Reactive Power Procurement: A Review of current trends[footnoteRef:2] [2: The paper is based on a report prepared for National Grid ESO under the context of the Power Potential project. However it represents the views of the authors and should not be taken to be the views of any other party associated with the Power Potential project.] By Karim L. Anaya[footnoteRef:3] and Michael G. Pollitt[footnoteRef:4] [3: Corresponding author, email: k.anaya@jbs.cam.ac.uk.] [4: Email: m.pollitt@jbs.cam.ac.uk.] Energy Policy Research Group, Judge Business School, University of Cambridge, Trumpington Street, CB2 1AG Abstract This paper reviews the international experience in the competitive procurement of reactive power and other electricity ancillary services. It involves system operators from different jurisdictions including Australia, the United States and Great Britain. The paper evaluates the different procurement mechanisms, related compensation schemes and looks at more competitive ways for procuring reactive power. It appraises a novel approach (from the Power Potential initiative in the UK) for contracting reactive power services from distributed energy resources using a market-based mechanism. A conceptual auction design applicable to the procurement of reactive power is also discussed. Our findings suggest that competition in reactive power is very limited in comparison with other ancillary services (such as frequency regulation and capacity reserves) and the energy market. The introduction of more market-oriented mechanisms and resources (such as distributed energy resources) for acquiring reactive and active power services by the system operator gives rise to new opportunities and new ways to deal with voltage stability issues. Power Potential brings the opportunity to trial the technical and commercial solution, new roles and new interactions required in the introduction of a competitive reactive power market. Introduction The electricity sector is undergoing a transformation due to the introduction of new market participants (e.g. distributed energy resources – DER) in wholesale and retail markets. Competition is increasing in both markets and established business models are being challenged [1]. However, competition in ancillary services[footnoteRef:5] (with a focus on reactive power) is still limited [2]. Reactive power is one of the ancillary services that system operators need to procure in order to maintain the local electricity network voltage within the required limits to secure the proper operation of the electrical power equipment[footnoteRef:6]. The lack of competition in this market can be explained by the fact that units of reactive power (known as VARs) do not travel well. This means that the source of reactive power should be placed closer to where it is needed due to high losses in transmission lines (i.e. it should be provided locally) [3]. Traditionally, transmission connected generators have been the main source of reactive power support provided by third parties[footnoteRef:7]. These services are contracted by system operators using different remuneration schemes. [5: Ancillary service refers to those supporting services that are required for the correct operation of the electricity system (transmission and distribution). ] [6: The voltage limit requirements can be altered by the different components of the power system such as loads and transformers, overhead lines, underground cables, others [4].] [7: This is especially in GB where most of reactive power services are provided by transmission connected generators via mandatory schemes [5]. ] Due to integration of renewable electricity generation in the system, demand for ancillary services is expected to increase. This is due to the displacement of dispatchable generation (e.g. coal, natural gas, hydro, biomass, nuclear power[footnoteRef:8]) in favour of intermittent generation like solar and wind power. Therefore, more interventions are needed to keep the right balance between demand and supply. Even well-organised markets such as the Nordic power market are experiencing significant increases in frequency deviations over recent years (circa 100% , 2003-2017). In order to reverse this effect the electricity system requires more activation, automatization and small bid-sizes [6]. A new balancing concept[footnoteRef:9] has been recently proposed by the Nordic TSOs and a gradual implementation will take place in the next five years. The New Nordic balancing model will help to deal with volatility created by the increasing share of intermittent generation. The energy transition is also challenging the German electricity market (due to increasing penetration of renewables and decreasing market share of conventional and controllable generators). All this increases the demand for ancillary services especially the need for more instantaneous reserve and more actions for the provision of reactive power [7]. [8: Nuclear is also an option but for economic reasons it is more used as baseload rather than for dispatch. ] [9: The new approach means the transition from a frequency-controlled balancing at a Nordic level to Modernised area control error (MACE), among others [27]. ] Ancillary services costs still represent a small amount of the total wholesale electricity costs, but this could change with the upward trend of renewable generation in the system (transmission or distribution-connected) [footnoteRef:10]. Increasing DER leads to a higher demand for reactive power [8] and to higher direct costs [9] due to the need to add reactive capabilities to the existing and new renewable generation (e.g. via power electronic converters in wind generation and inverters in solar PV). However, due to technological progress and new network codes and standards, it is expected that generators have this capability by default and at lower cost[footnoteRef:11]. On the other hand, DER may increase competition in the reactive power market by reducing the use of non-competitive procurement mechanisms and the acquisition of transmission connected assets by the transmission network operator. DER may deliver greater benefits due to the provision of multiple services, including reactive power [10]. In general and in comparison with the traditional transmission connected generators, DER may promote the proliferation of small bid-sizes. [10: In CAISO market, the share of ancillary service costs over the total wholesale costs of electricity is increasing with a share of 0.7-1.85% between 2015 and 2017 [28]. In Germany ancillary services represented 4-6% of the total market value (wholesale market) between 2012 and 2016 [29]. ] [11: In Germany, where solar PV can cover up to around 45% of momentary electricity demand [30], distributed generation units have been subject to specific rules for a long time (to ride-though low-voltage events, be remotely observable and dispatchable, etc.) and DSOs have been required to apply forecasting models (but without specifying the use and coordination with the TSOs) [31]. ] The majority of the current literature in reactive power is concentrated on the technical or economic aspects of reactive capability from different resources and enhanced procurement methods. Some studies