Power-to-heat on the reserve capacity market—Policy implications considering economic constraints and competing heat production

It is expected that occasions when electricity production is mismatched with electricity demand will increase in power systems containing large shares of variable renewable electricity (VRE) generation, i.e., wind and photovoltaic. These mismatches are caused by poor demand response flexibility and inherent weather dependency for VRE-generation technologies that can reduce power quality. This causes problems for the transmission/distribution systems and their operators (TSO/DSO). When electricity that exceeds the demand is generated, curtailment is a quick and an inexpensive fix. However, with large amounts of VRE, curtailing all generated surplus electricity would result in an inefficient management of the resources. Therefore, it is of interest to find solutions that enhance the total energy and economic efficiency by maximizing the use of resources.

In this study, the use of Power-to-Heat (P2H), in the form of large-scale heat pumps, is investigated as a strategy for managing the surplus from VRE production. The potential of P2H in district heating (DH) production has been examined in previous research and has proven to increase the useful share of VRE by up to 2–3 times the initial self-consumption.¹ The idea is that P2H consumes surplus electricity from VRE and produces heat. Thus, P2H can contribute to decarbonization of the DH sector by reducing the use of fuels, such as fossils and/or resource-scarce biomass, while simultaneously contributing with power balancing services. Large-scale P2H production units are already well established within many Swedish DH utilities.² Typically, P2H is used as intermediate load supply units to meet changes in the heat load. Preferably, they follow electricity prices, i.e., producing heat when the electricity price is low and being replaced by other production units when the price is higher. This enables DH production systems to be economically flexible.

In this study, we aim to further enhance the understanding of the economic applicability of P2H and also to find how heat production from P2H competes with the other DH heat and power production units and what policy implications this yields. This was done by taking into consideration the diversity of production units in three different DH systems. As is shown in the literature review, placing bids on the reserve market increases management complexity since heat produced from power reserve market sales occasionally is mismatched with local DH demands. Furthermore, for DH operators already active on the day-ahead and intraday electricity market, placing bids on the power reserve market means active participation on four different markets. Therefore, this study suggests a novel control regime for P2H production. The units are operated and owned by the DH utility but are controlled according to the reserve capacity market. P2H is integrated into the DH systems by buffering large-scale thermal energy storage (TES) units where the P2H-produced heat is stored and thereafter managed by the DH utility operator.

To fulfill the overall aim, the potential for P2H was investigated by quantifying the value of the produced heat in three different DH systems. This was done by answering the following research questions:

• What would the bidding price need to be to make a request from TSO/DSO economically justifiable in three different DH systems?

• What storage sizes are required for optimized P2H power balancing production?

• What kind of heat production does stored heat compete with?

The paper is structured as follows. The background is presented as a literature review in Sec. II. The method and materials used in the study are presented in Sec. III. The results are then presented in Sec. IV, and the results and the method are discussed in Sec. V. Finally, the conclusions and implications for policy-makers from the study are presented in Sec. VI.

In this section, the study is placed in a relevant context by providing a literature review of the research field. This is presented in four sub-sections, where Sec. II A contains studies of the technical potential for P2H and Sec. II B describes the markets and market barriers that challenge the implementation of P2H. Subsection II C focuses on ancillary service markets and the complexity this might cause. Finally, Subsection II D describes the competition between P2H units and other DH production units.

Lund et al.¹ propose that coupling thermal and electrical energy systems improves utilization of their inherent flexibility capacities in terms of demand-side-management, storage, or alternative use of surplus electricity for the integration of VRE. There are several studies that further support this concept. Meha et al.,³ for instance, show that P2H can increase the facilitation of VRE in energy systems while reducing primary energy use and decarbonizing the DH. This cross-sectoral interaction is defined in Ref. 4 as a “Smart Energy System” where synergies yield lower system costs. An important reason for providing negative power balancing via P2H using TES as flexible loads instead of using batteries is that TESs are up to a factor of a thousand cheaper.⁵ Even dedicating photovoltaic electricity for heat production in P2H units improves system revenue,⁶ and unpredictable peaks in VRE are reduced and stored as heat at low cost.⁷ 

Böttger et al.⁸ and Schweiger et al.⁹ studied the technical potential for P2H in Germany and Sweden, respectively. Both show that the potential for P2H increases with the amount of generated VRE, and that it is mainly limited by transmission constraints and access to TES. However, Ref. 9 also points out competing heat generation as a limiting factor. Bloess et al.¹⁰ give a comprehensive overview of the P2H research and acknowledge that P2H cost-effectively can decarbonize the heating sector, substitute fuel, and promote VRE integration. A case study presented in Ref. 11 demonstrated that a wind-power park in combination with P2H and TES had enough production flexibility to fully supply the annual DH demand when the production is controlled by an aggregator. In Ref. 12, P2H is shown to reduce CO₂-emissions more cost-efficiently than retrofitting of buildings. Furthermore, the same study also shows that electric boilers increase wind integration more than heat pumps. In Ref. 13, on the other hand, heat pumps are concluded to contribute more to decarbonization and VRE integration. Thus, the indications of P2H potential in existing research results are somewhat ambiguous. All the above-mentioned studies focus on the technical potential for P2H and show its applicability. However, the studies lack a comparison of different compositions of DH production units as well as what effects this might have on the production economy.

