In-depth case studies of flexibility initiatives
Photo: Unsplash.com
In this chapter, we present the 10 use cases. For each of them, we provide a case description, show the processes and actors involved in the value chain, and describe barriers and suggested ways to overcome the barriers. We also provide a rough estimate of the overall flexibility potential of the flexibility source in focus. Before we go into the specific use cases, we present the main learnings from the 10 cases and explain the value chain for flexibility.
Selected case studies
Several criteria were used when selecting cases to be studied in this project. They should focus on flexibility from a range of smart appliances like smart appliances with storage water heaters (continuous or night storage), thermal appliances (heating and cooling), electric radiators, electric boilers, electric and hybrid heat pumps, and air conditioners. The cases should have reached a certain level of maturity, i.e., how long they have been implemented. Newer cases were also preferred. The design choices in new solutions may be interesting even if the mechanism is not mature, and new mechanisms were expected to draw upon learnings from earlier solutions. It was also desirable to study innovative cases to study mechanisms utilising smart appliances in novel ways.
In addition to the criteria, the selection should also have a certain variation across:
- Different products/DSO needs
- Different flexibility sources/load types
- Different providers (i.e., contractual arrangements, aggregation solutions)
10 flexibility initiatives were selected across the four countries. The cases are presented in the table below, highlighting previous work they build upon, their purpose and focus, the actors and loads participating, and plans going forward. Two of the cases are permanent solutions, while the rest are temporary pilots and demonstrations.
Table 6: Overview of selected case studies
Builds on | Purpose and focus | Participants and loads | Plans | |
Byggfleks (2019-2023) |
|
|
|
|
Battflex (2019-2024) |
|
|
|
|
Norflex (2019-2023) |
|
|
|
|
Sthlmflex (2020-2023) |
|
|
| |
Effekthandel Väst (permanent) |
|
|
|
|
E.ON Energidistribution’s flexibility markets (permanent from 2023/ 2024) |
|
|
| |
FUSE (2020-2023) |
|
|
| |
EcoGrid 2.0 (2016-2019) |
|
|
| |
Helen Electricity Network and Fingrid (2025-2027) |
|
|
|
|
DSO Elenia (2020à) |
|
|
|
|
Main learnings from the 10 case studies
Norwegian DSO Lede: Byggfleks - Validating DSF technology for grid operation (2019-2023)
The pilot successfully tested a technical solution and systems for the activation and aggregation of flexibility from commercial buildings. Lede found that activation limited to 1-2 hours did not affect user comfort.
Feedback from building managers suggests that complicated control systems should be avoided, and if not, guidance and knowledge-building should be prioritised. Lede also experiences challenges related to variations in appliance smartness and interoperability. Older appliances had a limited ability to integrate and communicate with top control systems. However, once an integration was set up, the logic could be reused to integrate new assets.
Norwegian DSO Eidsiva (now Elvia): Battflex - Voltage control in weak grids (2019-2024)
Smart ESWHs were installed in a selection of households to assess how market-based optimisation impacts grid voltage levels and whether smart automation can mitigate potential grid voltage issues in local grids. The pilot found that pure price optimisation could lead to rebound effects. Controlling the ESWHs based on a combination of voltage control and price optimisation, on the other hand, led to reduced electricity costs without affecting the voltage level negatively. Only implicit flexibility was tested (optimising the water heaters based on Nord Pool spot prices).
No technical barriers were reported by the interviewees. This may be related to the flexibility provider (OSO) was also the supplier of the smart appliances and the technical system for activating flexibility. The flexibility provider only had to integrate with a single type of appliance, for which they had designed the interface themselves.
Norflex (Euroflex) - Large-scale demonstration project – flexibility market (2019-2023, 2023-2026)
Several important issues have been identified in Norflex, where possible solutions are being tested in Euroflex. One example is baseline estimation. The flexibility providers experienced difficulties estimating the baseline for the different smaller flexibility sources, especially for household appliances. As a response to this, a new product is being developed and tested in Euroflex, called MaxUsage. This product lets the flexibility providers bid on a total maximum capacity being used for their portfolio as a whole.
Another example is coordination and value stacking (letting flexible resources participate in several markets). Participants experienced coordination as essential to ensure efficient usage of flexibility sources. This concerns coordination between TSO and DSO, but also between local and regional DSOs. It is important to develop solutions allowing flexible resources to participate in both TSO and DSO markets. If market liquidity is split between markets, the worry is that much-needed profitability is lost. An easy solution for forwarding bids was tested briefly in Norflex. This leads the participants in the direction of a Flexibility register, providing an overview of distributed flexibility assets and their properties. Participants in Euroflex will pilot such a register and also determine rules on how and by whom the resources can be activated across markets.
Finally, building a functioning flexibility market takes time. Participants state the importance of predictability and using similar solutions and market setups for different local markets. It is also important to remember that household customers are primarily incentivised to be flexible due to volatile power prices (implicit flexibility). Explicit flexibility is something they engage in as an add-on, often after being informed by their utility company (flexibility provider).
Sthlmflex - R&D project with common market design and products (2020-2023)
The grid issues (restricted grid capacity) that the pilot was established to solve turned out to be significantly smaller than expected. This was due to two main reasons: i) mild winters and ii) high electricity prices. The need for flexibility is strongly dependent on weather conditions. A mild winter gave a smaller need for flexibility than anticipated. Furthermore, the Russian invasion of Ukraine led to electricity price spikes, resulting in lower consumption and decreased need for flexibility.
The case study identified several regulatory matters that can prevent well-functioning use of local flexibility in Sweden. Grid tariffs favouring a flatter consumption curve can decrease the profitability of offering flexibility. Furthermore, Swedish DSOs have stronger incentives to invest in grid capacity than to solve grid issues with operational measures, like flexibility. Finally, the structure of subscriptions Swedish DSOs pay towards the TSO also disfavours using local flexibility.
The pilot also revealed important learnings regarding the loads and procurement solution. Domestic heat pumps showed promise as they are numerous, and their electricity consumption is easy to predict. These loads respond too slowly to participate in reserve markets, but can participate in local DSF markets where a longer response time is acceptable. However, all the flexibility that was purchased was not actually delivered. This raised questions about the reliability of FSPs and discussions on how to validate deliveries in a proper way. Finally, an important learning from Sthlmflex was that tight communication between the aggregator and DSO prevents misunderstandings and redundant work.
Effekthandel Väst - Permanent DSF Platform (Permanent from 2022)
Previous pilots have shown that market liquidity is essential to ensure FSPs invest in technology, routines, and competence-building needed to participate in the market. The DSOs "invested" in the market by choosing to solve grid issues through the market, even if they had other alternatives to solve an issue. Still, the DSOs preferred bilateral agreements to use the local market to solve grid issues. This is seen as a barrier to sufficient market liquidity.
An important learning is that baselines should be adjusted according to the appliance. This is, for example, important for fridges and freezers, as food hygiene rules restrict their flexibility potential. Rigid baselines could prevent profitability for loads. If communication protocols are to be standardised, it is important that the standards give room for the baseline to be adjusted to the loads’ unique consumption profile.
It is important that flexible loads are located where the grid issues are. Loads located in areas with high population density are more likely to help solve a grid issue. This makes loads like fridges and freezers at supermarkets interesting.
E.ON Energidistribution’s flexibility markets (Permanent from 2023/2024)
E.ON owns and operates 70 local grids in Sweden and has established local flexibility markets in nine of its grid areas. E.ON experiences grid issues in several of its grid areas due to increased electricity demand and a higher degree of intermittent power generation, combined with low available flexibility potential from large industrial customers. Before establishing a new flexibility market, E.ON investigates whether flexibility can solve the grid issues at hand, whether a local DFS market is the best way of procuring flexibility, and whether this is the most cost-effective tool. After deciding to establish a new market, E.ON invests in the market by guaranteeing a number of years the market will be operated. They also set minimum availability payment levels. Ensuring market liquidity from the start is seen as essential to make potential participants trust that it will be profitable for them to go through with the necessary investments.
E.ON moved the management of its flexibility markets to its line organisation in 2022 and considers flexibility as an integral part of handling the grid in the future. They use the same processes and guidelines throughout the organisation. This includes the definition of baselines, prequalification of resources, and sanctioning of failure to deliver.
E.ON has also focused on developing automated solutions for the calculation and publication of flexibility needs, making it possible for FSPs to integrate with the API and enabling a smoother validation process. They have experienced that aggregators are crucial due to their know-how, hardware, and software regarding integration towards different market platforms and appliances. Aggregators also play an important role in defining viable baselines, increasing how much of the flexibility potential of the resource is being used.
FUSE project - Research pilot/potential study (2020-2023)
The FUSE project showed that it is technically feasible to shift charging demand to align with grid constraints without impacting user convenience. More specifically, it was shown that controlled smart charging can reduce peak loads by 20 per cent. This has a significant impact in high-density urban charging environments.
Another important learning is that latency issues in data transmission can significantly affect the responsiveness of charging control. Delays exceeding 500 milliseconds in load adjustment can reduce system efficiency. Standardised communication protocols will help ensure real-time interoperability between chargers, aggregators, and grid operators.
Different metering approaches to validate load reductions were tested in the project. As expected, sub-metering at the charging station level provides more accurate verification than aggregated household consumption data. Sub-metering is, however, more costly.
Bornholm’s Energi og Forsyning (DSO): EcoGrid 2.0. Full-scale pilot/research study (2017-2019)
The EcoGrid 2.0 project demonstrated that heat pumps and electric radiators can provide flexibility, but response times and control precision varied significantly. Furthermore, heat pumps exhibited delayed activation and rebound effects, making them less suitable for fast-response flexibility services. 9.7 per cent of flexibility activations failed due to lost signals or delayed data transmission. The issues could have been avoided with optimised control algorithms, standardised communication protocols, and full access to APIs.
Consumer engagement was limited due to low financial compensation. Households were compensated through reduced electricity costs. The average annual reduction was, however, only 100 DKK per household.
Both the DSO and the TSO used the flexible resources to solve issues in their grids. A lack of market mechanisms for resource allocation and coordination frameworks resulted in conflicts where the same resource was sometimes needed for both system-wide balancing and local congestion management.
Finnish DSO Helen Electricity Network and Fingrid: Pilot of marketplace for DSF (2025 – 2027)
DSO Helen Electricity Network’s and TSO Fingrid’s common marketplace for congestion management goes live in Q1 2025. The marketplace is mainly used for predictable congestion situations. Many flexible loads can’t be used for reserve markets, e.g., due to the reaction time requirements. The new marketplace will harness these loads, especially for congestion management, both for DSOs and TSOs. In case of a sudden grid fault situation, Fingrid will first use other congestion management methods, and the common marketplace comes in later if the congestion continues.
Value stacking (being able to participate in several markets) has been mentioned in many case studies as needed to ensure profitability for resources participating in local DSF markets. The common marketplace in Finland has enabled the necessary TSO-DSO coordination and gate closure times to allow for value stacking. In practice, the bids that are not accepted for DSO’s ShortFlex product can be used automatically for TSO’s ShortFlex product. The gate closure times also take into account the gate closure times of other markets, especially mFRR/aFRR capacity and energy markets.
Finnish DSO Elenia: DSF service for the day-ahead market based on smart meters (2023 à)
A significant number of loads, such as ESWHs and electric underfloor heating from single and double-family houses, are already connected to smart electricity meters’ control relays and have good demand response potential, without any need for additional devices or installations. Today, the loads are mostly controlled according to a simple day/night grid tariff, but the second-generation smart meters offer possibilities to develop services responding to DSO tariffs, spot prices, intraday, and local DSF markets.
The solution for controlling the relays according to spot prices was piloted by Elenia, and most of the customers who participated report that their living comfort was not affected. The next generation of smart meters can respond fast, but the communication infrastructure limits participation in the reserve markets. However, Elenia’s second-generation smart meters (Aidon’s meter) are capable of controlling loads also according to frequency.
Overall flexibility potentials
A rough estimate of the overall national flexibility potential of the flexibility source in focus is provided per case study. These estimates should not be interpreted as exact calculations but rather as indicative numbers on aggregated flexibility potentials from smart household appliances in the Nordic countries. The estimate and hence the extent of smart appliances’ ability to support the power system can help inform policy discussions on how far policies and regulations should go in terms of facilitating demand response from the appliances in question.
Table 5 summarises the estimated potential per appliance, case study, and country. The potentials are in the same order of magnitude as the installed capacities in some of the largest power plants in the Nordic countries. The method of estimation varies somewhat between the case studies. The values are therefore not directly comparable. For more details, the reader is referred to the subchapters below.
Table 7: Estimated potentials per appliance, case study, and country*
Potential | ESWH | Ventilation | Underfloor heating | EV charging | Heat pumps | Freezers and fridges | |
ByggFleks* | ∽ 1 GW | ||||||
BattFlex | 2,5 GW | ||||||
Norflex/Euroflex | 1-6 GW | ||||||
Sthlmflex | 2-8 GW | ||||||
Effekthandel Väst | 0,6 GW | ||||||
E.ON Energidistribution’s flexibility markets | N/A | ||||||
FUSE | 0,9 GW | ||||||
EcoGrid | 0,2 GW | ||||||
Helen Electricity Network | 1,7-3 GW | ||||||
Elenia | ∽ 1 GW |
*For Byggflex, the estimated amount refers to commercial buildings. Since E.ON Energidistribution’s flexibility markets are analysed primarily from a market perspective, the potential of single appliances was not determined.
Value chain for flexibility
Before entering the case studies, we provide an overview of the value chain for flexibility, including all aspects that need to be in place for DSOs to procure flexibility as a solution to grid issues. The description of the value chain makes it easier to understand the processes and systems used in the different case studies, and it also makes it easier to see where the barrier lies.
We have defined the value chain for flexibility according to 8 steps, shown in the figure below. In addition, there is a need for i) platforms and systems, ii) regulation, and iii) TSO-DSO coordination along the whole value chain. The case studies typically do not cover the whole value chain but focus on steps 5-7. Each of the case studies includes a flexibility value chain, explaining the process and systems used in the specific case study.
Figure 1: Value Chain for Flexibility
Explanation of the 8 steps:
- Grid development. For the TSO/DSO to be able to use flexible sources as a part of their service, they need to plan for it. This means including flexibility when planning/developing the grid. This requires that the TSO or DSO can trust that the necessary flexibility is available when needed (time, location, volume, etc.). It also requires that grid developers have the relevant data available.
- Grid operation. This step includes the necessary actions to make it possible to use flexibility as part of operating the grid. This can, e.g., be related to including information on flexibility in operating systems (SCADA/NIS/GIS).
- Assess the need for flexibility. The TSO or DSO assesses when and for what they need flexibility, meaning which situations in the grid and when this happens. This will, in many cases, require analysis tools and/or operation planning.
- Set the price/value for flexibility. Usually, there are different ways a specific situation in the grid can be solved, where some will include using flexible sources. The TSO or DSO should have incentives to choose the cheapest solution that solves the problem and will to an assessment on the price they are willing to pay for the needed flexibility.
- Mechanism/measure. Establishing a mechanism/measure enabling the procurement of flexibility. This can be a marketplace, bilateral agreements, conditional connections, etc. Prequalification of resources is necessary when using a marketplace as a mechanism.
- Delivery of flexibility from FSP. The flexibility service provider takes action to be able to provide their flexibility (because they expect it to be profitable). For this to happen, you need:
- A willing Balancing Responsible Party (BRP) for aggregators
- Sufficient income potential and profitability for the end-user, flexibility service provider, and balancing responsible party
- Sufficient information and understanding
- Activation and verification. Establishing systems that make sure loads are activated and that can verify that they were, and that there is a standardised way to do this. Important parts of this are:
- Availability - loads actually being available when they are needed
- Baseline - what would the power consumption/production have been without activation
- Verification – systems for verifying activation
- Settlement and payment. Establishing systems for settlement and payment, in addition to penalties to actors who do not fulfil their market obligations.
Norway: ByggFleks
General case description
ByggFleks was a part of a project called StrømFleks, led by Norwegian DSO Lede. StrømFleks demonstrates how DSF from different sources can be utilised to manage local issues in distribution grids. Apart from Lede, stakeholders included Horten municipality, Vestfold and Telemark County, property developer Format Eiendom, research institution SINTEF, and smart meter supplier Aidon. The entire project was conducted from 2019 to 2023 within a limited grid area. The knowledge and results from the project were transferred and continued by Lede in other flexibility projects in other types of commercial buildings. Important information sources for ByggFleks are the project’s final report
Prosjekt «Storskala uttesting av forbrukerfleksibilitet» (Lede & Menon Economics, 2024). Project ID 18/16282
In ByggFleks, the focus was on the grid customers’ side of the value chain, corresponding to steps 6 and 7 in the value chain outlined in Chapter 3.1, and to test the technical solutions and systems for activation and aggregation of flexibility from commercial buildings. The project aimed to reduce and distribute peak loads from various appliances. No grid issues were specified, as the main focus of the project was system testing and to gain a better understanding of commercial buildings’ flexibility potential.
Four commercial buildings with Building Automation and Control Systems (BACS) installed participated in the pilot
Information regarding capacities per load and ByggFleks-specific numbers obtained from dialog with Lede
Table 8: Flexibility resources in ByggFleks
Building type | Appliances made available for DSF | BACS status |
High school 1 | Room heating, water heating | New building, fully integrated BACS |
High school 2 | Room heating, water heating | Older building, partly integrated BACS |
Elementary school | Room heating, ventilation, and water heating | Partly integrated BACS |
Sports hall | Room heating, ventilation, and water heating | Partly integrated BACS |
Value chain for DSF
Systems in use
The DSF value chain is illustrated in the table below.
