the effective utilisation of electricity dynamic pricing and load shifting in cases of constant conditions at the DHN side. However, both DH thermal energy costs and temperatures can vary throughout the day and season, which affects the overall system efficiency and operating costs. Future investigations could consider these aspects and their correlation with various substation operating schemes by incorporating them into the MPC decision-making process. 3.4 ENVI The ENVI Live-In Lab, located at Environment Park in Turin, serves as an innovation hub dedicated to ecological transition and sustainable development. Environment Park is a technology park in Turin dedicated to sustainable innovation and ecological transition. It hosts companies, research centers, and laboratories focused on renewable energy, energy efficiency, and sustainable mobility. As part of the HYPERGRYD project, its energy monitoring infrastructure supports data collection on electricity, heat, and water consumption for ICT tool development and validation. Figure 22 Live-in lab Environment Park (Turin-Italy) D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 37 In particular, the complex consists of 10 buildings, including 5 office buildings, 4 laboratories, and a canteen, all interconnected by an internal distribution network for electricity, heating, and cooling. Figure 23 Main buildings within Environment Park Within the HYPERGRYD project, the lab has been utilized to validate a multi-carrier energy model aimed at integrating thermal and electrical networks with renewable energy sources. The system includes a District Heating and Cooling (DHC) network, heat pumps, photovoltaic installations, a hydropower system, and all crucial components for testing the effectiveness of the developed solutions. This demonstration is part of Task 5.4, which focuses on validating the multi-carrier energy dynamic model at ENVI. The methodological approach was structured around several key activities to ensure a robust validation process. Task 5.4 played a central role in coordinating the demonstration, while Task 3.5 contributed by developing the modeling tool applied in the ENVI LiL. The process began with real-time data acquisition from the SCADA system, allowing continuous monitoring of energy system performance. The integration of a digital twin, developed using ENCO’s SAInt tool, enabled the simulation of various operational scenarios, providing detailed insights into the interactions between thermal, electrical, and renewable energy sources. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 38 Figure 24 Model of the medium and low voltage electric distribution system of ENVI in ENCO. Red points represent nodes with electric demands Figure 25 Mod Model of the thermal distribution network of Envi Park in SAInt. Nodes describe the supply or demand maximum hydraulic pressure Active stakeholder involvement, including feedback from ENVI network operators and technology providers, was essential for refining the model’s usability and adapting it to real-world conditions. The validation process involved multiple network configurations, stress-testing different operational strategies under varying energy demand and supply scenarios. The feedback loop with WP4 - D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 39 HYPERGRYD Digital Twin Platform as a Service ensured continuous improvements to the modeling tool. The validation at ENVI LiL focused on three main use cases, each addressing critical challenges in multi-carrier energy network management: • Islanded Mode Operation: this scenario tested the ability of ENVI to function autonomously, minimizing reliance on the external grid. The results demonstrated that a combination of hydropower and photovoltaics significantly contributed to self-sufficiency while ensuring grid stability. • Transition to 4th-generation DHC network: the implementation of reversible heat pumps and optimized thermal distribution strategies resulted in increased efficiency and a reduction in energy losses. Lower supply temperatures improved overall system performance. • Integration of multiple energy carriers: the demonstration evaluated the combined operation of CHP and thermal storage, enhancing flexibility and enabling more effective load balancing between electricity and heat networks. The results confirmed that a multi-carrier approach allows for better demand-side management and grid optimization. The demonstration at the ENVI Live-In Lab successfully validated multi-carrier energy models, confirming their potential for seamless integration of renewable energy sources with thermal and electrical networks. The findings underscore the importance of advanced ICT tools, particularly digital twins, in optimizing hybrid energy grids. Collaboration among stakeholders, driven by the structured activities of Task 5.4 and Task 3.5, proved essential in ensuring technical viability, economic feasibility, and overall system efficiency in modern Smart Hybrid Grids. Additionally, as emphasized in deliverable D5.5, the availability and utilization of accurate data play a decisive role in enhancing system performance, making real-time data management an indispensable element for future energy communities. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 40 4 Guidelines for replication 4.1 Modular HP with PCM storage (installed in KEZO LiL) 4.1.1 Purpose The PCM-integrated heat pump system is a demonstrator for a new emerging technology aimed at small houses and apartments. The studies have demonstrated that coupling heat pumps with PCM storage can lead to significant energy savings and improved system performance, our research, for instance, indicates that integrating PCM storage with heat pumps can enhance the Seasonal Coefficient of Performance (SCOP) by up to 13%, depending on the building's heating demand. The technology demonstrator shows the technology is feasible and beneficial at the scale of households, with the following advantages and disadvantages: Advantages • The modular heat pump is designed to regulate power and heating capacity in a wide range, made possible thanks to the modular aspect which allows to disconnect individual heat pump modules. • Regulation of efficiency is facilitated by the possibility to switch each compressor on and off, which eliminates the need to use expensive solutions, such as e.g. frequency inverters. • Short-term heat storage based on phase change material RT57 (PCM) can store the heat for DHW, allowing it to be used for peak-shaving and peak-shifting when there is high heat water demand, e.g. in the evenings. • Combining the modular heat pump with PCM storage gives more possibilities to physically fit the device in a limited space thanks to its compact format: with a base of 70 by 70 cm, the device has a height of 2.3m. In this configuration the requirement is a room of height min. 2.5m. • Electrically, the modular heat pump uses 3 phases and is capable of equally loading phases. • PCM storage is smaller than water TTES storage, so its main customer group are flats, small houses and offices, which have their own heating system and DHW production. • Increased reliability and redundancy: parallel connection of the HP modules allows each module to work independently. If one unit would get damaged the others can work and ensure continuous heating or cooling. • Scalability and flexibility: Adding additional modules in this parallel configuration is relatively easy, which allows to adjust the system to changing heating or cooling requirements without the need to do big changes to the infrastructure. • Increased efficiency during partial loads: Working with many smaller units might be more efficient in case of partial load. The units might switch on or off depending on the needs, maintaining optimal efficiency and reducing the energy usage. Drawbacks and barriers • Increased initial investment: the purchase of the system is more costly considering the equipment and installation. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 41 • Complexity of system management: coordinating the operation of multiple units needs advanced management systems to ensure equalised distribution of loads and prevent short cycles. • Spatial requirement: each additional module takes some physical space. While each unit is rather compact, locations with very limited space may restrict the number of modules. In addition, while compact, the entire construction is still rather bulky and potentially problematic to physically fit through doors, etc. This could be mitigated, through e.g. on-site assembly, but the economic feasibility of this would need to be verified. • Regulation issues and safety: R290 (propane) is classified as a flammable cooling agent, which requires following strict safety norms. For example, in outdoor uses the volume of the R290 agent is limited to 152 grams per 1 cooling circuit to minimize the danger of combustion. • To summarise, the modular heat pumps in parallel configuration have many advantages, such as increased resilience, scalability and efficiency, but they also have some challenges like higher initial costs, increased complexity of the system, space limitation and elevated safety requirements. Careful planning and design are required to maximise the advantages, while at the same time reducing the problems. 4.1.2 Observed potential technical barriers for replication During the testing phases of the demonstrator, several technical aspects and issues, some of which need to be considered while others may be a barrier for replication and need specific attention for practical implementations, were observed. These observed aspects and issues are • PCM material and its characteristics. An appropriate PCM material should be selected depending on the heat pump and recipient parameters, because the RT 57 used here has a phase transition temperature of approx. 57 °C, which defines that the circuit has to be supplied with water of 6265 °C. This is too high as a parameter for a modular heat pump, causing ineffective condenser cooling and thus high condensation pressure. This leads to overheating of the refrigeration compressors. Too high condensation pressure significantly reduces the efficiency of the heat pump COP. • Heat exchanger in the PCM storage. It is necessary to select and design a suitable heat exchanger in the PCM storage made of a material other than aluminium as the aluminium plate heat exchanger poses challenges at the stage of production and operation. The process of welding plates in the heat exchanger is time-consuming, expensive, and requires a good specialist in this field. Thick aluminium sheets were used, which cause a significant increase in the weight of aluminium and a limitation/reduction of the amount of PCM material in the exchanger. During operation, the aluminium plates may also leak and PCM may leak into the water. The compatibility of the PCM material with the materials