From an international perspective, the Swedish situation, with existing large-scale heat pump capacity, is unique. The installed thermal capacity of heat pumps in DH since the 1980s is 1527 MWth, of which 80% of this capacity is still in use. The average coefficient of performance (COP) for these units is 3.3, and the most common heat source is sewage water.² The thermal power of electric boilers accounts for another 1151 MWth.¹⁴ This means that a significant power balancing resource is available if it is used in a sector-coupled way and designated for ancillary balancing services.

In Sweden, there are about 500 DH systems. Many of these are small, and heat is produced in biomass-fueled heat-only boilers (HoB). 35 DH systems have waste-to-energy recovery,¹⁵ and 83 systems contain combined heat and power (CHP) production from biomass.¹⁶ Some systems use industrial waste heat (IWH), and 72 systems contain P2H units (heat pumps and/or electric boilers).^2,14

Electricity markets are similar around the world. In this study, the Nordic electricity market, Nordpool, is considered. The Nordpool short-term market consists of three temporarily differentiated market areas: the day-ahead market (Nordpool spot), the intraday market (Elbas), and the reserve capacity market. Furthermore, the reserve capacity market also consists of three market areas. These are based on requirements for response to distribution network frequency deviations but also duration and repeatability of the capacity offered. These markets are Fast Frequency Reserve (response within seconds, lasting up to 30 s, restored within 15 min), Frequency Containment Reserve (response within minutes, lasting up to 1 h), and finally, Frequency Restoration Reserve (response within some minutes, lasting at least 1 h).¹⁷ 

On the day-ahead market, the actors place bids on the production and consumption for each hour of the next day. As the market closes at 12:00, the system and area market prices are calculated based on all bids. After the auction closes, however, operation starts at the earliest after 12 h, and the last hour of operation as late as after 36 h. Hence, during operational day, the intraday market is available until 45 min before the hour of operation to mitigate deviations in the actors' placed bids.

During the actual hour of operation, the TSO acts as balance-responsible and activates bids placed on the reserve capacity market. The type of deviation determines what bids will be dispatched since they vary in response time and duration.

Placing bids on a market with large shares of VRE is challenging due to the weather-dependency of the resources, as exemplified in Ref. 18. In demand-response projects, such as the Swiss Tiko,¹⁹ bid-placing has been made easier by the aggregation of many small users and the placement of their bids for nighttime. Bloess et al.¹⁰ conclude that several studies find the impact on electricity prices to be small for P2H. This is partly in line with the findings of Salpakari et al.,²⁰ where they show that implementing P2H primarily improves the amount of useful VRE sold to the day-ahead market significantly. However, Åberg et al.²¹ show that a strong correlation between the electricity price in the day-ahead market and the P2H-production exists, but the overall correlation between VRE production and electricity prices is weak. This implies that day-ahead market price is a blunt instrument for controlling P2H to consume surplus production from VRE.

A review of the P2H potential in Europe²² concludes that profitability is significantly improved when P2H is combined with CHP and TES. However, they also conclude that current fees and tariffs prohibit a cost-efficient use of excess electricity in P2H assets via the day-ahead market. In the literature, several barriers to increased use of P2H have been highlighted: electricity taxes,²³ biofuel tax exemptions making biomass-based heat-only boilers more competitive,²⁴ and grid network tariffs.^23–26 In Ref. 27, a dynamic tariff is shown to increase the operating hours for P2H compared to a constant tariff and also indirectly increases decarbonization of the heat production. The dynamic tariff controls the consumption of electricity toward low demand periods and thus remedies capacity shortage-related issues. However, it does not specifically target consumption to VRE surplus. In Ref. 23, tariffs designed to increase the use of P2H are concluded to cause reduced production in CHP units due to competing heat production, i.e., less flexible electricity generation. This means that the design of taxes, fees, and tariffs significantly impacts the potential for P2H.

Actors in the electricity market are assumed to be economically rational, i.e., they strive to maximize profit (for producers) or utility (for consumers). This rationality implies that actors are risk-averse. To minimize risks for long-term investments, forward market operators offer different contracts on a hedge market for all actors in the electricity market. The purpose of a hedge-contract is to guarantee the buyer a less volatile electricity price, e.g., negative prices or unexpected high prices. The benefit of such contracts is an in-advance known price for the electricity during the negotiated time of the contract. An effect of the hedge market is that the incentives for DH operators to realize the concept of P2H as a market-based, price-sensitive solution for power balancing may be significantly reduced.

An alternative would be to encourage utilities to place their P2H-capacity-bids in the balancing reserve capacity market, as suggested by Lund et al.²⁸ Participating in the reserve capacity market, however, increases operational complexity for the DH utility by needing to act on multiple markets.

Participating in ancillary service markets may improve revenues, but it also adds additional degrees of complexity in planning through necessary interaction with additional actors and markets.²⁹ On a local heat market, there may be different actors forecasting the demand, the production, and the distribution of heat (e.g., Varmelast in Copenhagen³⁰) These must, in turn, be coordinated with the potential bid-placing of the DH operator on an electricity market. In Ref. 29, it is concluded that participation in the reserve market is technically viable but not feasible, due to the multi-stage/multi-player process, i.e., operational complexity. This is partly addressed in Ref. 22, where P2H technologies on the control reserve market via contractual agreements between the system operator and the P2H owner are currently mainly applicable for industrial-scale installations.