Table 9: Physical value chain for DSF in ByggFleks
Load control is a unit installed on an appliance to allow it to connect and communicate with a central control system, either through a hardware connection or cloud-to-cloud integration. In ByggFleks, some appliances were already equipped with a load control, while other appliances were not, and local adjustments had to be made to integrate these appliances.
Building Automation and Control Systems (BACS) are systems used to control and automate loads or systems within buildings. A single BACS can manage all loads in a building, or a building may feature multiple BACS or independent loads operating without a BACS. BACS is usually operated by a building manager, and various sorts of interfaces exist to monitor and control the connected loads. A separate top control system is another option for integrating building appliances and/or BACS. In ByggFleks, a top system from Envo was utilised to connect some of the appliances and to merge separate BACS systems. The Envo control system had all the required protocols for integrating the appliances (HAN, Modbus, Enbus, Bacnet, etc.), enabling control and overview of production/consumption of all building appliances within a single system. The different routes taken to integrate the building loads are illustrated in Figure 4. To integrate existing loads that do not have standardised interfaces via a local integrator such as BACS and/or Envo is necessary to make these loads available for flexibility purposes.
Figure 2: Integration of building appliances through various routes in ByggFleks
The DSO used a platform called the FlexTools Platform to aggregate flexibility from the end-users by integrating with the BACS and/or top control system. The platform could integrate with the BACS and/or the Envo control system through a hardware connection (RTU channels
Flextools offers a Remote Terminal Unit (RTU) with controllable outputs that collects measurement data and acts as an interface, compatible with various Building Management Systems/BACS. RTU is an electronic hardware device that can be used to directly connect various electronic devices to an external system.
Customer management and involvement
For each building, a schedule indicating at what time periods DSF would be available was developed in collaboration between the DSO and building managers
Feedback from building managers and owners suggested some areas of improvement, primarily related to user-friendliness and further incentives to utilise flexibility. In buildings where it was deemed necessary to install a separate top system to aggregate all the appliances, building managers had to deal with dual control systems and interfaces in parallel to manage the various appliances. Further, as there were no market mechanisms or payments for DSF, the users expressed limited interest in further participation in the project without further incentives.
Barriers and enablers for DSF
The main challenges encountered in ByggFleks were related to variations in appliance smartness and interoperability. According to the interviewees, older appliances had a limited ability to integrate and communicate with top control systems. Some interviewees suggested that appliances in older buildings may use proprietary protocols and, therefore, cannot be programmed for load control or accessed. In newer assets, however, BACS integration was deemed manageable despite various interfaces. Once integration is done, the logic can be reused to integrate new assets. For interface standardisation not to be an issue, buildings should operate on the most common communication protocols. Apart from interoperability, no barriers related to the appliances themselves were identified in the project.
Integration and interoperability issues: At the Delivery-Activation stage (steps 6 and 7) of the flexibility value chain, many older building appliances cannot readily connect to overarching control systems because they use proprietary protocols or lack load-control units. This technical mismatch means device “smartness” and interoperability vary widely, curtailing the demand-side flexibility that commercial buildings can offer and risking lock-in between customers, original-equipment manufacturers, and balancing-service providers. Work-arounds, such as installing extra gateways or interface modules, do exist and can be reused once set up, but they add cost for building owners and demand time from service providers. Because today’s appliances rely on a multitude of protocols, gateways, and APIs, interviewees view the resulting complexity as a systemic hurdle. The most effective enabler is forward-looking standardisation of appliance-to-system communication, ensuring new equipment ships with open, harmonised interfaces that eliminate bespoke integrations and unlock greater flexibility potential.
Lack of mandatory load-control provisions in building regulations: At the Delivery stage (step 6) of the flexibility value chain, current rules do not require that appliances installed in new commercial buildings come equipped with load-control capabilities. Without this regulatory push, many devices enter service unable to respond to external control signals, which sharply limits the DSF that building operators can provide to the grid.
Unclear stakeholder roles and dual-selling risks: Across the Implementation-to-Settlement stages (steps 5 through 8) of the flexibility value chain, multiple parties can be involved in installing, integrating, and operating building appliances, each of whom might technically act as a flexibility service provider and bid the same resource into different markets. While this was not a problem in the ByggFleks pilot, which confined itself to technical tests of integration and activation, interviewees flagged it as a potential challenge if the project scales into a market-based arrangement. Without clear lines of ownership and control, the same kW of flexibility could be “sold” twice, undermining market integrity. The enabling measure is to establish explicit, transparent rules that define who owns each asset, who is authorised to dispatch it, and how participation across multiple flexibility markets is monitored and enforced.
Lack of compensation for flexibility activation: At the Settlement stage (step 8) of the flexibility value chain, building owners received no payment when their appliances were activated to provide DSF. Without a direct economic reward, owners quickly lose interest in taking part in further pilots, and the overall pool of commercial-building flexibility remains underutilised. The most effective enablers are (1) clear bilateral contracts in which the DSO and asset owner spell out the terms under which availability and activations will be paid for, or (2) migrating the assets onto a competitive market platform where transparent price signals automatically remunerate each activation.
Complex, multi-interface control systems: During the Delivery-Activation stages (steps 6 and 7) of the flexibility value chain, many commercial buildings relied on several overlapping control installations and user interfaces. This patchwork made it hard for facility managers to obtain a clear, real-time view of energy use and appliance status, hampering day-to-day optimisation and limiting the building’s DSF potential. Without systematic training or onboarding for building managers when new systems or concepts are introduced, the knowledge gap widens over time, discouraging deeper DSF engagement. The key enabler is a structured competence-development program that equips building managers and operators to exploit their systems’ full flexibility.
Overall potential for DSF from commercial buildings
Electric System Water Heaters (ESWHs) were the resources that provided most of the flexibility in ByggFleks. The buildings investigated in the project were schools and a sports hall, and thus, most of the hot water goes to showers. The ESWH capacities were 30- 105 kW
Information regarding capacities per load obtained from dialog with Lede.
Electric underfloor heating consists of electric heating cables cast in concrete and can be fully turned off for flexibility purposes without harming any components. The thermal inertia in the concrete allows for activation for at least one hour without users noticing. Beyond one hour, the availability depends on how densely the heating cables are packed in the flooring, how many people are in the building, and the outdoor temperature. Despite being installed at different times following different building stages, most of the heating cables already had load control and were integrated with a BACS. Their installed capacities and flexibility potential ranged from 35 to 300 kW.
The flexibility potential of ventilation depends on the air quality requirements of the users and the general air quality within and surrounding the building. In ByggFleks, some of the schools were located near garbage dumps or in areas with higher Radon levels, which meant that the ventilation system had to maintain a certain air quality level. Furthermore, parts of the testing were conducted during the COVID-19 pandemic, where hygiene and air quality requirements in schools were elevated. In contrast to other appliances tested, ventilation cannot be cut directly. Instead, the energy consumption is reduced by setting the system in recirculation mode or reducing the air flow temperature. Recirculation can only be done in short sequences without affecting the air quality. In this case, the maximum duration was set to one hour. In a building with multiple smaller ventilation systems instead of a few larger ones, the flexibility potential is expected to be larger, as you can rotate between zones. The ventilation system tested in ByggFleks had a total capacity of approximately 210 kW.
The DSF potential of commercial buildings is highly dependent on the building in question. For instance, the age of the building may determine how long room heating may be deactivated before it starts affecting the comfort of the users. The limited activation time for all appliances tested in the pilot leaves the question of how long the appliances can be activated without affecting user comfort open. The interviewees’ reflection is that the user comfort, especially hot water and room temperature, would be unaffected for a longer period than the time tested here. However, they also note that longer activation periods would result in rebound effects, which would coincide with other loads in the building. As newer buildings usually are more energy efficient than older ones, there also might be a limited potential for further energy optimisation efforts. Lastly, the BACS setup and the complexity level of integrating appliances with a top system will determine whether it is economically feasible to utilise the flexibility from the building’s appliances.
Overall potential
ByggFleks demonstrates that ventilation, room heating, and water heating can be flexible, provided that it is feasible to integrate these appliances with a sort of top system. Assuming that all schools have some form of ventilation system, water heating system, and heating, we can make the following assumptions regarding the theoretical flexibility potential in Norwegian school buildings from different energy appliances (Table 7).
Table 10: Overall potential estimation based on ByggFleks.
Appliance | Capacity per unit kW for different school sizes* | Number of buildings | Theoretical potential |
ESWH | Small: 30 kW Large: 105 kW | In 2024, there were 2692 elementary and middle schools and 418 high schools in Norway. Large (>300 students in elementary school, >900 in high school): 953 schools Small: 2157 schools | 164,7 MW |
Ventilation | 180 kW | 484,6 MW | |
Room heating | Small: 45 kW Large: 350 kW | 430,6 MW | |
Total potential | 1080 MW |
*Based on documented appliance capacities in ByggFleks. It is unclear how representative these capacities are of the entire appliance portfolio.
The size, age, and energy usage in schools vary significantly. The above estimates should thus be considered as rough. A study from Multiconsult
Tilstand skolebygg (Multiconsult, 2022)
TG scale from 0 – 3, where 0 = as good as new, and 3 = significant rehabilitation needed
Norway: Battflex
General description
The BattFlex pilot was led by the Norwegian DSO Elvia and was one of four initiatives under a large-scale demonstration project, IDE – Intelligent Distribution of Electricity. The project was conducted from 2019 to 2024 and led by The Norwegian Smartgrid Centre in collaboration with a consortium of partners, including the grid companies BKK, Elvia, Tensio, Agder Energi, Norgesnett, and Lede, as well as the NTNU university and EPOS Consulting.
The Battflex pilot aimed to demonstrate the potential, as well as evaluate the impacts and benefits, of utilising consumer flexibility from smart electric storage water heaters (ESWHs) and grid-connected batteries to address voltage quality issues in weak or vulnerable low-voltage networks. The project sought to verify that such solutions could provide a cost-effective alternative to traditional grid reinforcement.
The BattFlex pilot originated from the KAFFI project, conducted by Eidsiva Nett (now Elvia) in 2018 in collaboration with THEMA Consulting. The KAFFI project aimed to evaluate the need for and potential of activating local flexibility in the distribution grid. The study analysed alternative solutions for 17 selected transformer circuits, demonstrating that such measures could reduce costs by 75% compared to conventional grid reinforcements. The Battflex pilot was further shaped by insights from the Austrian LEAFS project and Ringerikskraft Nett’s EffektPILOT. Findings from the LEAFS project emphasised the importance of activating consumer flexibility in a "grid-friendly” manner rather than relying solely on price signals. This became central to the development of control algorithms for flexibility activation in BattFlex, optimising activation of ESWHs based on both price signals and dynamic voltage parameters.
As part of the pilot, 72 ESWHs from OSO Hotwater were installed in households across two chosen transformer circuits in Elvia’s grid area. In addition, three grid-connected batteries were deployed across three chosen transformer circuits in the low-voltage grid, also within Elvia’s grid area. The developed and tested solution for smart ESWHs incorporated control algorithms that automatically and autonomously optimise activation of the appliances based on both price signals and dynamic voltage parameters, testing how price optimisation impacts voltage issues in vulnerable low-voltage networks (more on this below). Since the focus of this report is on end-user appliances, the following sections highlight findings regarding the flexibility offered by ESWHs.
Findings from the BattFlex pilot have since been integrated into various initiatives and products. OSO Hotwater continues to sell smart water heaters with functionalities developed during the pilot, including the control algorithm for flexibility activation and a customer app, as further described in the sections below. Additionally, Enova's support scheme for smart ESWHs was developed in collaboration with Elvia following the Battflex pilot, confirming that such a scheme would provide significant added value and that smart ESWHs could deliver the desired functionality. Insights from the battery solution in Battflex were also applied in the IDE project's FBI pilot, which explored how voltage boosters, combined with batteries, can defer investments and help mature the technology for broader use.
For Battflex, important information sources include the pilot’s final report
Battflex Pilot Report “Nytteverdier fra smarte varmtvannsberedere” (Elvia, OSO Energy & EPOS Consulting, 2022)
Value chain for DSF
The DSF value chain is illustrated in the table below.
Table 11: Physical value chain for DSF in Battflex
Systems in use
In the pilot, smart ESWHs were delivered to the participating end customers with a hardware unit, the OSO Charge smart control unit, pre-installed on the ESWH. When installed, this hardware enables both the activation and measurement of flexibility from the water heater. For OSO water heaters manufactured from 2017 onward, the OSO Charge unit can be installed on the ESWH to retrofit it with smart functionality, even if it was not originally included.
Each OSO Charge unit is equipped with an internal measurement circuit that meets the Nordic resolution standards
As mentioned initially, OSO smart ESWHs were developed in the pilot to assess how market-based optimisation impacts grid voltage levels and whether smart automation can mitigate potential grid voltage issues. The pilot found that pure price optimisation could lead to rebound effects, in this context referring to a sudden surge in load when multiple water heaters reconnect simultaneously, exacerbating voltage issues in vulnerable low-voltage grids. In contrast, optimising for both price and voltage provided significant voltage support and added value for customers through reduced grid tariffs and electricity costs. To achieve this, OSO Charge incorporates so-called 'grid-friendly activation,' using control algorithms that optimise the ESWH's power consumption based on both local voltage and Nord Pool spot prices. OSO Charge autonomously activates flexibility according to predefined optimisation parameters, enabling operation without external control or customer involvement. The pilot tested only the activation of implicit flexibility. The algorithms are designed to prioritise safety (e.g., temperature control to prevent health risks), followed by primary performance (ensuring a consistent hot water supply), price optimisation, and, finally, system-level optimisation.
For price optimisation, market prices are retrieved from the Nord Pool API at 1 PM the day before. The machine learning algorithm then considers the price optimisation alongside the estimated future consumption of the appliance to plan heating for the next 24 hours, with optimisation occurring twice daily: between 2-3 AM and at 12 PM. The updated control profile is received by OSO Charge and runs locally on the device, meaning it does not rely on an internet connection to function.
Voltage optimisation enables grid-friendly activation by adjusting the setpoint for each appliance based on historical voltage data. The algorithm is dynamic and adjusts based on factors such as seasonal variations, providing voltage support based on the appliance’s grid location. OSO Charge enables the appliance to respond to voltage levels with a reaction time of 1.3 seconds, an accuracy/resolution of 5 MHz, and a reporting frequency of 10 Hz. Additionally, a fleet of OSO Charge-enabled water heaters can coordinate to reduce load during local voltage disturbances, ensuring a balanced aggregation of available flexibility.
As mentioned above, the algorithm prioritises price optimisation over system optimisation. However, grid constraints can override price-based optimisation to ensure grid-friendly activation at the local level (power and voltage levels). In the event of a significant voltage drop, OSO Charge attempts to deactivate to prevent instability. OSO has specific terms that allow for temporary disconnection of devices to support grid stability. However, if the ESWH remains within the acceptable voltage range, it continues to be controlled based on price optimisation.
Oso Charge also has the functionality for explicit activation of ESWHs, such as activation for delivery into FFR markets, although this functionality was not tested in Battflex. The internal measurement circuit in OSO Charge is technically approved for FFR, as its resolution and response time meet the requirements for participation in these markets.
Customer management
Through the OSO inCharge app, customers gain access to detailed consumption data and full control of their ESWHs’ consumption. The app offers multiple modes, including automatic, manual, third-party control (which offers the ability to control water heating via OSO’s open cloud-based API), sleep mode, and renewable energy mode, enabling customers to choose how their ESWHs operate based on their preferences. In the app, customers can define their flexibility preferences and determine the level of risk they are willing to accept regarding how aggressively the ESWH optimises based on price, with a slightly increased risk of reduced comfort. When connected to the app, OSO has access to individual-level data for the ESWH and, as previously mentioned, can explicitly activate the appliances; however, this is only available with the user's approval and was not tested in the Battflex pilot.
As part of the smart water heater rollout in the pilot, Elvia developed a digital process for managing customer communication and flexibility agreements. Customers also received compensation in the form of free smart ESWHs when participating in the pilot. While the customer relationship was owned by OSO, Elvia handled both compensation and the management of agreements for participation in the pilot. The pilot demonstrated that smart ESWHs, optimised for both price and voltage, provided value to customers by reducing both grid tariffs and electricity costs.
Barriers and enablers for DSF from ESWHs in Norway
The interviews identified several barriers to utilising consumer flexibility from smart electric storage water heaters (ESWHs) in Norway, as summarised below. Notably, no significant technical barriers were reported in the Battflex case. The interviewees indicated that integrating, aggregating, and activating ESWHs for flexibility did not present technical challenges. However, this finding should be seen in the context of the pilot, where the flexibility service provider (OSO) was also the supplier of the appliances and the technical system for activating flexibility from these. Hence, the FSP only needed to integrate with a single type of appliance, for which they had designed the interface themselves.
Rebound effects: Interviewees highlighted that ‘rebound effects,’ in this context specifically referring to a momentary surge in load when multiple water heaters are reconnected, and resulting voltage issues, as potential consequences of utilising flexibility from ESWHs. The project found that pure price optimisation led to rebound effects and negative impacts on grid voltage levels, which further aggravated voltage issues in vulnerable low-voltage grids.