used in the heat exchanger is also important. The exchanger must be cheap at the production stage, durable during operation and compatible with the PCM material. • Monitoring the temperature at the supply to the PCM storage. The PCM storage must maintain an appropriate temperature range to operate properly. If the system overheats or operates outside the desired temperature range, the PCM may not change phase as intended. It is D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 48 Energy Communities (LECs), which supports the modernization of district heating systems to 4th and 5th GDHC. Demonstration and validation of the ICT tools for the management of heat and energy networks in the local energy communities were carried out at the Live-in-Lab SONNE in Austria. The main objectives were to optimize energy flows and integrate renewable energy sources (RES) into integrated heat and electricity networks. The key points and benefits resulting from the implementation of ICT tools and the analyses performed are as follows: • The use of ICT tools such as BIM-GIS, exergoeconomic tools, AI, and edge-computing algorithms enabled accurate modeling of the DHC network and RES. Thanks to simulations using ICT tools, real-time data from PV systems, and energy prices, peer-to-peer (P2P) trading tests were conducted. These showed significant benefits for local market participants. • The simulation of the local energy market showed that the activation of P2P trading contributed to improving the electricity and heat costs by 145% and also increased selfsufficiency (up to 124% in summer due to PV production). The use of intelligent trading strategies, such as virtual heat pumps, yielded reduced costs and significantly improved selfsufficiency rates. • The use of ICT tools enabled the analysis of various scenarios in electricity, gas and heat networks, which contribute to the reduction of heat losses, the costs of using devices and facilitates the integration of distributed RES systems. • Continuous feedback from project participants, including households, municipalities, DHC operators and electricity grid operators, has allowed for the improvement of HYPERGRYD tools and services. Collaboration with different stakeholders has revealed important technical and non-technical aspects related to the implementation of solutions. • The results indicate that, despite the interest in sustainable initiatives, there are barriers such as administrative complexity and high upfront costs that can hinder full participation in advanced energy management systems. Understanding user needs, simplifying implementation and transparent communication of benefits are key to engaging local communities. • Addressing regulatory barriers, user education and cost-sharing mechanisms are essential to ensure wider adoption and long-term success of innovative energy solutions. The conclusions of work carried out within Task 5.6 at ENVI Live-In Lab in the Turin Environmental Park underline the importance of innovation in sustainable development and ecological transition. The project validated the effectiveness of energy models and tools based on multiple energy carriers, which enable the integration of renewable energy sources into thermal and electrical networks. The analysis carried out at ENVI LiL provided valuable information on three key operational scenarios: autonomous system operation, transition to a fourth generation heating network and integration of different energy carriers. The results showed that the appropriate combination and mix of PV and hydropower with energy storage contributes to the system’s self-sufficiency, while the use of heat pumps and heat distribution optimization strategies increases efficiency and reduces energy losses. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 49 Within the HYPERGRYD project, a Digital Twin Platform was implemented for SONNE and ENVI LiL. It was developed by IDP to improve the usability and stability of different tools developed by GSY, KTH, GET and ENCO. The platform is a digital twin of real heating and electrical grid with historical data, real-time data and continuous feedback for simulation of the LEC. Demonstration activities were carried out within in Task 5.3 at KEZO Live-in-Lab in Jabłonna in Poland, focused on defining KPIs for benchmarking new technologies and validating developed models and control strategies. The tested technologies, including a modular heat pump with PCM storage, and sorption storage, showed many possible applications. The main conclusions from these studies are as follows: 1. Modular heat pump with PCM storage: • The water source heat pump (WSHP) should have the borehole storage (BTES) as the lower heat source. Such a solution is proposed in the new generation 5th GDHC systems. • The tests of modular HP showed a low efficiency due to: i) additional heat exchangers causing heat losses; ii) high inlet temperature at the PCM storage (above 60°C) which is at the maximum temperature achievable with the heat pump with R290; the maximum value of COP achieved during heating mode was approximately 2.0, and the total COP during charging mode of PCM obtained below 1.5. • It was shown that the PCM storage is charged with relatively low power and for a long period of time of approx. 1 hour, but it has the ability to quickly discharge to provide DHW at up to 15 l/s. The maximum energy storage efficiency in the PCM reached around 85.5%. 