In Germany, CHP units can be contracted to operate above a certain power output level, in order to ensure the possibility of providing power balancing by reducing the output level. Reference 8 shows that electric boilers can provide the same power balancing service cost-effectively and also reduce CO₂-emissions. Heat pumps could, in the same way, provide this service as well. Electric boilers have instantaneous response (<30 s), unlike heat pumps, which have some lead time before becoming fully operational. Thus, different P2H technologies are suitable for power balancing on different segments of the reserve markets.³¹ 

A hybrid system used for frequency control, with batteries in combination with P2H units (electric boilers), is investigated in Ref. 32. The conclusion is that such a solution is significantly more cost-effective than a stand-alone battery system. This is because the P2H units allow the battery capacity to be halved and the cost-savings of this compensate for extra investments in P2H units and grid fees. Furthermore, the hybrid system in Ref. 32 is integrated with an existing CHP plant, which supports the finding in Ref. 22 that access to multiple technologies is an advantage. However, in Ref. 32, the lack of TES in the system is a significant constraint, especially during negative residual peak loads where the heat demand limits the P2H production. This emphasizes the potential for improving system performance and production flexibility by adding a TES, which is in line with what is stated in Ref. 11. The disconnection of P2H-production from the heat demand that a TES provides allows heat production to be less limited by the heat demand and the planned production strategy. Therefore, this study suggests a novel approach as to how P2H capacities can be placed as bids on the reserve capacity market by utilizing the flexibility that TES provides.

With P2H-capacities on the reserve capacity market, the produced heat is a bi-product from power-balancing electricity consumption. This means, first, that production does not primarily supply the heat demand, and thus heat often needs to be stored, and, second, that the produced and stored heat from P2H units competes with heat produced from other production units.²³ Thus, if CHP production is limited by this competition, power supply capacity for balancing purposes is potentially lost, yielding reduced flexibility in the energy system. Flexibility to provide both positive (supply) and negative (consumption) power balancing is essential in energy systems with large shares of VRE. In Ref. 33, it is concluded that power system balance is best achieved if installed P2H and TES capacities linearly correlate with wind power production levels. The authors did not, however, consider the effect of inflexible power generators, such as nuclear power and/or heat-controlled thermal power plants without sufficient output ramping, which cause negative electricity prices on the day-ahead market, as explained in Ref. 34. If P2H units are operated by a DSO or a TSO, as suggested in this study, competition with other heat production units at the utility can be expected since the total production is limited by the heat demand.

The studies in the literature review focusing on ancillary services lack consideration of the diversity among DH systems (see Table I). DH utilities are either considered generic CHP units that provide bulk thermal production or fossil-fueled units targeted for replacement. This makes P2H a potential, and to some extent improved, heat production competitor. On the other hand, DH systems are local entities with quite different compositions of production units, and P2H would probably be less competitive in DH systems with waste incineration where the operator has a negative fuel cost, i.e., gets paid for waste management.

There appears to be a research gap in how different compositions of DH production units affect the potential for P2H and how production unit competition could impair power-producing flexibility. In this study, this research gap is covered by simulating three different DH systems with P2H capacities acting on the reserve market and controlled by a DSO/TSO. A future scenario with a large share of VRE is investigated. The P2H units are applied as an “add-on” production disconnected from the regularly planned heat production, but integrated with the DH system via a buffering TES. This would be a novel method for DH operators to offer P2H capacities to the reserve market.

Furthermore, the impact of competition from P2H on the DH revenues and thereby the likelihood for this to be implemented as a business model, is analyzed. Finally, the required sizes of TES units to support P2H balancing capacity are analyzed as well as their possible implications on production strategies.

Sections III4 presents the study design and the materials used for the simulations. Initially, a general description of the modeled and investigated system is presented (Sec. III A). Thereafter, the algorithms used in the simulations (Sec. III B) and the definition of the Willingness-to-pay (Sec. III C) are presented. This is followed by descriptions of the different datasets used for the simulations (Secs. III D–III G).

The DH system studied is based on a generic city with an annual heat demand of 1400 GWh situated in Scandinavia. The system is assumed to be connected to a national power system including large shares of VRE. Figure 1 shows a schematic description of how the thermal and electric energy systems are connected in the model. The algorithm considers the heat demand separate from the electricity demand. The two demands, however, interact in three different ways via the electricity market. First, there is the electricity price that determines expected revenue for sold electricity from CHP and, thus, impacts the merit order for the heat production units. Second, CHP produced electricity (depending on case) is sold on the electricity market. Finally, heat pumps consume electricity on dispatch from TSO via the reserve market. (For clarification: Henceforth, in the paper, the term P2H refers to additional large-scale heat pump capacities that is exclusively used in the reserve capacity market, operated on dispatch from TSO/DSO. Whenever heat pumps are explicitly mentioned this then refer to existing capacities at the DH utility that are part of the utility's assets and are not used in the reserve market.)

As shown in Fig. 1, all heat produced from P2H is stored and does not affect the production merit order. However, all heat that has been stored will be used to supply the heat demand at some point. If the heat demand at hour i is greater than the level of production in technology U_k, heat from storage is used. If the stored heat is insufficient to meet the heat demand, technology U_k increases its production output, if possible, to supply the demand. If the heat demand is still not met, another technology, U_k+1, is activated.

Electricity will only be produced in CHP units when the heat and electricity demands coincide. Without electricity demand, the turbine will be bypassed to increase the heat output.