In Battflex, the issue of potential rebound effects was mitigated through implementing an automatic and autonomous voltage control algorithm in the smart control unit of each appliance. Combining this with price optimisation, grid-friendly activation was achieved, and rebound effects were prevented.
As an enabling measure, the project therefore recommends avoiding large-scale installation of solutions on the customer side that cannot be activated/controlled based on voltage parameters and do not have the necessary autonomy to prevent exacerbating issues related to voltage quality and voltage levels in the low-voltage grid. The project suggests that ensuring that suppliers of local flexibility solutions implement "grid-friendly activation" can be achieved through various means, including technical or regulatory requirements, collaboration with suppliers, and economic incentives.
Although the interviewees did not specifically address the cost of implementing voltage control compared to a solution based solely on price optimisation, the report and interviews suggest that there are no significant technical or economic barriers to implementing such solutions. The approach primarily involves programming control algorithms to activate based on both dynamic voltage parameters and market prices. Like price optimisation, this method also provides value to customers through reduced grid fees and lower electricity costs.
Regulatory restrictions hindering market access: Interviewees highlight that unlocking significant flexibility from end users in Norway requires that more flexibility providers participate in the market. So far, local flexibility in Norway has mainly been demonstrated through pilot projects, and there is a lack of scalable, commercially viable solutions that incorporate profitable and efficient business models for delivering flexibility.
Several interviewees noted that current regulatory restrictions for participation in TSO markets (see paragraph below) make it challenging for independent Balance Service Providers (BSPs), such as OSO, to offer flexibility from their customers in both TSO and DSO markets and thus discourage such providers from developing effective, scalable business models for flexibility.
In the FCR markets, current regulations require that the flexibility provider also hold balance responsibility for the regulated asset
A new model for balance responsible parties and providers of balancing services is planned to be introduced by the end of 2024 (Statnett, 2024) https://www.statnett.no/for-aktorer-i-kraftbransjen/nyhetsarkiv/ny-modell-for-balanseansvarlige-og-leverandorer-av-balansetjenester-planlegges-innfort-ved-utgangen-av-2024/
In the aFRR and mFRR markets, interviewees highlighted that the minimum bid size required by current market regulations, which ranges from 1 to 5 MW in FFR depending on the contract type, makes it difficult for actors providing flexibility from small household appliances to participate. For OSO, this minimum bid requirement means they cannot aggregate enough loads to deliver flexibility directly into these markets. Instead, a larger aggregator would need to take responsibility for submitting bids on behalf of OSO’s ESWHs in the reserve markets, adding administrative burden for OSO.
While OSO’s ESWHs are technically approved for reserve markets, under current regulations, this was deemed too complex to make participation worthwhile for OSO. Moreover, current TSO market regulations prevent BSPs from participating in both local flexibility markets and reserve markets with the same capacities. More on this will be discussed under the Norflex pilot, which is mentioned later in the report.
The interviewees highlighted that utilising flexibility from appliances is easier when there is a certain level of homogeneity in the market (homogeneous appliances). This suggests that unlocking the flexibility potential of household appliances in Norway might be challenging if one single actor (for instance, a BRP) needs to integrate, aggregate, and manage a diverse range of appliances.
Looking at potential enabling measures, changing the current regulations to allow independent BSPs to more easily sell flexibility from their end customers directly into both local and reserve markets could promote entrepreneurship and facilitate the entry of new flexibility providers. This would unlock flexibility from a broader range of loads and support more profitable business models for both providers and end users.
Interviewees highlighted that regulations better suited to providers offering flexibility from aggregated and small household appliances would be particularly beneficial. They pointed out that the regulatory barriers faced by BSPs in delivering flexibility present a greater challenge to unlocking the flexibility potential of ESWHs than the lack of standardisation and ‘smart readiness.’ Additionally, the pilot project demonstrated that overcoming technical barriers may be more achievable if the flexibility provider is also the supplier of the appliances and/or the flexibility solution. This suggests that such regulatory changes could help mitigate both regulatory and technical barriers.
Specifically, interviewees emphasised that BSPs should be allowed to deliver flexibility in the FCR-D and FCR-N markets without requiring an agreement with a BRP, as is currently the case for FFR markets, given that the potential imbalance costs for the BRP would be minimal also in these markets. One interviewee further suggested that participation in the mFRR CM market should also be permitted if the activation price is set sufficiently high.
Additionally, it was proposed that the minimum bid size requirement in the reserve markets should be adjusted to allow for aggregated loads as an exception, thereby enabling smaller actors to participate. It was also suggested that making reserve market bid data publicly available could help new entrants establish themselves in the market. Providing access to market data, also for actors who are not yet participating in the market, could give them valuable insights into market dynamics and the potential earnings from participating in these markets, enabling them to better position themselves and develop viable business models.
Inefficient customer incentives for providing flexibility: Based on interviews, commercialising flexibility solutions requires business models that facilitate end-user participation without complex prequalification processes and are perceived as simple and reliable. Furthermore, these models should enable efficient digital compensation and management of end users, eliminating the need for direct communication between the flexibility provider or DSO and each end user.
According to interviewees, the current Enova scheme, which provides support for end-users purchasing smart water heaters, may be too complex to effectively incentivise the adoption of smart ESWHs over non-smart ones. The application process is handled on an individual basis, and it requires, among other things, that the ESWH be compatible with smart home systems.
As for enabling measures, interviewees highlighted that it might be beneficial if equipment suppliers could pre-qualify their products for participation in the support scheme, streamlining the process for end users and making it a 'no-brainer' to choose smart appliances. One could imagine that such a process might be easier to streamline if the appliances followed a standardised code of conduct and had an energy label indicating their smart readiness.
Overall potential for DSF from ESWHs
According to OSO, a typical water heater has a capacity of 2–3 kW. With around 2.7 million private water heaters in Norway, ESWHs collectively represent over 5,000 MW of potential capacity, assuming an average of 2 kW per unit. Of these, 200,000 are OSO units with smart functionality potential, contributing 400 MW to the total capacity.
Test results from BattFLEX indicate that approximately 50% of the installed capacity in a fleet of ESWHs can be aggregated for flexibility during peak load hours. This means up to 2,500 MW of flexibility could be activated at any time, with 200 MW coming from OSO’s smart-enabled units. ESWHs provide instantaneous flexibility, as they can be easily switched on and off without any issues.
BattFLEX results indicate that ESWHs can remain off for most of the day without users noticing, allowing for up to 18 hours of downtime within 24 hours. OSO estimates that each unit provides an average daily flexibility of 12 kWh, resulting in a total estimated potential of approximately 32.4 GWh per day from all Norwegian water heaters, with 2.4 GWh coming from OSO’s units alone.
Norway: Norflex
General description
The purpose of the Norflex project was to increase the use of DSF through new market-based solutions to manage local grid bottlenecks. In total, 31,470 trades were carried out, amounting to 1.39 GWh from over 4,000 assets spread across southern Norway
Whitepaper Norflex (2020-2023)
Sluttrapport Storskala demonstrasjon av fremtidens energisystem (Norflex, 2023)
The asset portfolio spans from industrial and commercial buildings to batteries, EV chargers, and household appliances. For the scope of this report, the following sections will primarily focus on aspects related to EV chargers and other household appliances.
Value chain for DSF
In Norflex, the focus was on developing a functioning marketplace for local flexibility trade, corresponding to steps 4-8 in the value chain outlined in Chapter 3.1 above. The value chain and main stakeholders are illustrated in the table below.
Table 12: Value chain for DSF in Norflex
Market Platform and Market Participants
The demand for flexibility from DSOs and the supply of flexibility from flexibility providers are matched through the market platform NODES. NODES acts as the intermediary for buyers and sellers, which facilitates the reservation and activation of flexibility. The platform also allows transaction automation, including trade signals for activation, validation of delivery, and settling of the transaction. There is no direct contact between the DSOs and flexibility providers. Both DSOs and flexibility providers connect to the platform through open APIs.
Flexibility providers include both large individual actors selling their own flexibility (>1kW) and aggregators who mediate aggregated flexibility from smaller customers. The FSPs are responsible for all customer contact and recruitment and are not required by NODES to disclose their business models and profit strategies. In the Norflex project, only aggregators with balance responsibility (BRP) could participate, but in Euroflex, independent aggregators (BSP) are also allowed to participate
In Euroflex, a FSP can be either a BRP or a BSP. A BRP is responsible for managing any imbalances from their customers. Large BRPs can also prequalify for Statnett’s reserve markets. A BSP is an independent aggregator that does not hold any balancing responsibility but instead offers technical solutions to aggregated end-user flexibility. BSPs were recently granted access to Statnett’s mFRR and aFRR markets, proviso that the BSP has a balancing agreement with a BRP.
Registration of Flexibility Resources and Bidding
Flexibility providers prequalify their resource portfolios in NODES using Flextools, where information such as the main metering ID and capacity per asset is registered. Grid companies use this information to allocate the assets to various nodes in their network.
Following prequalification, flexibility providers can submit bids to the platform. A bid includes information about volume, duration, time, location (allocated through prequalification), price, and direction (up/down regulation). Each bid must have an associated baseline in the event of a matched trade (more on baselines below). Flexibility is reserved for two hours before delivery, and the flexibility provider receives a signal to activate. The delivery of flexibility is validated and settled on the platform two hours later. If there are deviations between the reserved and measured capacity, the payment is reduced accordingly. However, lack of delivery does not result in further penalties for the FSPs.
Products
The flexibility products in Norflex were
Whitepaper Norflex (2020-2023)
- ShortFlex: Physical delivery of flexibility. Market opens for trading 7 days ahead of physical delivery, and bids are cleared 2 hours ahead of delivery. The orders had a duration of one hour. The minimum bid size was 1 kW. FSPs were paid for activation.
- LongFlex: Reservation of flexibility. LongFlex contracts were offered with different durations, either weekly contracts (LongFlex week) or contracts that spanned over multiple weeks (LongFlex season). FSPs were paid for both availability and activation (ShortFlex).
Additionally, a third product, MaxUsage, is under development and is being tested. This is in response to FSP feedback indicating that estimating the baseline in the other products is challenging, particularly for household flexibility
Whitepaper Norflex (2020-2023)
Flexibility from Electric Vehicles (CIRED, 2022)
For both ShortFlex and LongFlex, the traded quantities increased over the pilot period as more FSPs prequalified their asset portfolios and started to bid in flexibility on the platform. In total, 1395 MW of flexibility was traded. The trades had a total value of 12.5 million NOK
Whitepaper Norflex (2020-2023)
Whitepaper Norflex (2020-2023)
DSO-TSO coordination
To ensure efficient usage of flexible assets and to handle local bottlenecks, coordination between Statnett and DSOs is seen by the interviewees as an important prerequisite to establishing functioning local markets for flexibility. Equally important as the coordination between TSO/DSO is the coordination between local distribution grid companies and regional DSOs. Without any clear framework on how bids can be forwarded from one market to another, or through simultaneous access, interviewees remark that it will be challenging to avoid the splitting of liquidity between markets. Furthermore, limited market access for smaller FSPs may limit their income potential and thus also their interest in developing and offering solutions to aggregate flexibility from small assets.
A central part of the ongoing workstreams in Euroflex is therefore to investigate how different markets should be integrated and how products and services must be designed to make it easier for FSPs to offer services to multiple markets. In Norflex, forwarding of uncleared bids from the local market to Statnett’s mFRR market was briefly tested
NorFlex project demonstrate integration to Statnett’s mFRR market (nodesmarket, 2022)
Sluttrapport Storskala demonstrasjon av fremtidens energisystem (Norflex, 2023)
Euroflex (Statnett, 2024)
Customer involvement
Household customers were recruited to the pilot by the FSPs directly. FSPs were also responsible for all customer communication and support. In Norflex, most of the flexibility came from private EV chargers, which were integrated with the FSPs’ own cloud-based data hubs. Similar set-ups are being used by other technical aggregators in Euroflex to aggregate flexibility from more EV chargers, as well as other household appliances such as heat pumps and heat panels. Some customers integrate their entire house into the platforms, while others only integrate one or two appliances.
The communication between the FSP and customers happens through an app, where the customers can monitor their own consumption and define how they wish to be flexible. Depending on what appliances are connected to the hub, the customer can define how they want these assets to be used. These settings can be dynamically changed or overruled by the consumer.
According to the interviewees, most household customers were incentivised to become flexible due to higher spot prices and price volatility, especially in southern Norway after 2020. The price variations and information campaigns from the FSPs made customers aware that there was money to be saved if they became flexible. The FSPs have different business models, where customers either achieve lower energy bills by allowing the FSP to optimise their consumption based on grid tariffs and spot prices, or by receiving points or discounts to use on services or products provided by the FSP, allowing the FSP to trade their flexibility in other markets. The compensation the consumer gets for participating in balancing markets or local flexibility markets is thus not directly related to the prices quoted in the markets. The FSPs report that even though many customers are willing to participate in the flexibility markets, the spot price signals are so strong that most of the capacity activated through their platform is implicit flexibility.
Other important learnings from the project
The key learning point from Norflex is that it takes time to build a functioning flexibility market, both from a technical, economic, and regulatory perspective. The interviewees emphasise the need to establish more permanent solutions and market set-ups across geographical regions to increase predictability for customers, FSPs, and DSOs going forward. The Network Code allows member states to assign the responsibility of setting up and organising local flexibility markets, and a national decision on how local flexibility markets are to be organised is not expected before closer to 2030. Thus, some interviewees suspect that many potential FSPs and flexible resources are pending their participation until such a national decision has been made.
Secondly, on the technical side, a lot of resources went into establishing cohesive, digital value chains. Automated solutions in the interfaces between buyer and marketplace, and marketplace and seller, are necessary to scale activities.
Lastly, as mentioned previously, baseline estimations for smaller loads, such as household appliances, were a challenge in Norflex. Several FSPs reported that the baseline estimation methods could be improved, as there were, in some cases, significant deviations between the estimate and what was actually delivered. For some loads, such as EVs, the unpredictable consumption pattern of the load was partly to blame for the deviation. However, for most loads, the main reason was that data was collected through the customer’s main metering device instead of from the dedicated metering devices installed on the appliances, and the loads are stochastic in nature. The household appliances tested constitute only a small portion of the customers’ total energy consumption. Thus, it was tricky to verify whether flexibility from the assets had in fact been activated due to “noise” from the rest of the household consumption.
Without reliable baseline models, there is a risk that FSPs will be either overpaid or underpaid for their deliveries. The value of flexibility in the short term is thus highly dependent on adequate baseline calculation and forecast models. Currently, the task of defining and calculating baselines is the responsibility of the system operators. The Norflex project, therefore, recommended several measures to improve the estimations going forward, including:
- Stricter requirements for dedicated submeters for individual loads and more granular data collection to verify delivery
- Linking the baseline estimation model to the pending flexibility register for more accurate forecasts
- Allowing all interested parties, including FSPs, to submit new calculation methods for approval
- Developing new products, such as MaxUsage, for appliances where baseline estimation has proved to be more challenging than for other appliances
Barriers and enablers
The following enablers and barriers were identified, both from the literature and from interviews with relevant stakeholders.
Unresolved national framework for local flexibility markets: At the Mechanism/Measure stage (step 5) of the flexibility value chain, the EU Network Code lets each member state decide who will set up and run local flexibility markets, yet Norway’s final decision is unlikely before 2030. This regulatory limbo discourages DSOs from planning such markets, while lingering uncertainty over future market design disincentivises FSPs from investing in new solutions.
Lack of coordination for dual-market participation: At the Mechanism/Measure stage (step 5) of the flexibility value chain, FSPs cannot offer the same capacity in both local flexibility markets and the TSO’s balancing markets because no framework yet exists for forwarding bids or aligning activation rules across market layers. This set-up fragments liquidity, and the limited on-ramp to reserve markets especially sidelines smaller FSPs, reducing their appetite to aggregate distributed resources. The main enablers are two ongoing initiatives: Euroflex, which is designing market products that span local and national platforms, and the Statnett–Elhub national flexibility register, which will catalogue assets, codify activation priorities, and give FSPs streamlined access to participate in multiple markets with the same resources.
Fragmented communication protocols and APIs: At the Delivery stage (step 6) of the flexibility value chain, there is still no common standard for how appliances expose data or accept control signals. Each FSP therefore has to build a bespoke integration for every new device, the effort and cost varying widely with the underlying API. Sluggish interfaces can, in some cases, even bar an appliance from fast-response markets. While FSPs treat integration know-how as part of their value proposition and reuse connection logic once it is in place (often opting for cloud-to-cloud links to bypass on-site hurdles), they stress that harmonising rules, such as the forthcoming Code of Conduct for Energy Smart Appliances, would sharply cut onboarding time and expand the pool of controllable loads.
Spot-price-driven optimisation undermines grid-needs flexibility: Throughout the Delivery, Activation, and Settlement stages (steps 6 – 8) of the flexibility value chain, sharp volatility in the day-ahead spot market makes it more profitable for customers to time-shift smart-appliance use, particularly EV charging, purely on wholesale-price signals. Interviews reveal that EVs are connected for less time than earlier estimates assumed, and during those limited hours, customers earn more by chasing spot-price spreads than by answering DSO requests. Consequently, implicit flexibility based on spot prices crowds out explicit flexibility procured via DSO signals, reducing the controllable capacity available for local grid relief. A practical enabler would be to align or stack incentives, for instance, through local flexibility premiums or tariffs that reflect grid constraints and can be combined with spot-price optimisation – so that serving the grid is at least as lucrative as riding wholesale-price swings.