2. Sorption storage: • The sorption storage must be supplied from a high-temperature source HT with a temperature around 70-90°C (in the KEZO case the supplied temperature was 80-85°C. It is also necessary to provide the possibility for the discharge of medium-temperature heat MT, e.g. to the building heating system with fan coils or floor heating, and ensure a feed to the low-temperature system LT, e.g. in order to cool the facility. This is because the HT and MT circuits must operate during the charging of the storage (when the sorption process occurs), while during discharging phase it is necessary to operate the MT and LT circuits (desorption process). • The sorption storage has 2 operating modes: i) as a thermal storage, ii) simultaneous as a heat storage and heat and cold production. • The storage capacity was 8 kWh, and the maximum energy storage efficiency reached 67.3%. It was shown that the time of fully charging the sorption storage is approx. 1.5 hours with a power of approx. 5kW; similar parameters were obtained during the discharging phase. • The ability to store thermal energy both short-term and long-term was demonstrated. ICT tools to manage and optimize the operation of the modular heat pump of the PCM storage and the sorption storage enables greater flexibility in energy and heat management. Such actions significantly contribute to reducing CO2 emissions and increasing energy efficiency, which supports sustainable development goals. Test results and obtained KPIs indicate areas for further optimization, which may lead to improving the technologies and increasing their efficiency in the future. In D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 50 summary, the demonstration activity conducted at KEZO LiL provided valuable information on the performance of modern solutions in heating and cooling systems, highlighting both their potential and the challenges that need to be overcome to maximise the benefits of their application in integrated energy systems. Based on the analyses carried out at the Eurac Live-in-Lab in Bolzano, Italy, within the framework of Task 5.3, key information on the performance and potential improvements of Model Predictive Control (MPC) for decentralized District Heating (DH) substations was provided. The main conclusions drawn from the analysis are as follows: • MPC outperformed the conventional Rule-Based Control (RBC) approach by reducing electricity costs and achieving significant savings. This was achieved by shifting Thermal Energy Storage (TES) charges to periods of lower electricity prices while still meeting power demand. MPC effectively uses dynamic pricing and load shifting to optimize the operation of heat pumps and TES. This reduces operating costs and shifts energy consumption to off-peak periods, improving overall system efficiency and streamlining demand-side management. • Continuous substation component’s model training ensures the system adapts to changing conditions and remains accurate, maintaining optimal performance over time. Additionally, accurate thermal load forecasting is crucial for effective MPC operation. Using artificial neural networks (ANN) to develop a reduced-order model (ROM) enables accurate and adaptive forecasts, thereby improving system responsiveness and efficiency. • Research suggests that exploring different substation schemes, such as temperature-based transitions and the use of hysteresis methods, improves system efficiency by facilitating smoother operational transitions. These improvements contribute to better substation efficiency and reduced operating costs. • There is a need to optimize the size and operation of thermal storage to improve the system efficiency and performance. Moreover, integrating renewable energy sources with thermal and electrical storage in MPC can increase sustainability and reduce the dependence on grid electricity. In summary, the study highlights the significant benefits of MPC in DH systems, especially in terms of cost reduction, efficiency and adaptability. However, there are several areas for improvement, including equipment durability, renewable energy integration, and expansion into cooling applications, suggesting that future research should focus on these aspects to further enhance system performance. The HYPERGRYD project has successfully demonstrated the potential of advanced ICT tools for the optimization and integration of renewable energy sources in heat and electricity networks. Validation of these tools contributed to a better understanding of the challenges and opportunities related to the implementation of modern energy solutions. The involvement of stakeholders and continuous exchange of information played a key role in improving the models and adapting them to real conditions. Collaboration between project teams, as well as the use of advanced ICT tools such as digital twins, were essential to achieve technical feasibility and economic viability of the solutions. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 51 The conclusions from this project provide a solid basis for future actions aimed at transforming energy systems towards greater efficiency, cost reduction and improved grid stability. D5.8 Outcomes of HYPEGRYD demonstration, lessons learnt and guidelines for replication 52 6 References Grid Singularity (2024): Leveraging heat pump flexibility in local energy markets. https://gridsingularity.medium.com/leveraging-heat-pump-flexibility-in-local-energy-marketsac20874345af (Accessed on 23.11.2024).