The following setups are simulated: 1-HoB, 2-WtE, and 3-BioCHP. The compositions of these are shown in Fig. 2 and Table III, and further described in Sec. III E. Each setup is simulated for four different levels of P2H capacity that are placed as bids on the reserve market (see Table II). The results are analyzed by comparing changes in revenues for the three different DH-configurations. These analyses are used to discuss the probable bidding price to ensure economic efficiency for different DH systems, and the potential for P2H as balancing capacity.

$Here, IWH$ is industrial waste heat recovery for hour $i$⁠. $P WtE t h$ and $P CHP , bio t h$ are heat produced in CHP units fueled with waste or biomass, respectively. $P HoB , bio t h$ and $P HoB , fossil t h$ are heat produced in heat-only boilers fueled with biomass, and correspondingly with natural gas. $P D C t h$ is heat that is discharged from the TES and used to supply the heat demand. $γ$ is a binary function controlling the availability of heat in TES (see Sec. III F). The merit order for the production units is based on the respective cost for production which is described in detail in Sec. III G.

For the simulations, a set of conditions are defined that control the energy flows in the model. For instance, in each time step (1 h), the DH demand must be supplied. The merit order of the heat-producing units is based on production cost, where low cost is prioritized. In addition, P2H production is simultaneously controlled via the reserve capacity market. This means that the P2H production units are activated via directives from the responsible balance operator of the electric network, regardless of the DH demand. Thus, the P2H production is controlled by the power balancing demand, $P bal$⁠, and all heat produced in P2H is stored in TES. In the subsequent time step, the stored heat in TES is available for heat supply in the DHS and also affects the merit order for the heat-producing units. A decision tree for how the power balancing demand, $P bal$⁠, affects P2H production and electricity production in the CHP units is shown in Fig. 3.

$C tot BaU$ is the net production cost for the BaU-level, $C tot l$ is the net production cost for each level of P2H capacity above the BaU-level, and $P P 2 H e l$ is the electricity purchased from the reserve market. $c bid avg$ is the willingness-to-pay, or more precisely the highest average bid that the utility can place for purchasing electricity from the reserve market before BaU is cheaper.

An electricity profile was derived to represent the electricity balancing demand in a future scenario with a high share of VRE. It is assumed that no transmission limitations exist within Sweden as well as no export or import possibilities. The latter is motivated by the fact that a high share of VRE in neighboring countries would impair the cost-saving potential of cross-border transmission.¹ With the already large shares of VRE generation coming from Germany's sea-based wind power parks in the Nordic Sea, and Denmark's wind power capacity, the probability for future exporting conflicts is considered significant. It is thus of interest to maximize the national use of VRE before exporting any surplus. There is also reason to clearly identify the potential for P2H to contribute to power balancing consumption within Sweden under extreme conditions, such as without the possibilities of exporting, curtailment, and/or storage.

The electricity balancing demand profile is based on actual wind and solar power production data from Sweden in 2019.¹⁷ The data have been scaled so that the annual consumption in Sweden (≈136 TWh_el) is supplied to 60% (≈82 TWh_el) by wind power and 10% by photovoltaic power (≈14 TWh_el). The electricity consumption data are also from 2019³⁵ and is assumed to be representative for future consumption.

Figure 4 shows a conceptual description of the electricity balancing demand profile, P_bal. P_bal is equal to zero when the electricity demand and VRE production are matched. If P_bal > 0, there is an unsupplied electricity demand, a positive residual load (PRL), that requires additional electricity production from hydro, or thermal power or import from other countries. If P_bal < 0, a negative residual load (NRL), there is a surplus of VRE production that can be exported or stored. It is assumed that no curtailments are allowed. The focus in this study is on the NRL. The PRL is more relevant for DH systems with CHP production, since co-generated electricity from CHP units can supply PRL and thus reduce the total cost.

Figure 5 shows that there is a seasonal difference in occurrence of PRL and NRL peaks. The electricity demand correlates fairly well with the heat demand, yielding PRL peaks in the cold seasons. NRL occurs mainly during the low heat-demand season when the photovoltaic generation is at its peak.

A synthetic DH load is generated for a representative city located near the Swedish city of Växjö (56.9N, 14.7E). The heat load has a peak of 405 MW_th (see Fig. 6). The heat load is in the simulations supplied through heat produced in the three different DH system setups referred to as 1-HoB, 2-WtE, 3-BioCHP, also presented in Fig. 2, Tables II, and III. The first setup, 1-HoB, represents a system with biomass fuel-based heat-only boilers and no co-generation of electricity. This is common in smaller Swedish DH systems. The production cost of these systems is not directly influenced by electricity price levels. The second setup, 2-WtE, represents larger Swedish DH systems (e.g., Gothenburg, Västerås, or Uppsala in Sweden) and has a mix of heat supply units. IWH utilization and/or waste management (often referred to as Waste-to-Energy, WtE) are normally based on contracts that limit the flexibility to replace these with other production units in the short term for cost-purposes. The last setup, 3-BioCHP, represents a mid-sized Swedish DH system using solely biomass as fuel (wood chip) in combination with a heat pump and heat-only boilers for peak load supply. The production cost for this setup is more sensitive to electricity prices. It is assumed that the peak load units use natural gas. This is, at present, not the most common fuel used in Swedish DH production for peak load supply, but a conversion toward natural gas from coal and oil is generally considered as an intermediate step toward cleaner heating production in Europe.