Voltage fluctuations from abrupt EV-charging ramps: At the Activation stage (step 7) of the flexibility value chain, simultaneous start-and-stop behaviour in low-voltage networks, for instance, when many EVs chase day-ahead spot prices, may cause sharp load swings that trigger local voltage dips and spikes. FSPs can smooth the profile by introducing a “slow-release” strategy, gradually ramping charging power over consecutive intervals, and by co-optimising all controllable household appliances (where customers permit), so total demand stays within voltage-friendly limits.
Inaccurate baselines for small, distributed loads: Across the Mechanism, Delivery, Activation, and Settlement stages (steps 5 – 8) of the flexibility value chain, the Norflex pilot showed that baseline estimates for household appliances often differed from their actual use. The main reasons were drawing data from main meters instead of dedicated submeters and the naturally variable behaviour of the devices involved. When baselines are off, FSP payments can be misaligned, which lowers the appeal of providing flexibility and reduces settlement accuracy. Installing higher-granularity meters and adopting a standardised baseline method would improve reliability and help ensure fair compensation.
Fragmented financial incentives for smart appliances: Across the Nordic countries, national support schemes for smart-energy devices differ in size, eligibility rules, and application procedures. This patchwork leads to uneven deployment of controllable appliances, so the pool of flexible loads grows at different speeds in each market. For FSPs that operate cross-border, the lack of a consistent incentive landscape complicates scaling their aggregation and service models. A clearer, more harmonised set of financial measures would encourage households to adopt smart appliances at a similar pace everywhere and thus make it easier for FSPs to expand region-wide.
DSO revenue regulation discourages flexibility: At the Market-design and Procurement stages (steps 3 and 4) of the flexibility value chain, the current revenue-cap model ties a DSO’s allowed income mainly to its asset base, so grid investments boost revenues more reliably than operational expenditures such as procuring flexibility. Interviewees therefore feel the framework gives DSOs little motivation to weigh demand-side flexibility as an alternative to grid investments.
Household appliances as resources for DSF
In Norflex, the primary household flexibility resource was private EV chargers. FSPs integrated the EVs with their own cloud-based data hubs. Similar set-ups are being used by other technical aggregators in Euroflex to aggregate flexibility from more EV chargers, as well as other household appliances such as heat pumps and heat panels.
Any device that can receive and respond to external signals can, in theory, be operated flexibly. Most electric household appliances sold over the past ten years can communicate with external systems through WiFi, 4G, or radio signals. Appliances can also be made flexible by the installation of a local hardware device that is directly connected to the appliance in question and acts as an intermediate between the appliance and the external system.
Most appliances were integrated with the FSPs' data hubs through cloud-to-cloud integration. The FSPs are responsible for setting up the integration, which happens over various APIs. The interviewees remark that even though some API integrations are challenging, once it is completed, the logic can be reused. Therefore, finding solutions to integrate new appliances into their platforms is seen as part of their value proposition. Since the appliances and business models are still under development, standardisation of APIs and communication protocols is perceived as limiting further innovations instead of being helpful. On the other hand, the FSPs mentioned that some APIs are too slow for the appliances to participate in fast-response markets, which may potentially become an issue in the future.
In the data hub, consumption forecasts for both individual customers and the FSPs’ customer portfolio are made with AI based on a combination of historic consumption data, customer settings, and weather forecasts. These forecasts are used by the FSPs to plan their trading activities.
Once integrated, the customer can give the FSP permission to control the loads based on various signals through their app solution. For EV charging, for instance, the customers can specify that they want their car charged to X% within a certain time. Then, day-ahead prices are used to plan the charging cycle. If there is further room to manage the charging cycle, the FSP can trade the customers’ flexibility in other markets. For most customers, the largest part of their flexibility is activated based on spot prices, and not price signals from the flexibility markets.
EV charging as a resource of DSF
EVs are well-suited to provide short-term flexibility services by rapidly starting and stopping charging according to the needs of the power grid
Flexibility from Electric Vehicles (CIRED, 2022)
In Norflex, the FSP Tibber participated with approximately 1500 EV chargers
Flexibility from Electric Vehicles (CIRED, 2022)
One observation made about flexible EV charging in both Norflex and other projects is that it can pose challenges to the grid if not managed properly. If a lot of EVs start and stop their charging simultaneously from one hour to another, it may cause large voltage fluctuations locally
Flexibility from Electric Vehicles (CIRED, 2022)
DSF from other household appliances
Though not the focus in Norflex, other household appliances are being set up to deliver flexibility services to the grid in Euroflex, by both Tibber and other household FSPs.
According to another FSP, an average household can deliver 7.5 kW of flexibility, of which most comes from smart EV charging. Other smart appliances include heat panels, ESWHs, and underfloor heating. These appliances can be integrated using similar logic as the ones developed for EV smart charging. The flexibility potential of these loads is summarised in Table 8. The potential, especially the thermal flexibility, is dependent on thermal characteristics of the house, installation topology, and, most importantly, the comfort preferences of the residents.
Table 13: Flexibility potential from other household appliances
Based on information provided by interviewed FSPs.
Appliance | Capacity, kW | Response time, seconds | Max activation duration, hours | Minimum recovery time, hours |
ESWH | 2.5 kW | < 10 s | 4-6 h | 1-2 h |
Heat panel | 0.7 kW | < 60 s | 1 h | 1 h |
Heating cables | 1 kW | < 30 s | 3 h | 1 h |
Other appliances | 3 kW | < 30 s | - | - |
DSF potential from EV charging in Norway
As newer EVs can be smartly charged, a significant portion of the Norwegian EV park of nearly 790,000 cars has the potential to deliver DSF services to the grid
Depending on the EV and charging box, most EVs charge with a capacity of 3.7 – 22 kW, and need, on average, 2 to 10 hours to be fully charged
Sweden: Sthlmflex
General description
In 2018, the Greater Stockholm region had been facing local and regional grid challenges for years, with more connection requests than grid expansion alone could handle. Also, since a tax on combined heating and power (CHP) was announced to be introduced in 2019, a higher electric heating demand was projected as district heating prices were expected to increase.
Sthlmflex was thus established as a local flexibility market to deal with these challenges and set up as a collaboration project with Swedish TSO Svenska Kraftnät as project owner and the DSOs Ellevio and Vattenfall (later joined by DSO E.ON) as participants. The concrete issue the participants hoped to address was the risk of exceeding power subscriptions
In Sweden, every customer to the national grid signs a usage agreement, which governs how much transmission capacity the customer subscribes to. This agreement states the payable fee according to the grid tariff's capacity fee. On top, it is possible to apply for a temporary subscription, which is added to the annual subscription and applies to one single week.
Sthlmflex was operative starting in 2020. In 2022/23, the last active season as a research project, 4.641 flexibility resources participated through 10 providers. Of these, around 95% of resources were privately owned, mostly EV chargers.
Project results
An important result of the pilot market was that the need for flexibility is strongly dependent on weather conditions, with ambient temperature being the most important contributor, especially when these lasted for longer periods of time. The winter of 2020/21 was relatively mild in the Stockholm area, resulting in fewer call-offs than initially anticipated. On top of that, external factors like the spike in electricity prices caused by the preparation and execution of Russia’s full-scale invasion of Ukraine decreased the baseline consumption and thus the need for flexibility further
Another learning point was that all flexibility purchased by the DSOs had not been delivered, raising questions about the reliability of FSPs and the proper ways to validate deliveries.
In the end, the market was discontinued, mainly due to low liquidity. While the need for flexibility is still present and growing, the DSOs viewed other flexibility measures as more promising, such as bilateral contracts or conditional connections in the near term.
Value chain for DSF
Sthlmflex was conceived in part as a research project to demonstrate the different steps to be taken for flexibility deliveries. However, the project was also designed to solve a real and pressing underlying grid issue in a transition period until physical grid reinforcements came in place, thus covering steps 2-8 in the value chain described in Chapter 3.1.
Table 14: Flexibility value chain for DSF in Sthlmflex
The market platform used was NODES, with their products “LongFlex” (risk handling with long time horizon, not unlike bilateral contracts), “ShortFlex” (“free bids” used for short-term cost optimisation and “VeckoFlex/ShortFlex Availability” (2-3 days’ time horizon, including an availability fee). The product characteristics are displayed in Table 9.
Table 15: Product characteristics in Sthlmflex
Short Flex | LongFlex | VeckoFlex/ ShortFlex availability | |
Endurance | 60 minutes | ||
Market close | 2h before delivery | - | - |
Lowest bid | 0,1 MW | ||
Granularity | 0,01 MW (2022/23) | ||
Payment | Activation Pay-as-bid | Activation Pay-as-bid Availability Pay-as-bid | Activation Pay-as-bid Availability Pay-as-bid |
Availability requirements | During hours covered by bids | Days with -5°C or colder | Predefined number of hours per week |
Case study description: heat pumps as a flexibility source
The analysed case study consisted of two heat pumps in office buildings in Stockholm. They are owned and operated by Vasakronan, one of Sweden’s largest property owners. During Sthlmflex, Vasakronan used both heat pumps, car chargers, and ventilation as flexibility resources. NODES was used as trading platform, and communication with the heat pumps’ control units was handled by FlexTools. ProptechOS was utilised to model and aggregate several buildings for participation in the marketplace.
As a general result, less trading than expected occurred on the market, mainly caused by high electricity prices, which decreased the overall need for flexibility. More results are incorporated in the Section “Barriers and enablers”.
Overall potential of heat pumps as a flexibility source
Especially for single-family dwellings, heat pumps have a high market share in Sweden. According to Energimyndigheten’s statistics database, around 60 % of these houses are heated by different kinds of heat pumps. Table 10 shows the numbers for different kinds of heat pumps and building types.
Table 16: Number of installed heat pumps in Sweden
Number of installed heat pumps (1000s, 2023) | Ground source | Air-to-water | Air-to-air | SUM |
Single-family dwellings | 478 | 443 | 745 | 1667 |
Apartment buildings | 20 | 15 | 2 | 37 |
Commercial | 13 | 7 | 8 | 28 |
Total | 510 | 466 | 756 | 1732 |
Source: Energimyndigheten 2024: Ny energistatistik för byggnader
Assuming an average electric power input of 4 kW for smaller and 20 kW for larger heat pumps, this would yield a theoretical potential of 8 GW flexibility from heat pumps if all were available as FSPs. As most heat pumps to date do not fulfil the requirements to be used as FSP (mainly due to a lack of connectivity, which in part could be solved by retrofitting), and because not all heat pumps are situated in areas with a need for local flexibility, this number would have to be reduced in reality. Also, not all heat pumps in the country will always operate at nominal power at the same time, which further reduces the potential.
In a recent publication
Flexibilitetspotentialer till år 2030, Power Circle
Summing up, the total flexibility potential from heat pumps in Sweden can be assumed to be several gigawatts, with the spatial distribution in the country and the ability to retrofit remote control and measurement technologies as important issues to take into consideration.
Barriers and enablers
Enablers:
- High market penetration of heat pumps in Sweden enables considerable potential for demand flexibility, which can be unlocked by new standards when upgrading
- Demand characteristics: Heat demand dominated Sthlmflex and is easy to predict. Different demand profiles in different appliances can be complementary: heat pumps have a different flexibility profile compared to EV chargers or other appliances.
- Activation times: In local markets, the use of resources with longer activation times compared to balancing markets is possible since the market rules on activation times are usually less strict.
- Reusable integrations: Digitalising legacy equipment is costly, but the installed hardware and implemented integrations can still be used for other marketplaces or services.
- Tight communication between DSO and aggregator when building integrations prevents misunderstandings and redundant work.
- Green leases enable landlords to vary temperatures within certain limits.
Barriers:
- High entry threshold: The setup cost for administration and integration can exceed the benefits if the flexibility demand (and thus market liquidity) is low.
- Future power tariffs may promote a flatter power consumption curve. Since flexibility resources such as heat pumps or batteries need to run on higher power before and/or after having delivered flexibility, their power profile becomes more volatile when they are used flexibly. This volatility results in higher power peaks, and the cost incurred by these in a power tariff can decrease the profitability of using the resource flexibly.
- Low temporary subscription prices: In Sweden, DSOs and regional grid companies can apply for a temporary subscription to the TSO’s grid, which is added to the annual subscription, see section 3.6.1. In the case of Sthlmflex, the prices for these subscriptions were often lower than the flexibility bids, which were thus outcompeted. It should be mentioned, however, that temporary subscriptions are not guaranteed and thus cannot be seen as adequate replacements for flexibility.
- Low incentives to participate in flex markets: The Swedish regulation on the revenue of grid operators (Intäktsregleringen) defines the economic framework for investments in electricity grids and defines a cap for DSOs’ revenue. In connection with this, the Swedish Electricity Act (Ellagen) states that the revenue cap should take into account the extent to which flexibility services are used and improve the efficiency of network operations. However, interviewees expressed that this definition is insufficient and that making investments in physical grids is still considered more economically appealing.
- Complicated prequalification process:
- Prequalifications for both regional and local grids.
- Baseline- and measurement problems: Different appliance types feature different power consumption profiles and follow more foreseeable or completely stochastic patterns, which makes it harder to fit pre-defined baselines as suggested by market platform operators for certain appliances.
- Data availability and ownership: in Sthlmflex, the grid companies (rather than the FSPs) communicated with NODES and thus needed power-of-attorney for every resource to relay measurement and verification data. This added an administrative burden and led to discussions about what the recorded data can and cannot be used for
- Information disadvantage for actors other than the DSOs, who don’t have access to data such as customers’ power demand profiles for baseline optimisation through other channels, since there is no data hub
- Multiple communication interfaces and integrations towards different proprietary systems in appliances.
- High participation cost per unit for appliances, mainly due to requirements for individual measurement per unit, especially when no measurement hardware is present in the existing hardware.
- Endurance was more important for local markets than TSOs’ flex markets: In a cold wave, flexibility from heat pumps and other heaters may be used up after day 1, when room temperature moves below an acceptable threshold. Without reheating the building in between (e.g., when the grid situation relaxes during nighttime), flexibility cannot be delivered after that time.
- Low baseline consumption in modern office buildings: Newer office buildings typically adhere to stricter standards than legacy buildings, meaning that heating, ventilation, and air conditioning are optimised towards low energy consumption. This results in a comparably low amount of flexibility landlords can offer from appliances in these buildings.
Sweden: Effekthandel Väst
General description
Effekthandel Väst is a local flexibility market in the Gothenburg region, established to address the region’s future electricity demand and grid capacity challenges. On this market, DSOs Göteborg Energi and Mölndal Energi purchase flexibility services from connected customers. Participants in Effekthandel Väst include industrial enterprises, small companies, and private persons via aggregators, and large property owners. Among the flexibility sources traded are batteries, EV chargers, and fridges/freezers in supermarkets and warehouses.
Since NODES is used as trading platform, the same products are traded as in Sthlmflex (ShortFlex/LongFlex), except for the new product ”MaxUsage”, which limits power use and thus circumvents the need for a baseline calculation, with lower administration efforts as a consequence. Effekthandel Väst is active during winter (November 1st to March 31st), when electricity demand is highest. Minimum endurance is 1 hour, and the lowest bid size is 50 kW (single source or aggregated).
Learnings from other markets
Before starting the market, both CoordiNet and Sthlmflex were analysed. A key learning was that a market needs a sufficient number of trades to be interesting to market actors for them to invest in both technology, routines, and competence building. This was especially important since the market needed to work without external research funding. To achieve this, direct communication lines with potential customers were established, rather than choosing a top-down approach where mainly the needs of the DSOs were communicated publicly and widely. Also, the pricing scheme was designed to be more attractive from the start, with a price that considered the alternative cost for flexibility on the customer side and prices in competing markets (e.g., balancing markets). Purchasing flexibility was prioritised over other possible solutions to solve the capacity problem. This leads to more flexibility being purchased than strictly necessary and stimulates market activity.
Since Effekthandel Väst, unlike Sthlmflex, only needs to consider the local grid (and not the regional grid), the extra administrative burden for filling in power-of-attorneys can be eliminated. Also, fewer different kinds of appliances are currently in use on the market, which enables a slimmer prequalification process of flexibility resources.
Future development
Göteborg Energi is determined to expand the market in the coming years and believes that flexibility will be an important tool to handle the challenges its grid will need to handle in the future, alongside grid upgrades. The total estimated need for flexibility is about 100 MW. Also, new forms of flexibility resources are planned to be integrated with the market, mainly larger battery storage and, when available, aggregated electric cars via V2G.
Value chain for DSF
As mentioned above, Göteborg Energi and Mölndal Energi foresee a growing demand for flexibility in their grids in the future and plan the grids accordingly. Since flexibility is considered an integral part of handling the grid in the future, and because the market can be seen as mature after several operating seasons, all steps 1-8 in the value chain defined in Chapter 3.1 are covered by the project.
Table 17: Physical value chain for DSF in Effekthandel Väst
Systems in use
As mentioned above, NODES is used as market platform, with an API that can be accessed for all relevant market activity, such as registering resources, placing bids, or sending baselines and measurement data. For validation, the FSP can choose to upload their own data to NODES or let the DSO use the electric meter data directly. There are no specific requirements for communication protocol standards because the DSO neither communicates directly with nor controls the flexibility resource.
Instead, all communication takes place via the market platform. Buying and selling bids are matched, and the winning flexibility provider either receives an SMS containing all necessary details on the trade or chooses to integrate with the API.
Resources
The DSOs running Effekthandel Väst do not prescribe which types of resources can be used on the flexibility market, as long as they fulfil the requirements stated above.
In total, around 30 FSPs managing roughly 700 flexibility resources provided 30 MW of available flexibility in 2023/2024. On the appliance side, EV chargers and non-residential fridges/freezers can be mentioned, with heat pumps being added at a later stage.
The minimum bid size for the local market is 50 kW, which means that most resources need to go through an aggregator to participate. Minimum endurance is set to 1 hour. Separate measurements on the appliance level are not a requirement today, but this will probably change in the future. In the following, a case study for flexibility from freezers and fridges in supermarkets is described.
Case study description: fridges and freezers in supermarkets
Ambidex was founded in 2023 with a business idea founded upon research from Chalmers University of Technology. The company installs control modules in the PACs (Programmable Automation Controllers) of supermarkets’ cooling systems that both monitor and control the refrigeration system’s power consumption. Communication usually takes place via the open Modbus/Modbus-TCP protocols.
In Effekthandel Väst, Ambidex also acts as an aggregator, bundling together several supermarkets in the Gothenburg region in different portfolios which can be activated separately or together depending on where and how much the grid needs to be supported, and handling the integration towards NODES via an API. On the marketplace, the aggregated resources are traded in the LongFlex market (see Chapter 3.6). A typical supermarket in the portfolio can have an installed electrical capacity of around 150 kW, of which 50 kW are potentially flexible for longer periods of time, and 35 kW effectively brought to market.
Before being eligible to participate in Effekthandel Väst, the resources need to be prequalified, which is done by an activation test and reporting of the results, a simpler procedure as compared to the Swedish TSOs’ ancillary service markets.
The NODES default baseline calculation only considers the last 4 weeks’ power consumption. As fridges and freezers feature a power consumption profile that, among others, factors change with ambient temperature, week, day, and time of day, the match between the default and actual baselines is imperfect. Part of the integration into the market platform was thus to negotiate an adapted baseline procedure based on Ambidex’s data, which takes the flexibility resources’ unique consumption profile into account. This enabled Ambidex to deliver more flexibility than had been possible otherwise. In the future, Ambidex will look into both additional markets such as SvK’s ancillary service markets
In the interview, it was pointed out that a sufficiently low response time to meet these services’ requirements was possible to reach from a technological point of view.
Overall potential of fridges and freezers as a flexibility source
One way of calculating the total flexibility potential, i.e., the sum of implicit and explicit flexibility, from fridges and freezers is to create an annual consumption profile based on previous data, which leads to the assumptions that cooling demand is 1.4 times higher during high-demand hours (07-19) than at night and that the total annual electricity consumption lies around 575 GWh for cooling of goods in the service sector.
2021 numbers; Method and numerical results are presented in Energimarknadsinspektionens report Främjande av ett mer flexibelt elsystem
The subsequent calculation yields a total flexibility potential of 75 MW for Sweden, with an endurance of 1h and a recovery period of 2h. It is worth noting that the same report quotes a potential of 564 MW in residential white goods, i.e., fridges/freezers and washing machines/tumble dryers in households.
Another article
Funder, T. (2015). Supermarkets as an Important Smart Grid Application. In 16th European Conference, Technological Innovations In Refrigeration And In Air Conditioning
Barriers and enablers
The following enablers and barriers were identified, both from literature and from interviews with DSOs, technology providers, and aggregators.
Enablers:
- Implementation of the demand response network code (see Chapter 4), which is set to standardise market-specific parameters such as bid sizes, call-off periods, and communication protocols between different local markets. The main points interviewees want to see answered by rules in the demand response network code are standardised communication protocols, prequalification rules, validation and payment solutions, and harmonised baseline registration methods.
- Local gateways owned and operated by the aggregator relay information from and to the flexibility resource and ensure secure data transmission.
- Flexible baseline definitions: The possibility to discuss the definition of appliance-specific baselines with the market platform operators and agree on said baselines is seen as a huge advantage; without this possibility, it is questionable if fridges and freezers could have been brought to the local flexibility market while maintaining economic profitability.
- Spatial distribution of supermarkets in the grid: Unlike large industrial flexibility resources, supermarkets exist in most places and are correlated with population density, which in turn can drive the need for flexibility. Fridges and freezers are present in most of these supermarkets. Thus, the even distribution of these resources aligns well with the need to address grid issues locally.
- Existing communication protocols such as Modbus (as standardised in IEC 61158) enable quick integration towards different builds and generations of fridges and freezers, or their control systems, respectively. In Sweden, this is enhanced by a high market share of control systems from the same company (Danfoss). A future research question might be whether the current level of standardisation is sufficient for this application or not. If not, a solution might be to amend initiatives like the Code of Conduct for Energy Smart Appliances (see chapter 4.3.2) to include industrial-size freezers and fridges as smart appliances.
- APIs: When appliances are aggregated, manual activation of flexibility resources is not an alternative. The presence of an API, as when integrating with NODES, ensures that activations and market transactions can be automated.
Barriers:
- Food hygiene and respective legislation: temperatures in fridges and freezers are heavily controlled by food hygiene rules. When used as a flexibility resource, these appliances will not be able to deliver in situations where decreasing their power intake could mean a risk of exceeding temperature thresholds. In the current case study, this was handled as previously described by fine-tuning the resources’ baselines, meaning that no deliveries had to be cancelled.
- Missing communication standards for market platforms: As an aggregator, being present on different local flexibility markets with different rules and market platforms usually means high integration costs, which do not benefit from economies of scale. The definition of common standards could lower the threshold for participation in multiple markets. However, when defining these standards, it is important to ensure enough flexibility in questions such as baseline definition, see section on enablers.
- Market liquidity was also mentioned as a potential issue for Effekthandel Väst, although liquidity is higher than it was for Sthlmflex. For details, see the chapter on Sthlmflex.
- Limited testing infrastructure towards market platform: As of now, aggregators cannot test their entire code up to a bid being called, but have to rely on NODES’ test call. An extended test infrastructure could help avoid this problem.
- The Swedish regulation on the revenue of grid operators (Intäktsregleringen) defines the economic framework for investments in electricity grids and defines a cap for DSOs’ revenue. As was the case for Sthlmflex, interviewees in this case study also expressed the view that making investments in physical grids is more economically appealing than investing in flexibility measures.
- Lack of expertise: Resource owners, in this case owners of supermarkets, are often unaware of their own ability to participate in flexibility markets. Targeted information campaigns could be useful in tapping this potential.
- Non-aligned trading deadlines on different markets: Ancillary markets and local flexibility markets are not always synchronised in their closing times, which is especially important when bids from the local markets shall be forwarded to an ancillary market.
- Power tariffs: As described in the chapter on Sthlmflex, power tariffs can hinder the roll-out of flexibility markets since the power consumption profile will become more volatile. On top of that, interviewees pointed out that there are a lot of different power tariffs present in Sweden, since the DSOs have some freedom in designing them, and that there is no easily automated way to import data on the different tariffs. Keeping the optimisation algorithms for flexibility resources up to date is therefore a challenge. Dynamic power tariffs are being developed in research projects right now and will probably solve this problem.
- DSOs not favouring market solutions over bilateral contracts: DSOs might be overcautious and instead of implementing flexibility markets, cover their flexibility needs by bilateral contracts with single FSPs, which excludes a large number of potential FSPs.
Sweden: E.ON Energidistribution’s flexibility markets
E.ON Energidistribution has opened different commercial markets in its grid areas in Sweden. In order to analyse what the drivers behind this development were, a general description of the company’s flexibility markets, rather than one specific case study description, is performed in this chapter. The potential of single appliances on a national scale was therefore not determined. The information is based on publicly available sources, a stakeholder webinar
“E.ON Energidistributions flexibilitetsmarknader”, September 12th, 2024
General description
The German energy company E.ON entered the Swedish market as a DSO in the early 2000s by acquiring Sydkraft, which itself had previously bought several Swedish DSOs. Today, E.ON Energidistribution owns and operates 70 local grids in 139 municipalities, with a focus on southern and middle Sweden. Most of the company’s grids are situated in grid areas SE3 and SE4. The grid situation in these southern Swedish areas is challenging at times due to an increased electricity demand. Also, they feature a higher degree of intermittent power production and lower available flexibility potential from large primary industrial sites as compared to the northern Swedish grid areas SE1 and SE2. Also, E.ON Energidistribution’s grid tariffs are in the higher range of the Swedish DSOs. This is in part caused by the long distances between large-scale power generation in northern Sweden and the company’s grids in southern Sweden, and the losses resulting from the required long-range electricity transmission. To enable customers to connect faster and use the grid more efficiently, which can also keep down costs for electricity and tariffs for the customers, E.ON Energidistribution works strategically with both implicit and explicit flexibility. The work on explicit flexibility includes the setup of flexibility markets, whereas implicit flexibility is encouraged by flexibility mappings, which unveil existing flexibility potential for companies and property owners and help them to lower both electricity and grid tariff costs.
Local flexibility markets
Based on learnings from the EU-funded project CoordiNet, E.ON Energidistribution established its first local flexibility markets in 2023 in Bålsta, Vaxholm, Hässleholm, and Southern Skåne
Southern Skåne, Bålsta and Vaxholm had previously been run as a part of CoordiNet since 2019/2020
See network development plan: https://www.eon.se/content/dam/eon-se/swe-documents/swe-eon-natutvecklingsplan-2025-2034.pdf
E.ON Energidistribution addresses both homeowners (via aggregators), property owners, municipalities, and industrial/commercial actors as participants in their local markets. The tech-neutral market approach enables resources such as heat pumps, car chargers, battery storage, ventilation units, backup generators, and lighting systems to participate.
As a market platform, but also as a planning tool for both FSPs and DSOs, SWITCH, which was developed within CoordiNet and is owned by E.ON Energidistribution, is used. Three products are traded:
- “Direktorder” (direct orders) with a pay-as-bid activation fee
- “Tillgänglighetsorder” (availability orders), which adds a fixed availability fee and
- “Säsongstillgänglighet” (seasonal availability), which is comparable to availability orders, but focuses on longer time spans. The smallest bid is limited to 100kW, and an endurance of 1 hour needs to be guaranteed.
If the FSP fails to deliver at least 75% of their bid, no payment is made. Above 75%, the payment scales linearly with the percentage of delivery from 75 to 100%.
Although some resources are aggregated, a handful of industrial and commercial customers are big enough to participate on their own, their size ranging from the minimum bid size of 0.1MW up to 15MW, where the larger resources could, for example, be industrial-size heat pumps or steam turbines.
Value chain for DSF
E.ON Energidistribution expands the number of local flexibility markets and runs them on a commercial basis. Since the company considers flexibility an integral part of handling the grid in the future, and because the markets can be seen as mature after several operating seasons, all steps 1-8 in the value chain defined in Chapter 3.1 are covered by the market.
Table 18: Physical value chain for DSF in E.ON's flexibility markets
Learnings and status of new markets:
While E.ON Energidistribution participated in both CoordiNet and Sthlmflex, the management of their flexibility markets has moved from a project form to be handled in the company’s line organisation since 2022, meaning that the same processes and guidelines are used throughout the organisation. Among other things, this means that the definition of baselines and production plans, prequalification of equipment, and sanctioning of failure to deliver are handled more strictly now than in previous research projects, which is reflected by new and uniform market design and rules.
The opening of a new market is preceded by an analysis of whether a local flexibility market can solve the grid issues at hand. E.ON Energidistribution’s network development plan estimates the total flexibility need, and a team analyses whether a local flexibility market or other solutions are the most cost-effective tools to address the grid issues.
When scaling up from single projects to a larger number of markets, automation plays a crucial role since manual interventions scale badly. An automated calculation and publication of flexibility needs and the possibility for FSPs to integrate towards an API have therefore been a focus of development work and resulted in fewer manual interventions than in the project phase. Concerning the Switch platform, an iterative product definition and more accurate prediction algorithms have led to better demand forecasts, higher trust in flexibility deliveries, and smoother validation processes. Random samples are analysed to ensure that the reported baseline is in accordance with the actual delivery. The service provider also provides meter data, which can either be based on the billing meter or a submeter directly metering the flexibility resource. For aggregated resources, the aggregator is responsible for sending measurements for the whole group of resources.
Barriers and enablers
In line with the other Swedish case studies, interviewees pointed out liquidity issues, regulation on the revenue cap of grid operators, missing communication standards for market platforms, and a lack of expertise as barriers. Since these points have already been analysed in Chapters 3.7 and 3.8, they are not mentioned separately in this section.
Enablers:
- Aggregators: E.ON Energidistribution describes the aggregator’s role as crucial for successful flexibility markets. In part, aggregators own the know-how regarding integration both towards different market platforms and a wide range of appliances, often using their own hardware for measurement and communication. As described in the case study for Sthlmflex and Effekthandel Väst, aggregators also play a fundamental role in defining viable baselines, unlocking the flexibility potential a resource might have. A steep learning curve concerning baseline procedures has led to better matches between baselines, flexibility demand, and validated deliveries.
- Clear and consistent rules: Market rules, but also policies, laws, and regulations that shift quickly and unpredictably, can hinder the implementation of local flexibility markets because both DSOs and FSPs will shy away from necessary investments in an uncertain situation. E.ON Energidistribution has therefore tried to standardise the rulebook between their individual markets and stresses the importance that the same is done on a national and European level, e.g., by swift implementation of the Network Code on Demand Response (see chapter 4)
- Long-term planning certainty: When starting up a new market, E.ON Energidistribution states a minimum number of years the market will be guaranteed to operate alongside minimum availability payment levels. This reduces planning uncertainty for FSPs and lowers the threshold for market participation.
Barriers:
Political decisions on a municipal level might hinder certain flexibility resources from being brought to the market. A notable example mentioned in the interviews is power generators, which technically could be used as flexibility resources, but are not allowed to run on fossil fuels other than in crises due to guidelines the owner (a municipality) defined.
Long ramp-up to build markets: In the more mature local flexibility markets managed by E.ON Energidistribution, market liquidity was an issue at first, but became less problematic after the first couple of operating seasons. Ensuring market liquidity from the start should therefore be a focus for further market design efforts. Also, the aforementioned lack of competence falls in the category of ramp-up issues.
No intrinsic motivation for local flexibility markets: As opposed to the TSO’s ancillary markets, local flexibility markets are not an exclusive solution: Whereas ancillary services are critical to the operation of an electrical grid, and cannot be easily replaced by other measures, local flexibility markets need to justify their existence in competition with other solutions to grid issues such as bilateral contracts or conditional connections.
Inflexible leases: Many lease conditions for commercial buildings contain service descriptions defining acceptable windows for office temperatures or ventilation system operating times. These conditions can prevent the landlord from using the HVAC system flexibly, since this usually means that a wider temperature corridor needs to be accepted. A solution to this can be found in so-called green leases, as used by Vasakronan, see Chapter 3.6.
Appliances’ warranty conditions: Appliance manufacturers might add terms and conditions forbidding the end user from controlling the appliance by using external signals, with the threat of voiding the warranty if this is ignored. Warranty conditions should therefore be written in such a way that allows for flexible use.
Internal barriers: External factors, like the discontinuation of other local flexibility markets or the recent drop in prices for ancillary services, can lead to internal resistance against this solution, both in DSOs and potential FSPs.
Denmark: FUSE
General case description
The FUSE project was initiated in 2020 as part of Denmark’s efforts to integrate smart EV charging into the energy system, with funding from EUDP (2020-I). The project was designed as a three-year initiative, running from 2020 to 2023, with a focus on real-world testing and market integration of EV flexibility. The project is aimed at integrating smart charging of electric vehicles (EVs) as a flexibility resource in the power grid. FUSE was deployed in Frederiksberg Municipality, a dense urban area with high EV adoption and limited space for charging infrastructure. The project specifically explores how dynamic charging management can provide DSF and help stabilise the distribution grid, particularly in urban environments.
Objectives
The increasing adoption of EVs introduces significant new electricity demand, often concentrated in specific time periods, such as evening hours when people return home from work. Without coordination, uncontrolled charging could lead to localised grid congestion and increased peak loads. To mitigate this, FUSE investigates how flexible charging strategies can:
- Shift charging to off-peak periods, aligning with lower grid load and higher renewable energy production.
- Enable EVs to act as a controllable load, responding to price signals and grid conditions.
- Integrate EV charging into local flexibility markets, allowing it to compete with other distributed energy resources.
The project explored the flexibility potential from various charging settings, including:
- Private home charging: Studying how residential users respond to dynamic pricing.
- Workplace charging: Evaluating how daytime charging can provide additional flexibility.
- Semi-public chargers: Testing flexibility integration from shared EV charging stations.
Technical and market mechanisms
FUSE employs a multi-layered approach to studying and implementing smart EV charging:
- Dynamic Charging Management: Charging power is regulated in real-time to prevent grid overload and optimise energy use.
- Aggregator-Based Control: A centralised control system manages multiple EV chargers, ensuring coordinated demand response.
- Integration with Spot Prices: EV charging is optimised based on day-ahead electricity prices, ensuring cost-effective charging for users.
- Grid Load Balancing: Charging is dynamically adjusted to prevent local distribution grid bottlenecks, particularly in areas with high EV density
Stakeholders and participants
FUSE was conducted as a collaborative project involving stakeholders from grid operators, research institutions, mobility providers, and policymakers.