This section presents the characteristics and mathematical modeling of the different heat production units and the TESs that are included in the DH system setups for the simulations.

For the BaU scenarios and setups 2-WtE and 3-BioCHP, heat accumulating tanks are included to handle excess heat from CHP turbine bypass.

This section includes a presentation of data and descriptions of the components (electricity prices, fuel costs, and taxes/subsidies) that are used to calculate the production costs and revenues from the heat produced in the DH setups.

VRE production is assumed to have zero marginal cost. Thus, with a large share of VRE in the energy system, the electricity price is expected to be influenced. In order to contain the volume of VRE production assumed in this study, the intraday market system-price for the Nordic region is modeled by adjusting the hourly bidding curves retrieved from Nordpool⁴⁰ for 2019 (see Fig. 7). This rendered a representative system price, $c S P e l$⁠, for a future scenario influenced by VRE. The price is converted to USD/MW_el (exchange rate used is given in next section).

The fuel costs for the production units in the DH systems are shown in Table IV. Prices for pellets, biomass, and natural gas are based on a prognosis made by the Danish Energy Agency⁴¹ and thereafter recalculated to USD/MWh. Municipal solid waste management normally means that the DH operator is allowed to charge the municipality a depository fee, which here is described as a negative fuel cost (an income) for waste. The depository fees for municipal solid waste are taken from annual statistical reports from the industry organization Svenskt Avfall¹⁵ and recalculated to USD/MWh. The exchange rates used for monetary conversions are 6.25 DKK/USD, 0.84 €/USD, and 8.49 SEK/USD.

Several fuels and some use of electricity are taxed, and some fuel use and production units are subsidized. Biomass-fired CHP units are, for instance, allowed green certificates for every MWh of electricity produced.⁴³ These certificates can be traded on a market with bidding and offers where the buyers are actors obliged to compensate for nonrenewable production/consumption. This is a system aimed at increasing the amount of renewable power generation. The price for green certificates has been declining recently and for this study the price for February of 2021 was used, which was 0.25 USD/MWh.⁴⁴ 

Natural gas and other fossil fuels are taxed according to the emissions of carbon dioxides that they cause. If the producing unit is participating in the European Union emission trade scheme (EU ETS), a deduction of 9% is allowed on the CO₂-tax. These fuels are also taxed with an energy tax in Sweden.⁴⁵ 

All utilities with waste incineration plants are subject to a waste incineration tax of 4.88 USD/MWh (125 SEK/ton municipal solid waste).⁴⁶ 

The benchmark number is set to 47.3 allowances per TJ heat delivered to DH during the current fourth trading scheme. HAL is the historical activity level in TJ heat (the median of the last three years of production). CLEF is carbon leakage exposure factor, which for DH is set at 0.3. The allowances are distributed regardless of the type of fuel. Thus, utilities using non-fossil fuels for production, i.e., biomass, can sell surplus allowances since they are not necessary for their own production of CO_2,eq-emissions. The price for ETS allowances is set on trading markets. Market analysts predict that the price for ETS allowances will increase, since the total number of ETS allowances is planned to be reduced in the future.⁴⁹ The impact of trading with ETS allowances is not expected to be the most crucial part of an actor's total revenue. Hence, in this study, a fixed price of 53.57 USD/ton CO_2,eq is used to represent the current trading value of allowances.⁵⁰ All taxes and subsidies used in this study are shown in Table V.

The total production cost, including fuel cost, taxes/certificates, and expected revenue for sold electricity in CHP units, is shown in Fig. 8. The production costs define the merit order of production units for each of the three setups (1-HoB, 2-WtE, and 3BioCHP).

The merit order is determined for every time step. IWH is always given the highest priority due to the contract between actors binding the DH operator to always use this heat. Waste incineration is given second priority since this is partly a waste managing service that needs continuous operation, except during maintenance when the unit is shut down. As shown in Fig. 8, the order may change during the year depending on electricity prices and maintenance of units.

In this section, the results from the simulations are presented. Initially, the economic effects of having P2H on the reserve market are presented and analyzed (Sec. IV A). Second, the heat balances of the DH setups with P2H production units on the reserve market are compared to the BaU cases (Sec. IV B). Finally, the TES performances, required capacities, and role for the heat balances are presented and analyzed (Sec. IV C).

Figure 9 shows the changed production costs for the three simulated DH setups. No income for produced heat is shown since this is unchanged for all simulations, and thus not of interest. The figure shows five bars representing the different scenarios for each setup. The first bar is the BaU-scenario without additional P2H. The following four bars are the economic outcome for the scenarios with different additional levels of P2H capacities placed on the reserve market. The red numbers are the approximated total costs with the given assumptions for this simulation. Table VI presents willingness-to-pay, i.e., the average bid, $c bid avg$⁠, per MWh_el/h that the DH operators can be expected to place for buying electricity via the reserve market before BaU would be economically favorable. This means that if the bidding price is below this average bid, P2H would have a lower cost than BaU.

Setup 1-HoB (Fig. 9 and Table VI) shows the greatest willingness to pay before the cost becomes equal to the BaU-scenario. For the scenario with 60 MW_el P2H capacity, the amount of heat from P2H units also competes with heat produced in the base load unit. This is explained by the restrictions in the algorithm where the heat produced in P2H is used first to avoid increased production in operating units, and second, to postpone the start of additional production units. This means that P2H production is not used to replace heat produced in operating units. The effect of this is that heat is accumulated in the storage unit. At the end of the year, storage contains 72% more heat than at the start of the year. In the end, this results in unnecessary amounts of electric power being purchased for heat production, which ends up in storage not being used. This overproduction of heat becomes a cost which reduces the willingness to pay by 47%. A similar pattern with accumulation of heat in the TES unit is seen for the scenario with 40 MW_el P2H, with about 29% more heat accumulated at the end, compared to the beginning of the year.

For setup 2-WtE (Fig. 9 and Table VI), the results show that P2H is slightly less economically favorable for the DH operator compared to what was seen in setup 1-HoB. The produced heat from P2H units replaces production in the biomass-fueled CHP unit. This means that the cost for used biomass fuel is reduced and the depository fee for receiving waste is shown in the figure as an income. In the scenario with 60 MW_el P2H capacity on the reserve market, heat is accumulated in storage for setup 2-WtE in the same way as for setup 1-HoB. At the end of the year, 45% more heat is stored compared to the beginning of the year. This reduces the willingness to pay by 30%.

For setup 3-BioCHP (Fig. 9 and Table VI), increased P2H capacity yields increased replacement of expensive HoBs fueled with pellets. This means that the utility can offer higher bids for consumption of electricity at these levels of P2H capacity. There is no reduced income from sold electricity in the biomass-fueled CHP unit in setup 3-BioCHP, as was the case for setup 2-WtE.

Figures 10–12 show how heat produced in P2H units replaces heat from other producing units for the three setups, respectively. Similar to Fig. 9, the bars in these figures represent the BaU-case and the cases with different levels of P2H capacities from left to right. In Fig. 10, the presented results show that heat from P2H units for setup 1-HoB gradually replaces heat produced from both boilers, but primarily from the second pellet boiler. In the scenarios with 40 and 60 MW_el P2H capacity, heat from P2H units replaces about 5% of the production from the base load unit. This 5% is replaced during maintenance of the second boiler, which indicates that the heat demand is supplied with heat from TES. In the scenario with the highest level of P2H capacity, however, there is an overproduction of 158 GWh_th heat which causes the accumulation of heat in TES described in Subsection IV A.

Figure 11 shows the distribution of heat production for setup 2-WtE. The heat from P2H units primarily replace the production from the biomass-fueled CHP unit and second replace production from fossil-based production. This is due to the naïve strategy for the TES units from which heat is supplied as soon as it is available. Similar results can be seen for setup 3-BioCHP (Fig. 12). In the case with the highest P2H capacity, there is, similar to for setup 1-HoB, also an accumulation of heat of about 95 GWh_th in TES.

In Fig. 12, the results for setup 3-BioCHP show that the peak load unit fueled with natural gas is the last to be replaced with heat from P2H. In the scenario with 60 MW_el P2H capacity, the fossil-based heat production has been totally replaced by heat from storage. The heat produced via P2H primarily replaces heat from the heat pumps and second heat produced from pellet-fueled HoBs. For this setup, no over-produced heat is accumulated in the TES unit for any of the scenarios.

In Fig. 13, key figures for the use of the TES units in the different DH system setups and additional P2H capacity levels are shown.

Figure 13 a shows the storage capacity relative to the annual heat demand (left y-axis) and the physical volume (right y-axis). Table VII presents the absolute numbers for the required storage capacities for the different setups and additional P2H capacity scenarios. There is a significant difference in storage capacity and volume between the 40 MW_el P2H scenario and the 60 MW_el P2H scenario, in particular, for setups 1-HoB and 2-WtE. This is explained by an accumulation of heat that is over-produced in P2H units and not used. This does, however, not occur for setup 3-BioCHP, where all heat stored is also used.

It is shown in Fig. 13(b) that the relative losses are kept fairly constant and at low levels. Only for setup 2-WtE and the BaU scenario, high relative losses are shown. This is explained by a sporadic use of storage in this case, with merely a few short charging and discharging periods, leaving the TES unit idle for 90% of the year. For all other setups and scenarios, the relative losses stay below 1%. However, the actual losses do increase with an increasing TES volume and/or capacity, as shown in Fig. 13(c). For setup 3-BioCHP, an overall lower loss compared to the other two setups is shown. This indicates more efficient use of the TES capacity for this setup, which is also seen in Fig. 13(d), which shows the TES degree of utilization, η_U. Setup 1-HoB reaches the highest η_U for the scenario with 10 MW_el P2H, but then continuously drops with increasing P2H capacity. Setup 2-WtE also shows a continuously lower η_U with increased P2H capacity, starting at 3.9 (10 MW_el P2H) and ending at 0.9 (60 MW_el P2H). Setup 3-BioCHP yields a fairly constant η_U for all scenarios starting at 3.4 (10 MW_el P2H) and ending at 2.2 (60 MW_el P2H). This indicates that the storage capacity is utilized to a similar extent for all scenarios. Finally, in Fig. 13(e), the energy efficiencies, η_E, for the TESs are shown. The only setup for which the η_E increases with increasing P2H capacities, is for 3-BioCHP. For the other two setups, the η_E decreases significantly with the largest P2H capacity and also TES capacity, due to the accumulation of over-produced heat in the TES.

This study investigated how different compositions of DH production units affect the economic feasibility for placing P2H electricity consumption capacity on the reserve market. The study also investigated the required storage size for applying heat production from P2H units activated on request from the electric network balance operator in parallel with conventional DH production. This was done by simulating three representative setups of DH systems and applying different levels of P2H production capacities. The change in total production costs for producing heat was compared with a BaU scenario without additional P2H capacity.