- DTU Wind and Energy Systems (Project Leader)
Conducted grid impact analysis and demand modelling. - Radius Elnet (DSO)
Assessed grid constraints and charging flexibility integration. - Danish e-Mobility
Worked on policy recommendations and industry dissemination. - Spirii (Aggregator and Charge Point Operator)
Managed dynamic charging and real-time market integration. - Municipality of Frederiksberg
Provided test areas and urban mobility insights.
In preparation for this report, several stakeholders were contacted for supplementary interviews to qualify the findings and provide further context to the Danish cases. While no responses were received from most of the stakeholders, partial interviews were successfully conducted with Danish e-Mobility on 26 November 2024, as well as with DTU Elektro, which served as the primary project lead. These inputs are reflected in the references and analysis where relevant.
Value chain
The FUSE project aims to integrate smart charging infrastructure into the electricity system by enabling dynamic load management and leveraging demand-side flexibility. The value chain in FUSE consists of several components, including electric vehicle (EV) charging stations, energy management platforms, grid operators, and market mechanisms that facilitate flexibility services.
Table 19: Physical value chain for DSF in FUSE
Systems in use
The FUSE project utilises multiple systems to enable demand-side flexibility through smart charging infrastructure. Unlike a dedicated market platform, FUSE relies on existing market mechanisms and infrastructure to manage flexibility.
Grid Operations and Monitoring: Radius Elnet assesses grid conditions and determines when flexibility is needed to avoid local congestion issues.
Planning and Forecasting Tools: DTU Wind & Energy Systems' Charging Point Calculator models EV charging flexibility potential and forecasts grid impact based on usage patterns.
Aggregation and Control Systems: Spirii serves as the flexibility aggregator, coordinating charging schedules based on grid demand and price signals from the existing market structures.
End-User Appliances: Smart chargers at homes, workplaces, and semi-public locations dynamically adjust charging power based on real-time signals.
All communication between these elements follows standardised protocols to ensure interoperability. Aggregators interact with charging infrastructure through APIs and cloud-based platforms, enabling remote optimisation of charging behaviour.
Resources
The flexibility resources in FUSE consist primarily of EV chargers deployed at various locations. These resources contribute to grid flexibility by shifting charging demand to off-peak periods. Unlike market-based flexibility programs, where resources are traded, FUSE leverages demand-side management to optimise energy usage for consumers rather than generate direct financial revenue.
Flexibility Service Providers (FSPs): Spirii aggregates and optimises flexibility without introducing new financial transactions on dedicated platforms.
End-User Participation: Consumers benefit from lower electricity costs due to optimised charging, but do not actively trade flexibility as a revenue source.
Data and Control Infrastructure: EV chargers are connected to management platforms that allow automated control and optimisation based on price signals and grid conditions.
The approach in FUSE focuses on cost savings through smart energy management rather than trading flexibility as a market product.
Learnings from the FUSE project
The FUSE project has provided key insights into the practical implementation of demand-side flexibility through smart EV charging. One major learning point is the technical feasibility of shifting charging demand to align with grid constraints without impacting user convenience
Dansk ELBIL ALLIANCE, DTU - Smart fra start Anbefaling til fremtidens ladestandere
Flexibility capacity (MW) = Average flexible capacity per unit * Number of units * Utilisation factor
Assuming 900,000 residential charging points, each with 11 kW charging capacity (as specified in the report), and using simulated charging profiles from Table 3 in Section 2.4, approximately 40% of charging occurs during peak hours (17:00–20:00) without intervention. If a conservative utilisation factor of 25% is applied—representing the share of chargers in use and technically controllable during peak windows—the potential for flexible load shifting is:
900.000 x 11kW x 40% x 25% = 900 MW. This suggests that nearly 1 GW of demand could theoretically be shifted away from peak hours under optimal—but realistic—conditions. Such a level of flexibility could significantly ease grid congestion, particularly in residential areas with high EV adoption, and contribute to deferring costly grid reinforcements.
FUSE has demonstrated that demand-side flexibility from EV charging is technically feasible and, when effectively managed, contributes to reducing grid congestion. Further refinement of interoperability standards, addressing regulatory barriers, and improving real-time measurement capabilities remain key areas for future research and development.
Barriers and enablers
Regulatory challenges: The project has also identified key challenges in regulatory frameworks
Dansk ELBIL ALLIANCE, DTU - Smart fra start Anbefaling til fremtidens ladestandere
Communication and data integration issues: Ensuring real-time interoperability between chargers, aggregators, and grid operators has been a notable challenge. Standardised protocols are required to maintain effective communication between different system components
Dansk ELBIL ALLIANCE, DTU - Fremtidens ladestander - Smart fra start - Anbefalinger til den smarte ladestander, 2020
Dansk ELBIL ALLIANCE, DTU - Fremtidens ladestander - Smart fra start - Anbefalinger til den smarte ladestander, 2020
Measurement and verification constraints: Reliable measurement and verification are crucial for ensuring that flexibility services are effectively delivered. The project tested different metering approaches to validate load reductions. Initial findings indicate that sub-metering at the charging station level provides more accurate verification than aggregated household consumption data
Dansk ELBIL ALLIANCE, DTU - Fremtidens ladestander - Smart fra start - Anbefalinger til den smarte ladestander, 2020
Peak load reduction and capacity: FUSE demonstrated that smart charging strategies can effectively shift charging demand away from peak hours, reducing local grid strain
Dansk ELBIL ALLIANCE, DTU - Fremtidens ladestander - Smart fra start - Anbefalinger til den smarte ladestander, 2020
Denmark: EcoGrid
General case description
EcoGrid 2.0 was a large-scale research and demonstration project in Denmark that aimed at exploring and developing a market-based approach to demand-side flexibility (DSF). The project builds upon the success of the previous EcoGrid EU initiative
Objectives
The project aimed to:
- Develop and test an open flexibility market where households could sell their demand-side flexibility.
- Demonstrate automated demand response in residential homes without affecting user comfort.
- Reduce grid congestion and balance the system by integrating decentralised flexibility resources.
- Investigate business models for aggregators to bundle and trade household flexibility.
- Enhance digitalisation and forecasting through smart meters, machine learning, and data analytics.
By transitioning from a pilot project (EcoGrid EU) to a market-driven system (EcoGrid 2.0), the project sought to prove that demand-side flexibility could be a commercially viable solution for grid balancing and congestion management.
Market-based approach and aggregation model
How the Flexibility Market Worked
- Consumers' energy consumption was adjusted automatically based on price and grid signals.
- Aggregators pooled flexibility from multiple households and sold it to TSOs, DSOs, and balance-responsible parties.
- A Flexibility Clearinghouse Market (FLECH) was established, allowing both TSOs and DSOs to procure demand-side flexibility services.
The market supported two primary service types:
- Scheduled Services: Consumers agreed to shift consumption at predefined times.
- Conditional Services: Consumers provided on-demand flexibility when requested by the market.
Technical implementation and digitalisation
- Smart meters: Provided 5-minute data for accurate consumption monitoring.
- Machine learning algorithms: Predicted household flexibility potential based on historical data.
- IoT-based Home Energy Management Systems (HEMS): Automated electric heating control via cloud-based optimisation.
- Real-time market platform (IBM Bluemix): Enabled automated trading and settlement of flexibility services.
Stakeholders and participants
EcoGrid 2.0 was a collaborative effort involving multiple stakeholders from research, industry, and the energy sector.
- TSO: Energinet.
- DSO: Bornholms Energi & Forsyning.
- Aggregators:
- IBM: Developed automation tools and market integration strategies.
- Insero: Focused on consumer engagement and aggregator operation models.
- Technology & Research Partners: DTU (Technical University of Denmark), IBM, Uptime-IT, Insero, Krukow, and CBS.
- Private Consumers: 800 households participated, primarily with heat pumps and electric heating panels.
Value chain
EcoGrid 2.0 contributes steps (4)5 to 7 on the DSF value chain described in chapter 3.1, focusing on market integrations, activation, and verification. It demonstrated how household flexibility can be traded and utilised in real-time markets.
- Step 4/5: Establishing market mechanisms
- EcoGrid 2.0 developed and tested a market-based approach for demand-side flexibility.
- It included a dedicated market platform (FLECH) for aggregators to trade flexibility.
- TSOs, DSOs, and BRPs could procure flexibility services through scheduled and conditional bids.
- Step 6: Delivery of flexibility from providers
- Aggregators (IBM & Insero) managed flexibility from 800 households.
- Smart meters and automation enabled real-time load shifting of heat pumps and electric radiators.
- Consumers’ energy use was adjusted automatically based on market signals and system needs.
- Step 7: Activation, verification, and settlement
- Flexibility activations were tracked using smart meters with 5-minute resolution.
- Baseline models ensured accurate verification of delivered flexibility.
- Market settlement was automated, ensuring payments to aggregators based on actual performance.
Table 20: Physical value chain for DSF in Ecogrid 2.0
Core mechanisms
The EcoGrid 2.0 project functioned as a real-time flexibility market, where household electricity consumption—primarily from heat pumps and electric radiators—was adjusted based on system needs. The market operated on the Flexibility Clearing House (FLECH) platform, where aggregators submitted bids for flexibility, and the TSO or DSO Bornholms Energi & Forsyning procured this flexibility for balancing or congestion management purposes.
The bidding and activation process was structured in hourly cycles. Aggregators collected and forecasted available flexibility based on temperature, historical usage, and real-time consumption patterns. The market cleared within 15 minutes, selecting the most cost-efficient offers, after which flexibility activation was initiated. The cost-efficiency of an offer depended on factors such as the aggregator’s estimated cost, the expected energy rebound effect, and the location-specific needs of the TSO/DSO.
Flexibility forecasting
Aggregators used machine learning algorithms that incorporated historical household consumption data, weather conditions, and time-of-day factors to predict available flexibility. Load reduction potential varied based on outside temperature, with greater flexibility available on colder days. At -5°C, the average household could deliver 1.1 kW of load reduction, while at 0°C, this dropped to 0.8 kW
EcoGrid 2.0 - Main Results and Findings, September 2019
Consumer compensation
In the EcoGrid 2.0, consumers did not get directly compensated for the delivery of flexibility. Instead, financial benefits were realised through reduced electricity bills, as households could shift consumption away from high-price periods
ibid.
DSO-TSO market coordination
The DSO played a crucial role in determining when and where local flexibility was needed, particularly for preventing grid congestion. Unlike TSO activations, which responded to frequency deviations in real time, DSO activations were often scheduled days in advance based on congestion forecasts. This occasionally led to conflicts between TSO and DSO activations, as the same flexibility pool could be requested by both parties
ibid.
Activated flexibility volumes and market performance
Over the course of the project, aggregators initiated a total of 102,458 activations, distributed between the two main aggregators:
- Siemens: 19,682 activations
- GreenWave: 82,776 activations
Despite the high number of activations, approximately 9.7% of attempts failed, primarily due to communication errors between the platform and household devices. In terms of flexibility capacity, the participating households delivered up to 1 MW per activation, with individual heat pumps providing between 0.8 and 1.1 kW per event, depending on outdoor temperature and home insulation levels. The estimated total energy shifted during the project was over 5 GWh, with activations typically lasting 30–60 minutes
ibid.
Economic aspect
One of the key findings of EcoGrid 2.0 was that the economic incentives for household participation were limited. Consumers did not receive direct payments for their flexibility but instead saved on their electricity bills by shifting consumption away from peak-price hours. However, the average annual savings per household amounted to less than 100 DKK, which did not create a strong financial incentive. From a market operations perspective, aggregators generated revenue by trading flexibility in both TSO and DSO markets. However, price volatility and limited liquidity in the flexibility market restricted the overall profitability. The evaluation suggested that a more structured compensation model could enhance aggregator participation and market scalability
ibid.
Rebound Effects and Technical Constraints
A major challenge in EcoGrid 2.0 was the rebound effect
ibid.
Key findings on the rebound effect:
- Uncontrolled rebound effects resulted in a load increase exceeding the initial load reduction, negating part of the benefits of flexibility activations.
- Smart control algorithms were implemented in heating season 3, successfully mitigating rebound effects by gradually restoring heating instead of allowing a sudden power surge.
- Rebound effects were particularly pronounced on very cold days, where indoor temperatures dropped faster than expected, requiring more energy to restore comfort levels.
The final assessment recommended that future flexibility programs incorporate advanced load recovery strategies to minimise rebound effects, ensuring that flexibility services remain net beneficial to the grid.
Technical and infrastructure overview of EcoGrid 2.0
EcoGrid 2.0 relied on a combination of hardware, software, and communication infrastructure to facilitate real-time demand-side flexibility from residential consumers. At the heart of the system were heat pumps and electric radiators, which acted as primary flexibility resources, complemented by smart metering, cloud-based aggregation platforms, and real-time communication networks. Despite the success of the technical implementation, the project also highlighted key challenges related to device interoperability, data communication reliability, and standardisation.
Heating appliances
Heat pumps, used in 85% of participating homes, proved to be particularly effective due to their thermal storage capabilities. Depending on outdoor temperatures and insulation levels, heat pumps could deliver between 0.8 kW and 1.1 kW of flexible load per activation, providing a scalable and predictable source of demand response. Electric radiators, used in 15% of participating homes, provided an alternative flexibility source, with individual power capacity of 0.5–1.5 kW per unit. Unlike heat pumps, radiators did not benefit from thermal storage, meaning short-term flexibility activations had a more direct impact on indoor comfort. This resulted in higher override rates, where participants manually adjusted their settings to counteract flexibility activations, reducing overall effectiveness.
Metering and data collection
Participating homes were equipped with smart meters capable of recording high-resolution energy consumption data. While standard Danish smart meters provide hourly data, EcoGrid 2.0 required higher resolution (5-minute intervals) to allow real-time tracking of flexibility activation and settlement. These meters fed consumption data directly into the market platform, allowing aggregators and system operators to monitor the impact of activations and compare real-time reductions against baseline consumption models. One of the critical advantages of high-resolution metering was the ability to accurately verify delivered flexibility, ensuring that aggregators were only compensated for actual load reductions. However, the project also revealed challenges in data processing and integration, particularly in cases where communication failures resulted in missing or delayed metering data. Approximately 9.7% of flexibility activations failed due to loss of communication between smart meters and the central platform, highlighting the need for improved data reliability in future implementations.
Communication and control systems
A core component of EcoGrid 2.0’s technical implementation was the communication infrastructure that connected household appliances to aggregators and the market platform. Heating appliances were controlled via a mix of Wi-Fi, Zigbee, and proprietary cloud-based manufacturer interfaces, allowing real-time activation signals to be transmitted to and from consumers. While most modern heat pumps supported remote control, many older models lacked standardised API access, requiring aggregators to develop custom software layers to bridge these gaps.
The lack of standardised communication protocols across different manufacturers posed a significant barrier to seamless integration. While some devices used MODBUS or BACnet protocols, others relied on proprietary cloud platforms, making uniform control difficult. This fragmented technology landscape meant that aggregators had to develop multiple parallel integration methods, increasing the complexity and cost of system operation.
Key learnings from EcoGrid 2.0
EcoGrid 2.0 demonstrated the technical feasibility of household demand-side flexibility but also highlighted several challenges that need to be addressed for large-scale deployment. One of the key findings was that heat pumps and electric radiators can successfully provide flexibility, but response times and control precision varied significantly. Heat pumps exhibited delayed activation and rebound effects, making them less suitable for fast-response flexibility services without better control mechanisms
ibid.
Reliable communication infrastructure proved to be critical, yet 9.7% of flexibility activations failed due to lost signals or delayed data transmission. The use of multiple communication protocols across different appliances led to integration issues, and some devices had limited API access, requiring custom software solutions for market participation. The lack of standardisation in communication created barriers to automation, increasing the cost and complexity of aggregator operations.
From a market and economic perspective, EcoGrid 2.0 successfully demonstrated real-time flexibility trading, but low financial incentives for consumers limited engagement. Households did not receive direct payments, instead benefiting from reduced electricity costs, but with annual savings under 100 DKK per household, motivation remained low. Furthermore, the coordination between TSO and DSO activations revealed conflicts, as the same flexibility resource was sometimes needed for both system-wide balancing and local congestion management, leading to suboptimal resource allocation.
Overall, the project confirmed that household flexibility can play a role in future energy markets, but success will depend on improving device interoperability, strengthening data communication reliability, optimising financial incentives, and refining market coordination mechanisms.
Overall potential of heat pumps as a flexibility source
The project’s scalability assessment estimated that, if rolled out nationwide in Denmark, demand-side flexibility from residential heat pumps could contribute up to 200 MW of flexible capacity, mainly during winter months when heating demand is high. This represents approximately 2–3% of Denmark’s peak electricity consumption. Further analysis suggested that, under optimal market conditions, automated flexibility activation could reduce the annual peak demand by 5–8%, helping to delay grid investments and improve renewable energy integration
ibid.