The first research question asked was what bidding price would be economically feasible for the three different DH systems setups. Table VI shows the maximum average biding price, $c bid avg$⁠, that DH utilities can offer per MW_el consumed in P2H units, without heat produced in conventional heat plants, and according to BaU, being cheaper than the heat from P2H. Any bidding price lower than the figure given in Table VI will always be economically beneficial. The highest possible economic benefit is shown for case 1-HoB. Case 2-WtE shows a slightly lower economic benefit. For both these cases, heat is accumulated in the storage at high levels of P2H. This yields a requirement for large storage capacity and reduces the revenue. For case 3-BioCHP, the economic benefit is the lowest of the three cases. Nevertheless, the revenue is increasing in positive correlation with the amount of installed P2H capacity.

The kind of reserve market that is simulated in this study does not currently exist. The current reserve market mainly focuses on reserve capacity for electricity production, and not consumption. It is, however, already possible to place bids as “Frequency Containment Reserve consumption” for existing energy storage. In addition, as shown by Ref. 8, electric boilers could replace must-run CHP units for negative balancing purposes, and this solution can be applied today. Furthermore, considering the report from Energiewende³⁴ regarding negative electricity prices being caused by inflexible production units, applying P2H would not only increase the useful share of VRE, but would also minimize the risk for negative electricity prices.

When comparing the actual prices during 2020 from the Frequency Containment Reserve—Normal and Frequency Containment Reserve—Disturbance markets in Sweden, the mean price per MW was 20.3 and 19.4 USD/MW_el, respectively, with a standard deviation of 9.7 and 8.8 USD/MW_el.⁵² In comparison with the average bid price calculated in this study, these values are considerably lower, hence indicating a profitability for P2H. Note that the prices from the reserve markets are for production.

The price for P2H is very close to the production cost for heat based on pellets (Fig. 8) for setup 1-HoB. For setups 2-WtE and 3-BioCHP, this correlation is not as clear. This is because several different production units are involved, so the price becomes a result of multiple parameters.

The second research question concerns the required storage size and the third question is how the stored heat would compete with other production. The size of required storage is significant in this study. This means that the storage strategy applied in this study is inadequate in terms of minimizing storage size. In addition, existing rock caverns used as TES are normally around 90–120 000 m³. Volumes above 150 000 m³ are rare. For instance, in the study done by Hast et al.,⁵³ a TES capacity of approximately 1% of the annual heat demand is suggested. From the simulations in this study, the storage capacities range from between 0.8% and 1.3% for the scenarios with 10 MW_el P2H capacity, and up to between 13% and 27% for scenarios with 60 MW_el P2H capacity. This indicates that the operation strategy of the storage is important in order to minimize the required TES capacity, but note that this is also influenced by the production strategies for other production units. For instance, if the storage is used to replace heat from baseload units to a greater extent than what is seen here, the storage size could be further reduced. In addition, if CHP units are only operated when there is an electricity demand, i.e., the production primarily supplies the electricity demand and not the heat demand, the storage could supply heat when there is no electricity demand and this would shorten the retention time for the heat as well as reduce the required size of the TES units.

With the naïve storage strategy, mainly intermediate units and peak-load units are replaced by heat produced from P2H. For setup 1-HoB, this means that heat produced in the second heat-only boiler was primarily replaced. Furthermore, for setup 2-WtE, the results show that heat produced in the biomass-fueled CHP unit was gradually replaced by stored heat previously produced in P2H units. This means that P2H production that is controlled by the electricity reserve market can reduce the CHP power balancing capacity due to heat supply competition from stored P2H-produced heat. This is especially the case when P2H is integrated into DH systems with waste incineration plants and CHP plants as intermediate load supply units. This is in line with the findings in Ref. 23. For setup 2-WtE, actually all power-balancing electricity production from the biomass-fueled CHP unit is replaced by heat from P2H. For setup 3-BioCHP, however, the CHP power-balancing capacity is not limited to a similar extent. This is explained by the fact that for setup 3-BioCHP, the biomass-fueled CHP unit supplies the base load, and the heat supply replaced by heat from P2H is primarily units used for intermediate and peak-load supply, i.e., heat pumps, pellet-fueled HoBs, and fossil-fueled HoBs.

Previous studies have concluded that revisions of tariffs, taxations, and grid fees are necessary for P2H to be realized as a power-balancing instrument. This is an important aspect for the business models applied by DH utilities. As pointed out in the literature review (Sec. II), placing bids for P2H capacities on the reserve capacity market is not a part of the current business models for DH utilities. The participation in a new third market area [(1) heat market, (2) day-ahead market, and (3) reserve capacity market] implies increased workloads and, most likely, a need for new competence and know-how. Since P2H is potentially an efficient way to decarbonize and/or reduce fuel use in the heating sector, P2H is a technology that policy-makers should consider thoroughly. One way would be to support investments in P2H-technology and TESs as a means of reaching climate targets of reduced CO₂-emissions. Another important issue is to create incentives for DH utilities to be active on reserve capacity markets by for instance, compensating extra for the P2H electricity balancing service. Such compensation could counteract barriers created by tariffs, taxations and/or grid fees. An extension of this compensation scheme could include the power-balancing electricity production from CHP units.

The method used to derive the system price for electricity in this study yields a fairly representative price for a power system with a large share of VRE production. However, the occurrence of negative electricity prices is not captured here, which has been a reality in, for example, Denmark.³⁵ Negative electricity prices are to a large extent caused by other inflexible production units in the power system, such as nuclear power plants and conventionally operated CHP plants.³⁴ In this study, however, these inflexible production units are assumed to have been phased out, thus reducing the risk for negative electricity prices. Nevertheless, conventionally operated CHP plants are still included in the model; furthermore, since electricity price is determined on a market considering several regions, there will be an impact from remaining inflexible technologies from surrounding regions. Hence, negative electricity prices could still be expected at some level. If negative electricity prices had been included these could affect the results, since this would possibly change the merit order and replace current base-load heat supply with heat from heat pumps to a larger extent.

The results show that in order to estimate the potential for P2H, it is necessary to consider existing production facilities. For setup 2-WtE, the heat produced from P2H replaced heat from the biomass-fueled CHP unit, and reduced the CHP power production. In the simulations the power production was gradually reduced from 115 GWh_el in the BaU scenario to 0 GWh_el in the 60 MW_el P2H capacity scenario. In setup 3-BioCHP, a similar trend was seen, but to a lesser extent. Nevertheless, the aspect of possible power-balancing production from CHP units in DH systems needs to be considered if and when P2H is implemented for balancing purposes. This is because of the consequences of heat supply competition, especially with CHP units.

A parallel P2H production setup as investigated here may be feasible, but depends strongly on the storage strategy and/or the production strategy of the other heat producing units. This means that coordination with other production technologies may reduce the required storage size.

Future policies, for increasing the share of renewables in power system, are recommended to consider incentives for P2H electricity consumption on reserve capacity markets through, for instance, compensation for power-balancing services. Such compensation could also help to overcome market barriers caused by tariffs, taxation of electricity consumption, and grid fees. Compensation should be designed to also include power-balancing services from CHP units. Furthermore, as P2H will require access to large-scale and long-term TESs to reach its full potential, investment subsidies for these are also important. Investments in P2H should be considered as a significant measure to decarbonize DH sectors and/or reduce biomass demand. Thus, the suggested system changes in this study could potentially contribute to reaching environmental targets related to biodiversity as well as reductions in GHG-emissions.

Suggested future work is to develop a production strategy that allows heat produced in P2H to replace base-load production and thus reduce required storage capacity, without increasing the risk of reduced revenues. In addition, the possibility to use stored heat for other purposes than local heat load supply, i.e., sorption chillers cooling production or carbon capture from flue gases should be further investigated.

The overall conclusion from this study is that the potential for P2H to be economically feasible for DH operators, when acting on the reserve market, depends on the composition of production units in the DH system. All investigated system setups reduce the average heat production cost when offering P2H capacities on the reserve capacity market. The average heat-production cost reduction ranged between 1.25 $/MWh_th and 1.08 $/MWh_th depending on DH system setup. Furthermore, large additional P2H capacities tend to yield an accumulation of over-produced heat, which requires additional heat loads, other than the existing DH load, in order to make it economically beneficial.

The maximum average bidding price per MW_el that the utility can offer on the reserve market before the BaU is cheaper (research question number one), depends significantly on the heat production unit composition of the DH system. This is because of the different heat production costs for the setups. For setup 1-HoB, the maximum average bidding price is 150 $/MW_el, while the corresponding bidding price for setup 3-BioCHP is between 60 $/MW_el and 80 $/MW_el. To engage all DH operators into offering P2H electricity consumption bids on the reserve capacity market, the average bidding price should be below 60 $/MW_el, i.e., the lowest indicated average bid cost in the scenarios.

A conclusion corresponding to research question number two is that P2H units in combination with TES enables production of heat independently of the heat load, and is here shown to be a feasible way to facilitate P2H on the reserve market. The results indicate that the required storage sizes increase relative the installed P2H-capacity. This is explained by a decreasing utilization of the storage units at higher P2H capacities. For a P2H electricity consumption capacity of 60 MW_el, and all stored heat is used, the required storage capacity is about 13% of the annual heat load.

In response to research question three, the results show that heat produced in P2H units will replace heat mainly from intermediate and peak-load supply units. Such units are generally biomass-fueled CHP or heat only boilers as well as fossil-fueled boilers. Waste incineration and industrial waste heat mainly supply base load and are therefore not replaced to significant extent by P2H-produced heat. It is, however, noteworthy that replacing intermediate load CHP production may impair the possibility for providing CHP power-balancing capacity from the DH system.

A final reflection is that in future energy systems, it is reasonable to expect an increased utilization of (and storage possibilities for) heat from, for example, industrial and commercial processes, solar thermal, and geothermal heat sources. It is therefore crucial to create TES operation strategies and DH system setups that are dynamic in terms of storing a variety of different heat sources. The possibilities and limitations of multi-purpose storage of heat in one single TES, however, needs further investigation.

The authors have no conflicts to disclose.

Svante Monie: Conceptualization (equal); Formal analysis (equal); Methodology (equal); Resources (equal); Software (equal); Writing – original draft (equal). Mohammad Reza Hesamzadeh: Writing – review & editing (supporting). Magnus Åberg: Formal analysis (equal); Supervision (equal); Writing – review & editing (equal).

The data that support the findings of this study are available from the corresponding author upon reasonable request.