Barriers and enablers
Table 21: Barriers and enablers in Ecogrid 2.0
Category | Barrier | Enabler | Potential solution |
Appliance response | Heat pumps had slow reaction times and rebound effects, limiting fast flexibility. | Thermal storage in buildings mitigated rapid temperature drops | Optimised control algorithms to gradually restore load and reduce rebound peaks |
Communication | 9.7% activation failures due to lost signals and protocol inconsistencies | Standardised APIs and improved cloud-based control | Industry-wide adoption of common communication standards to enhance compatibility |
Market and economy | Low financial incentives reduced consumer engagement | Aggregators successfully traded flexibility in real-time markets | Higher compensation schemes or dynamic pricing structures to improve consumer participation |
Coordination | TSO and DSO competed for the same flexibility, leading to inefficiencies | Flexibility successfully tested for grid balancing and congestion relief | Clearer market mechanisms for resource allocation and better coordination frameworks |
Integrations and standardisation | Lack of uniform protocols increased integration costs | Machine learning-based forecasting improved bid accuracy | Regulatory push for standardisation in device control interfaces |
Finland: Helen Electricity Network’s and Fingrid’s marketplace for congestion management
Flexibility markets have been seen in Finland as a necessary tool for congestion management. Prior knowledge for preparing the markets has been gained from technical pilots with simulated market design in INTERFACE and OneNet projects. TSO Fingrid and the DSO Helen Electricity Network are launching together the first pilot of the live marketplace for congestion management with a go-live target in Q1 of 2025.
General description
The project creates and pilots a common marketplace for congestion management for Fingrid and Helen Electricity Network (DSO). Although it is a pilot project, the market will be used operationally by the system operators to solve congestion. The pilot project tests the functioning of the flexibility market for congestion management, defined products and market rules, and the market’s liquidity. The plan is to continue the piloting phase until the end of 2027. The longer-term target is to develop a common national marketplace by the experiences gained from the pilot.
About Helen Electricity Network Ltd. (DSO)
Helen Electricity Network is part of the Helen group (included as a sub-group in the Helsinki city Group) and has 430,000 customer measurement points and 36,000 connection points. Further information about the network can be found in the Electricity distribution network development plan of Helen Electricity Network Ltd
Grid Issue
Fingrid’s main challenge is the transmission grid capacity from northern and central to southern Finland and from western to southern Finland. The case study of Fingrid in the project is to solve 400 kV grid congestion in normal, fault, and disturbance situations. Fingrid uses the congestion management market as part of other congestion management methods.
Helen Electricity Network is challenged by rapid changes in the electricity production (shutdown of CHP power plants) and consumption (e.g., electric boilers and electrification of transport), which might create a deficit in the 110 kV network. The problems could occur rather seldom and infrequently, e.g., in an N-1 situation. The maximum loading of the area is expected to double by 2030, up to ca. 1600 MW. This challenges the capacity of the local 110 kV network. In addition to maximum loads, energy transmitted to customers is estimated to increase from 4,436 GWh in 2023 to 8,294 GWh in 2033
Helen Electricity Network Ltd. Electricity distribution network development plan 2024.
Value chain for DSF
The value chain of flexibility of the marketplace with the main stakeholders is presented below.
Table 22: Physical value chain for DSF in Helen Electricity Network's and Fingrid's marketplace for congestion management
Flex Mechanisms
Fingrid has defined five congestion zones, i.e., procurement areas of flexibility, for the marketplace (North, Central, South, Southwest, Uusimaa)
The DSO’s main solution is temporary flexible connections, but the market-based congestion management opportunities are an additional possibility. Helen Electricity Network Ltd. has created four congestion zones for the marketplace for congestion management.
Brief description of the marketplace for congestion management
Fingrid and Helen Electricity Network have submitted the terms and conditions of the marketplace to the Energy Authority for approval in October 2024. When the terms and conditions have been approved, trading in the marketplace will begin, estimated in Q1 of 2025. The description here is based on the situation in December 2024, and the information is based on Fingrid's and Helen Electricity Network’s info event on 17.12.2024 and interviews of Fingrid’s and Helen Electricity Network’s representatives.
The marketplace responds mainly to predictable congestion situations. In Fingrid’s case studies, in a sudden 400 kV network fault situation, other congestion management methods, such as mFRR special regulation, are used first, and the marketplace comes in later if the congestion continues. Also in Helen Electricity Network’s case study, in addition to forecasted congestions, e.g., the exceptional switching status caused by a sudden fault and the ensued, prolonged repair work might cause flexibility needs. The project aims to harness especially such flexibility resources that are not capable of participating in the TSO reserve markets.
The chosen market platform is NODES, which facilitates the trading of flexibility between grid operators and flexibility providers. The market products are LongFlexTM for capacity and ShortFlexTM for energy markets. The capacity product LongFlexTM allows system operators (SO) to reserve flexible capacity in a capacity market auction to ensure that flexibility is available when it is needed. The accepted bid for LongFlex obligates to give an offer to the ShortFlex product. Both LongFlex and ShortFlex bids are given separately for the up- and down-regulation. The market operates on a pay-as-bid principle. The idea is to make it possible to develop product specifications based on experiences gained during the pilot project. The bids are selected per flexibility area in price order, starting with the most cost-effective from the network company's perspective. The system operator can set a price cap for the product.
Helen Electricity Network aims to use the capacity product LongFlexTM. Helen Electricity Network’s ShortFlex product’s trading happens by an auction at D-1 10:00 (EET/EST) before the day-ahead market closes, and thus the trade is not reported to eSett for the imbalance settlement. The DSO accepts the ShortFlex bids needed at 10:00 the day before the delivery. Since the trade is not reported to eSett for the imbalance settlement, the balance responsible party (BRP) and the electricity supplier need to take into account the trade done with the system operator in their trading activities, e.g., in the day-ahead market.
Fingrid’s ShortFlex product is continuously traded after the day-ahead market closes, from D-1 at 15:00 (CET+1/CEST+1) until 60 minutes before the delivery (T-60 min). Fingrid reports the energy trade for eSett. The reporting is planned to be implemented latest 13 days after the transaction. In both products, the invoicing of the transactions between the SO and the flexibility service provider (FSP) is managed by NODES. The bids not accepted for Helen Electricity Network’s ShortFlex product can be used automatically for Fingrid’s ShortFlex product.
Participation in the new TSO-DSO congestion management market and reserve markets of the TSO with the same flexible resource is possible. However, the same capacity of a resource cannot be bid to the congestion management market and to a reserve market (mFRR, aFRR, FCR-N, FCR-D, or FFR) for the same hour. A flexibility service provider decides which markets to participate in with its flexible resources. The need for flexibility in the new congestion management market will be significantly lower compared to the reserve markets, where flexibility is needed every day for balancing.
The gate closure times of the TSO-DSO congestion management market are set taking into account the gate closure times of other markets, especially mFRR/aFRR capacity and energy markets:
•The DSO’s congestion management energy market auction is held D-1 at 10:00 (EET/EEST) for all hours of the next day. The auction is held after the results of the mFRR and aFRR capacity markets that have been received earlier in the D-1 morning.
•The trading in the continuous congestion management energy market ends 60 minutes before delivery, and thus, the flexibility service providers know the result of the market before the gate closure time of the mFRR (45 min before delivery) and aFRR (25 min before delivery) energy markets.
The market time unit is planned to be 60 minutes, and the minimum bid size would be 0,1 MW for the ShortFlex and LongFlex products. However, some FSPs have expressed that there would be more flexible resources available for shorter bids. The FSP defines the preparation time required for its flexible resources and can set an expiry time for its offer in the ShortFlex product. The verification of the traded flexibility by the measurement data happens after the transaction.
The FSP could use the baseline provided by NODES based on the measurement data and calculation methods of NODES, or send its own prognosis of the baseline. The future EU-level network code for demand response is expected to include top-level principles for baseline and the process for defining it.
FSP must agree with the BRP and the electricity supplier of the flexible resource to participate in the marketplace. The FSPs are first pre-approved before their trading is accepted. The FSPs are queried about the technical information of the flexible resources, but prequalification tests of the flexible resources are not performed.
DSF potential from household heat pumps
DSF potential and possible barriers connected to household heat pumps and electric heating were mapped by interviewing flexibility service providers (FSPs).
A very indicative estimation of the number of heat pumps in Finland at the end of 2024 is as follows:
- Ground source heat pumps: 160,000 pcs.
- Air-to-water heat pumps: 90,000 pcs.
- Exhaust air heat pumps: 40,000 pcs.
- Air-to-air heat pumps: 1,040,000 pcs.
The estimate is based on the sales statistics of the Finnish Heat Pump Association.
Heat pumps in Finland are mostly installed in residential buildings. Based on Finnish energy performance certificate (EPC) database statistics, around 75% of ground source heat pumps are installed on single-family residential buildings, and below 5% on other than residential buildings.
If assumed that in winter months the average power demand of ground source, air to water and exhaust air heat pumps, together with their electric resistance heaters, would be around 2,5 kW, the total power demand would be around 700 MW. If assumed that the average power demand of air-to-air heat pumps would be around 1 kW in winter months, the total average power demand of air-to-air heat pumps would be around 1,000 MW.
During the coldest days (average temperature under -20 °C), the average power demand of heat pumps is a lot higher, and the differences of the heat pump types grow wider. If simplified and assuming that the average power demand of ground source, air to water, and exhaust air heat pumps together with their electric resistance heaters would be around 5 kW on the coldest winter days, their total power demand would be around 1,400 MW. If assumed that the average power demand of air-to-air heat pumps on the coldest winter days would be around 1,5 kW, the total power demand of air-to-air heat pumps would be around 1600 MW.
With these assumptions, a very rough theoretical estimation for the power demand of heat pumps in Finland could be somewhere around 1,700 MW on a normal winter day and around 3,000 MW on the coldest winter days. These numbers are indicative estimates including significant uncertainties, and, therefore, they should not be taken as trusted results. The demand response potential is weakened by, for example, the fact that older heat pumps are not easy to connect to FSPs’ services.
The demand response potential of heat pumps varies a lot according to outside temperature and the properties of the building, heat distribution system, additional heating sources, and the heat pump itself. On very cold winter days, air source heat pumps use more complementary heating (usually electric resistance heaters integrated to a heat pump or other electric heating) to support the heating with a compressor than ground source heat pumps, which increases air source heat pumps’ power demand on the coldest winter days. Fireplaces (masonry heaters) are also widely used as complementary heating on the coldest days, which affects DR potential. On the coldest winter days, the room temperature drops more quickly when the heating stops, especially if the insulation of a building envelope is weak, the ventilation system doesn’t include heat recovery, and there is low thermal mass on the heat distribution system and on the building itself.
The heat pump heating in Finland’s building stock is increasing, and electric heating is decreasing. The estimated development of electricity consumption from 2020 to 2030 expressed in the long-term renovation strategy, is presented in the table below.
Table 23: The estimated development of electricity consumption of electric and heat pump heating in Finland’s building stock from 2020 to 2030
Building type | Heating method | Consumption in 2020 | Consumption in 2030 |
Detached and semidetached houses | electric heating heat pump heating | 9 607 GWh 4 958 GWh | 7253 GWh 7115 GWh |
Terraced houses | electric heating heat pump heating | 1,385 GWh 591 GWh | 1 132 GWh 824 GWh |
Apartment houses | electric heating heat pump heating | 1136 GWh 81 GWh | 810 GWh 836 GWh |
Non-residential buildings | electric heating heat pump heating | 2106 GWh 183 GWh | 1810 GWh 953 GWh |
Source: Finland’s long-term renovation strategy
Project learnings: Barriers and enablers
Barriers and enablers were mapped by interviewing FSPs connected to the DSF of residential buildings. This review was not restricted only to the marketplace for congestion management but also included a wider DSF context.
Barriers: household heat pump solutions
Finnish FSP Kapacity.io was interviewed for the use of home-sized heat pumps as a DSF resource. Kapacity.io provides services to energy companies and property owners through their DSF cloud platform for heat pumps and electric heating. The company operates internationally and participates in many DSF markets across Europe. The DSF resources are connected to the Kapacity.io cloud. Kapacity.io’s service gives control commands directly to the device manufacturer's cloud platform without an intermediary to harness the extensive control capabilities provided by each manufacturer. Kapacity.io was acquired by North American FSP EnergyHub at the end of 2024.
The home heat pumps are connected to the Kapacity.io cloud via an internet connection (LAN or Wifi), so additional devices are not needed. Most homeowners can bring the service into use independently. A wide range of heat pumps in the market is supported.
The identified barriers and respective opportunities:
- The local DSF marketplace pilot: FSPs’ DSF resources are spread over a wide area. Difficulty in achieving a significant capacity for local DSF from a specific area in Helsinki. The condition that a bid can include only resources from one BRP could also limit the opportunities.
- The possible duration of DSF with heat pumps varies a lot according to the weather conditions (e.g., weak DSF potential in -30°C) and when the previous activations have taken place (when and how long). For heat pumps, DSF for two hours could be possible, but for six hours, it is not feasible. However, the potential duration depends strongly on weather conditions and building properties.
- Verification:
- Measurement of the heat pump's electricity consumption should be accurate enough to fulfil verification requirements. There is some variation in the accuracy of the electricity consumption measurement integrated into heat pumps. The interviewee estimated that there could be some variation in the accuracy between heat pump manufacturers. There is also variation in the measurement accuracy between different installations (different operating conditions) with the same heat pump model. The accuracy could not be verified, but one manufacturer's representative estimated that the measurement error could roughly be inside the tolerance of +-10%.
A weak accuracy can be seen as a barrier to fulfilling verification requirements. The accuracy can be solved with external electricity consumption measurement, but this raises costs for smaller household size heat pump installations. Harmonisation of the methods to estimate heat pumps’ electricity consumption and requirements for minimum accuracy could be seen as a possible opportunity to remove these barriers.
- The verification requirements for DR differ between different DSF markets. Harmonisation of requirements and conditions in general would help FSPs' operation in different DSF markets.
- Internet connectivity:
- Heat pumps’ internet connectivity depends on the model. The connection is often as an additional module in cheaper models, which increases the adoption threshold. Requirements for internet connectivity are a possible opportunity to remove the barrier.
- API interfaces: There is a lot of variation in API-interface arrangements among heat pump manufacturers, and also, bad API practices exist. Variation complicates and increases the work for connecting DSF services to the cloud services. Kapacity.io has published a “wish list” for the HVAC device cloud compatibility, which includes an example on how to build an industry-standard API for controlling heat pumps by third parties
. The following control and monitoring possibilities of the heat pump via the manufacturer’s cloud service must be enabled to ensure compatibility with the service: - Control the supply temperature setpoint
- Control the domestic hot water temperatures (if the heat pump provides domestic hot water)
- Monitor heat pump domestic hot water temperature (if heat pump provides domestic hot water)
- Monitor heat pump supply temperature
- Monitor alarms of the heat pump
Based on the above, harmonisation of heat pump manufacturers’ API-interfaces for controlling heat pumps via the manufacturer’s cloud by third parties and minimum requirements for control and monitoring capabilities could be seen as an opportunity.
- Information about device manufacturers’ software updates in time for the FSPs. A software update of a heat pump model may cause a need for an update to the DSF service. If the FSP doesn´t receive information about the software update of the heat pump model in time from the manufacturer, the heat pump’s compatibility with the DSF service could be disturbed. In the worst-case scenario, this might cause disturbances to the operation of the heat pump.
- The installation and connecting the device to the internet: Some installers could benefit from training on the DSF possibilities of heat pumps. The ease of connecting the heat pump to the Internet varies between manufacturers. Some residents need help connecting the device to the internet.
Barriers: home ESWHs, electric heating, and air-to-air heat pumps
Finnish FSP Optiwatti was interviewed for the use of home ESWHs, electric heating, and air-to-air heat pumps as a DSF resource. Optiwatti has “dozens of megawatts” of DSF resource in its service. The brief introduction of the service and systems in use:
- OptiWatti controls home heating equipment centrally: Electric radiators, underfloor heating, ESWHs, and air source heat pumps. In addition, car heating, electric car charging, and solar panel production information can be integrated into the same control system.
- The heating of each room can be controlled separately, and the room temperature can be programmed with an hour granularity.
- The service automatically optimises heating based on indoor and outdoor temperature, weather forecast, and electricity exchange price. Aggregated control of the resistive load-based resources for the FCR-D (up) market at the development stage (as part of Horizon Europe STREAM project).
- Equipment assembled to provide the service:
- Sensors: measure the temperature and humidity in each room
- Smart central unit: controls the heating and connects to the internet (Optiwatti cloud)
- Smart radio-controlled relays: control heating devices
- Accessories: Air source heat pump ir-controller, hot water boiler control relay
The identified barriers and respective opportunities:
- It is assumed to be more economically feasible to offer fast responding DSF service with resistive load to e.g., FRC markets than for the local DSF market. If FSP is accepted in yearly FCR markets, there is no possibility of taking part in the local DSF market in the same period with the same DSF resource. However, FSP could take part in the local DSF market together with the hourly FCR markets.
- FSP has some difficulties in obtaining customers’ information, e.g., on the electricity supply point, electricity supplier, and BRP. The information is needed for FCR and local DSF markets.
- Suggested solution when FSP utilises the electricity meter’s control relay: FSP would obtain the customer information from Datahub through the customer's authorisation. The Government's bill suggests that a load control interface for electricity meters load control relay, would be built to the Datahub, but the project has not yet started, and the information FSPs could obtain from the Datahub in this context has not been defined.
- ESWHs: FSP doesn’t have real-time information about the temperature or electric current/power consumption of the ESWH. The information would help the FSP to estimate the DR capacity of ESWHs and to verify the fulfilled flexibility. A readable signal from ESWH, whether electric current is being consumed, and the hot water temperature information from the ESWH would help here. Possible requirements for the minimum information obtained from the ESWH could be seen as an opportunity.
- Air-to-air heat pumps: It is difficult for FSP to obtain information straight from the air-to-air heat pump, because heat pumps lack an easy standard communication interface for this (e.g., Bluetooth, USB-C). A two-way communication should be easily enabled, and the beneficial data points from the heat pump should be available: e.g., target temperature, fan speed, and the power consumed. The information would help to know if the control commands for the heat pump are functioning, to estimate the demand response capacity, and to verify the fulfilled flexibility.
Communication protocols of air-to-air heat pumps are manufacturer-specific. The harmonisation of the communication protocols would help connect air-to-air heat pumps to the FSP’s service.
- Electric radiators: A variety of connectivities exist. Standard connectivity would help to connect radiators to the FSP’s service (e.g., Bluetooth, USB-C, or standard Wifi). The information on the power consumed would be beneficial on the same basis as in air-to-air heat pumps.
The development of the Code of Conduct for Energy Smart Appliances (ESA) can be seen as an opportunity here. In the mapping of use cases of the Code of Conduct (version 1.0), the monitoring of power consumption is defined as a mandatory feature for heat pumps, local space heaters, water heaters, and ventilation. The implementation of the capabilities of Energy Smart Appliances can be, e.g., realised physically in the ESA or represented as a digital twin in the manufacturer's cloud
Code of Conduct on energy management related interoperability of Energy Smart Appliances (V.1.0)
Finnish FSP Cozify was interviewed for additional information on communication and controllability of electric heating and heat pumps. The key takeaways:
- The problem with the lack of two-way communication properties of the devices was also verified by Cozify. E.g., the MODBUS communication module for a single heat pump could cost around 300 EUR. A basic example for the importance of information from the device controlled: if known when the radiator heats up and the heat pump cools the air, the service could, e.g., prevent the heating when the cooling is on. The information also helps to estimate the DR capability of the device at a specific moment.
- Heat pumps have not been designed with demand response in mind, and the requirements for controllability for DSF purposes are a somewhat new issue for the manufacturers.
- The interviewee has seen some challenges in controlling devices via the manufacturers’ cloud services: if, e.g., the user's password expires, the service might not work before the password has been changed.
Finland: DSF possibilities of smart electricity meters load control relays, case Elenia
General description
About Elenia
Elenia is the second largest DSO in Finland with 440,000 customers and 76,600 km of electricity network
Today, smart electricity meters’ load control relays are used to control electric loads mostly according to the DSOs’ night tariffs. The next generation of electricity meters, which are currently being installed by DSOs, are called AMR 2.0 meters (AMR = Automatic Meter Reading). The Government Decree mandates load control functionality as a mandatory feature also in the new 15-minute resolution AMR 2.0 meters for single-family and two-family houses. Electricity meters must be changed to AMR 2.0 meters by 4.7.2031.
The government bill proposes the commercialisation of the load control of DSOs’ AMR 2.0 meters for use by FSPs. The regulation would enter into force on 1 September 2026. Residents could authorise FSP to issue load control commands via the load control interface, which would be built to the Datahub. The role of DSOs is seen as an enabler and procurer of local demand side flexibility, not as a main FSP.
Elenia has carried out extensive development work together with Aidon (producer of smart electricity meters) to enable the feasible use of load control relays in AMR 2.0 meters for demand response. Elenia piloted the DSF possibilities of smart electricity meters in 2018-2020. Elenia today offers free of charge DSF service, where household customers are able to control the loads connected to the AMR 2.0 meter’s control relays automatically according to the cheapest day-ahead market prices. In the future, the AMR 2.0 load control relays could also be used for the evolving local DSF markets and for optimising the loads according to the DSOs’ power tariffs.
Estimated development in the grid area and grid issues
In the future, the population is estimated to decrease in Elenia’s network area, but electricity consumption is estimated to increase significantly, especially due to the electrification of industry and transport. Production is estimated to increase significantly due to the increase in solar and wind power. The number of new electricity storage connections is expected to continue to grow. Increasing consumption and production will require significant investments in the network. The key indicators of the electricity network over the next ten years are shown in Table 23.
Table 24: Key indicators of the electricity network over the next ten years
Current state, 31.12.2023 | Forecast, 31.12.2033 | |
Energy transmitted in the network area, MWh | ||
Energy transmitted to network service customers | 5 789 601 | 7 902 000 |
Energy received from network service customers | 2 928 679 | 12 170 000 |
Number of places of use | 440 025 | 450 000 |
Distributed generation | ||
1) Nominal power, kW | ||
1.1) Connected to the high-voltage network | 1 246 000 | 3 986 000 |
1.2) Connected to the medium-voltage network | 150 800 | 642 000 |
1.3) Connected to the low-voltage network | 16 026 | 401 000 |
2) Number of production sites | ||
2.1) Connected to the high-voltage network | 26 | 75 |
2.2) Connected to the medium-voltage network | 89 | 526 |
2.3) Connected to the low-voltage network | 16 026 | 43 053 |
Number of connections used in public charging of electric vehicles | 266 | 602 |
Value chain for DSF
A simple physical value chain of utilising Elenia’s AMR 2.0 meters’ load control relays as they are utilised today is presented below.
Table 25: Physical value chain for DSF in Elenia's control relay service
The following example describes a possible general, not DSO-specific value chain in the future, in a case where load control relays of AMR 2.0 meters would be commercialised, the control interface would have been built to Datahub, and the loads connected to the AMR 2.0 meters’ load control relays would be aggregated by FSP to the DSOs’ marketplace for congestion management. Today, Helen Electricity Network Ltd. is the only DSO in Finland, which is piloting the marketplace for congestion management, but the longer-term target is to develop a common national marketplace by the experiences gained from the pilot (the marketplace pilot is described in Chapter 3.10).
Table 26: Physical value chain for DSF in Elenia’s case with commercialised control relays
The flexibility mechanisms and capabilities
Elenia today has the service for utilising the AMR 2.0 control relays in day-ahead markets. The household customer with AMR 2.0 meters’ load control relays can choose to use the spot-price control in Elenia’s cloud service. The end user is responsible for verifying that there is a controllable load connected to the meters’ control relay. The cloud service chooses the next day’s cheapest spot prices and sends the control calendar to the ARM 2.0 meter, which activates the control relay(s) and the appliance(s) connected to the relay(s) according to the calendar. The most typical loads connected to the relays are ESWHs and underfloor electric storage heating.
AMR 2.0 meters’ load control relays will be commercialised in the near future. The government bill proposes the commercialisation of the load control of DSOs’ AMR 2.0 meters for use by FSPs. The regulation would enter into force in September 2026. The load control interface would be built for the Datahub. FSPs would send the control commands (e.g., a control calendar) to the common Datahub’s control interface. Each DSO would integrate its electricity metering system into the interface. The control interface would send the control commands onward to the DSO’s system, which activates the AMR 2.0 meter’s control relay. The end user (household) would authorise the FSP in the data hub for controlling the AMR 2.0 meter’s relays.
It is expected that the load control interface will be used first, mostly in the day-ahead market. In the day-ahead market, the required reaction time is easy to fulfil, economic potential is high, most household customers understand the principle, and there are no challenges concerning verification, baseline definitions, or balance management.
In terms of technical reaction time, the local DSF markets have good potential, but economic potential is expected to be somewhat weaker than in the day-ahead market. Using the load control relays in the intraday markets requires faster reaction time, which is still reachable with AMR 2.0 control relays, and the economic potential is good. In the legislation, the load control request’s turnaround time is set to a maximum of six hours. However, in most cases, the turnaround time of control commands is practically no more than a few minutes, and at its fastest, only tens of seconds. The control commands will be implemented in the DSOs' systems on a “best effort” basis to reach a faster reaction time than legislation obligates.
The use in reserve markets challenges the reaction time of the AMR 2.0 control system and, therefore, doesn’t seem to be feasible in the short term.
The systems in use
Elenia’s AMR 2.0 meter’s load control system includes the following systems:
- Elenia uses Aidon’s AMR 2.0 meters with Aidon’s head-end system (HES)
- Elenia’s metering data management system (MDM)
- Elenia’s smart grid platform with a user interface, including the control of load control relays
- The signals are sent to the AMR 2.0 meter using a mobile network connection (LTE-M or NB-IoT)
Today, only DSOs can send control commands for the AMR meters control relays, but according to the government's bill, control relays will be commercialised in September 2026. The FSPs would be authorised by residents in the Datahub to issue the load control commands. FSPs would send the control commands via the load control interface, which would be built into the Datahub. The control interface conveys the commands to the DSO’s system.
DSF potential from loads controlled by AMR 2.0 meters
Technical potential of the AMR 2.0 meters’ load control relays – ESWHs and electric underfloor storage heating
In Finland, the new AMR 2.0 meters are obligated to have the control relay when they are installed in detached and semi-detached houses. The rough estimations of DSF potentials connected to the relays are presented below. The loads connected to the control relays consist mainly of ESWHs and underfloor electric storage heating.
The DR potential of ESWHs and electric underfloor storage heating in Elenia’s grid
Elenia’s 70,000 customers have the time-based tariff (day-night tariff), and the maximum DR potential of the electric load connected to load control relays in the grid area is estimated to be around 125 MW. The controllable capacity connected to the load control relays of the meter is estimated to be 75 MW in summer and 120 MW in winter. The controllable load per customer is estimated to be 3 kW on average
Vanguard Consulting, 2024.
National DR potential of ESWHs and electric underfloor storage heating connected to electricity meters’ control relays
The estimated number of detached and semi-detached houses with electric heating in Finland is around 506 000 and terraced houses 26,000
Statistics Finland.
Järventausta et. al. (2015) estimated that in total, there would be around 2,900 MW installed total capacity of electric heating, which can be controlled by electricity meters’ load control relays. The electric load already connected to the relays was estimated to be 1800 MW. After the study, some of this load may have been disconnected from the meters.
The national flexibility potential of the devices installed in the load control relays today is estimated to vary as a monthly average between 950 MW in March and 540 MW in July
Vanguard Consulting, 2024.
Typically, electric underfloor heating is combined with an air-to-air heat pump and often also with a heat-retaining fireplace. This brings uncertainty to the estimations.
The development of local DSF possibilities
Elenia today offers a free-of-charge DSF service, where homeowners in Elenia’s grid with new AMR 2.0 meters can control the loads connected to the relays automatically according to, e.g., time of the day or cheapest day-ahead market prices. In the future, AMR 2.0 meters’ load control relays could be used for the local DSF markets as they evolve, and for optimising the loads according to the DSOs’ power tariffs. Elenia estimates that the DSF potential of AMR 2.0 meters for congestion management is growing significantly towards the end of the decade
Elenia’s Network Development Plan.
Coordination with DSO needed
There are many details that need to be agreed upon before the FSPs will be able to launch the services using the AMR 2.0 load control relays. If the meter breaks down, who will ensure that the customer has hot water? If the control command to the meter does not get through, is it the responsibility of the DSO or FSP to send the message again?
Elenia’s interview on 26.11.2024
Project learnings: Barriers and enablers
Elenia’s pilot on electricity meters’ DSF possibilitiesThe intelligent power grid as an enabler of the virtual power plant platform, Final report. 2020 Elenia’s and Vattenfall’s interviews on 26.11.2024 and 31.12.2024
The intelligent power grid as an enabler of the virtual power plant platform, Final report. 2020
Elenia’s and Vattenfall’s interviews on 26.11.2024 and 31.12.2024
Elenia piloted the DSF possibilities of new generation electricity meters’ load control relays in 2018-2020 together with Vattenfall, Empower, Telia, and Visma Consulting. The loads involved were ESWHs and electric heating in 76 homes in Elenia’s grid. The pilot homes were fitted with new electricity meters, enabling load control with a faster reaction time.
Electricity supplier Vattenfall controlled the loads according to the day-ahead market, and FSP (Empower) according to the mFRR, with part of the pilot group. When no more system changes and repairs were made, the throughput rate of control commands was high, 98%. The reaction time with 88% of commands was less than one minute, and of those commands 80% reached under 30 seconds reaction time. At its longest, receiving the message took about 42 minutes from sending the first control command due to poor signal coverage. The maximum controlled average power per customer varied from 2,6 kW in June to 4 kW in winter months (more info in the table below).
The pilot also tested how long the total time is from sending the control signal to receiving the power change message back from the electricity meter. When the first power change message was received successfully, in 90% of the controls power change message was received within 3 minutes after the control command was sent.
Customers responded monthly to a survey about the impact of the test controls on living comfort. Primarily, the test controls did not affect living comfort, and to the extent that they did, most of the perceived effects were positive.
Table 27: The controlled power in Elenia’s pilot
Maximum monthly controlled power and average power per customer The intelligent power grid as an enabler of the virtual power plant platform, Final report | ||
Maximum total controlled power (kW) | Maximum controlled average power/customer (kW) | |
June | 21 | 2,6 |
July | 168 | 3,1 |
August | 203 | 3,6 |
September | 203 | 3,8 |
October | 138 | 4 |
November | 159 | 3,2 |
December | 204 | 3,9 |
January | 176 | 4 |
February | 179 | 4 |
March | 175 | 3,9 |
Maximum | 204 | 4 |
Barriers and opportunities of DSF connected to the use of load control relays
To identify the barriers and opportunities of the DSF connected to the use of load control relays, representatives of Elenia, Optiwatti, and Vattenfall were interviewed. The recognised barriers and opportunities are presented below.
The barriers
- The reaction time of the service utilising AMR 2.0 meters’ control relays today limits the use for TSO’s reserve markets. In addition, the measurement resolution of the meter should be more frequent. Elenia’s AMR 2.0 meters send metering data in 5-minute resolution for the measurement information system, which combines it to a 15-minute time series. It would be possible to program the meter to send the data, e.g., in 1-minute resolution, but the amount of data increases, and the speed of the transfer needs to be fast enough. Limitations to the reaction time could come from the capability of the mobile network (e.g., NB-IoT). If the data is sent to thousands of customers at the same time, the deviations could increase to some extent.
Elenia’s AMR 2.0 meter (Aidon’s meter) is capable of controlling loads also according to frequency. This means that the load would regulate up and down according to the frequency when the frequency control is “turned on”. The frequency control is “turned on and off” by the control calendar sent to the meters' Head End System. The control calendar sent to the meter could, e.g., turn the frequency control on at 8 am, turn it off at 10 am, turn it on again at 2 pm, and off at 3 pm. - In the future, it could be possible to use the loads connected to the control relays to the FCR market by sending the meter a frequency control calendar in advance. The calendar would tell the meter to control the loads of the relay, for example, between 12 noon and 8 pm based on the measured frequency. The development referred to above should be done to fulfil verification requirements of the FCR-markets (shorter metering resolution and sending the metering data for Fingrid almost in real time).
- Challenges with poor signal coverage in some areas. The load control function requires a better connection than traditional day-night control. Although the throughput rate of control commands to the ARM 2.0 meter has been high (98-99 %), some connection failures exist due to poor mobile network connection (NB-IoT or LTE-M) to the meter. It is possible to improve the connection with an additional antenna. Poor connection can prevent the use of LTE-M, which would be faster than NB-IoT, which has better signal reception.
- The operation of an electronic thermostat, with e.g. weekly program, may be disrupted if the heating is controlled too frequently by cutting the power directly with the electricity meter’s control relay (e.g., the use for FCR-N markets). When the loads are controlled based on the hourly spot price, the control frequency has been estimated to remain at a relatively safe level for the electronic thermostats. The problem with frequent controls could be bypassed by implementing a control from the electricity meter to the thermostat, if the thermostat is compatible with the signal from the meter (usually a potential-free contact). However, physical wiring from the meter to each thermostat is expensive to implement.
- The space heating, which consists of many heating sources, could be challenging in DSF markets, because it is more complicated to estimate how electricity consumption behaves.
- Lack of information: Elenia conducted a survey for 1000 customers who had an AMR 2.0 meter and load control connected to the meter (night-tariff) and had not adopted the spot price control. Many respondents were not aware of the service despite the information sent to customers. 60% of customers who were not aware of the service responded that they will adopt it.
- The weak knowledge of the household customers: The household customer needs to understand the principle of how the load control works and how he/she benefit from offering the resources for the local DSF. The service based on the local DSF could be more difficult for the customer to understand than the service based on the spot prices.
- Only the customer knows if there is a load connected to the relay suitable for the DSF service. Sending the electrician to ensure the electric connections and the loads/appliances connected to the control relays would be expensive for the FSP. This causes some uncertainty for the FSPs in estimating the DSF potential of their resources. The FSPs have a responsibility to inform the household customers what kind of devices can be controlled by their service.
- Some FSPs don’t recognise AMR 2.0 control relays as an interesting way to control the electric loads, which have capabilities to react fast enough for the TSO’s reserve markets, if controlled straight by the FSP’s own control system.
Opportunities
- A good DR potential with the loads already connected to the control relays. The national DR potential is estimated to be 540 MW in July and 950 MW in March. No need for additional devices/installations. The expectation is that electricity suppliers will widely offer DSF services that utilise AMR 2.0 meters’ load control relays when the use is commercialised.
- There is a lot of variation in electrical switchboards and couplings. Still, electricians have been able to install the loads correctly in the load control relay.
- Elenia’s AMR 2.0 meters’ control relays can offer the reaction time needed for local DSF and, e.g., for intraday markets.
- In Elenia’s pilot, the test controls primarily did not affect living comfort (91% of respondents), and to the extent that they did, most of the perceived effects were positive (57%). In general, customer satisfaction with the load control was rated 4,5/5.