evaluate the impact of the grid integration of intermittent generation on reactive power in terms of quality issues and planning [11,12]. Other studies assess the capability of DER in the provision of reactive power to utilities and microgrids [13,14] or evaluate the current methods for reactive power compensation and propose enhanced methodologies including the optimisation of active and reactive power costs [15,16]. A different group of studies explore competitive market models for reactive power procurement for long and short-term supply using optimisation algorithms [17,18]. Due to the proliferation of DER, different studies evaluate also the capabilities and influence of solar PV inverters in reactive power and voltage support [19,20] and the capabilities of electric vehicles for fast vehicle-to-grid reactive power dispatch [21]. Most recent studies discuss a joint active and reactive power market model at distribution level [22,23]. A limited number of studies refer to specific surveys of ancillary services including reactive power [24], some concentrate only on contestable ancillary services such as frequency regulation and reserves [25] and others focus on a mix of both [3, 26]. Current trends and new developments in the management and procurement of reactive power (including compensation schemes, enhanced network codes, role of DER), and the procurement of reactive power services by system operators from DER, need further discussion in the current literature about reactive power. This is because reactive power remains a poorly understood aspect of electricity supply, with little attention given to the practicalities of how to develop existing reactive power procurement methods. Existing procurement methods, often involving the payment of fixed prices, were developed at time when reactive power was not very significant. Now we face a situation where demand for reactive power is rising and there are more potential sources of reactive power. This gives rise to opportunities for a much more competitive procurement regime for reactive power than hitherto. This study reviews the international experience in the management, procurement and compensation mechanisms for reactive power in its role of supporting system operation as an ancillary service. We look at system operators from Australia, USA and GB and other key European jurisdictions (including Nordic countries). It also explores a pioneering initiative (the Power Potential trial) that is being implemented by National Grid Electricity System Operator in GB (NGESO) and UK Power Networks, the largest Distribution Network Operator (DNO) in the UK. Power Potential seeks to contract with DER for the provision of reactive and active power services in the southern region of GB (in and around London) using a competitive mechanism. Some specific lessons are identified for the future procurement of reactive power by system operators from DER using market-based mechanisms. The structure of this paper is as follows. Section 2 discusses the role of reactive power as an ancillary service, the economics of ancillary services, key upgrades in network codes that enhance DER reactive power capabilities and the future role of DER in reactive power support and other markets. Section 3 explores current methods for managing and procuring reactive power by selected system operators in Australia, the USA and GB. Section 4 describes the Power Potential initiative. Section 5 discusses the main findings of this study and the lessons for the procurement of reactive power using market-based mechanisms. Section 6 concludes. About Reactive Power 2.1 Reactive Power as an Ancillary Service The safe operation of the electricity network creates a need to procure reactive power to compensate for the reactive power impacts of generators and loads at different nodes. Reactive power ancillary service constitutes a separate market from the active power (or real energy) market. Traditionally system operators have the primary responsibility to acquire these services mainly from generators (independent or affiliated) using dispatch instructions (a voltage schedule), however some exceptions may apply. In Australia, the primary responsibility is given to Transmission Network Service Operators instead, see section 3.1 for further details. An intermediate approach is observed in PJM where the transmission providers administer the purchases and sales of reactive power supply with PJM as a counterparty [32]. In contrast with other ancillary services such as frequency regulation or capacity reserves where market-based mechanisms are used for their acquisition, reactive power is less exposed to competitive mechanisms. Thus, Independent System Operators (ISOs) from the USA[footnoteRef:12] procure operating reserves together with energy when clearing either the day-ahead market, real-time market or both markets [33]. In Australia, frequency response ancillary services (FCAS) are procured together with the energy market [34]. This practice is called co-optimisation which results in better price formation and brings important savings, 30-50% of ancillary services costs [35]. Some initial ideas of co-optimisation are emerging in Europe based on the EU Guideline on electricity balancing [36]. In GB, some ‘Response and Reserve’ products (e.g. Fast Reserve, Firm Frequency Response, and STOR) are acquired by NGESO using tenders, however co-optimisation has not been put in practice. By contrast none of these markets procures reactive power competitively as a matter of course[footnoteRef:13]. Table A1 from the Appendix compares the different ancillary services procured by ISOs from the USA and NGESO from GB. [12: In this paper ISOs refers to both: Independent System Operators and Regional Transmission Operators. ] [13: The exception is Australia, however in this case AEMO acts a procurer of Last-Resort, see Section 3.1. ] We observe that frequency response services and reserves are usually procured in the day ahead market (DAM) or real time market (RTM) in the USA, while in GB the procurement period is longer (one month is the minimum). Shorter procurement periods favour the use of co-optimisation, allowing to clear both markets (energy and ancillary services) simultaneously. In GB, NGESO is implementing a frequency response (FR) auction trial with weekly auctions. One of the aims of this trial is to allow greater participation especially from variable generation (e.g. solar, wind). There are limited competitive mechanisms for the procurement of reactive power. In the USA competition is null and in Europe is very limited, with some few exceptions. In Belgium, Elia procures reactive power via tenders only[footnoteRef:14]; while TSOs from GB (NGESO) and Denmark (Energinet) use occasional tenders but mainly for covering additional reactive power needs (i.e. enhanced reactive power service – ERPS - in GB). Reactive power services are not incorporated in the dispatch process and are usually provided under mandatory procurement by TSOs (via bilateral contracts), excluding Belgium and partially in Netherlands [46]. The local nature of reactive power and the limited number of potential providers (usually transmission connected generators) contribute to the lack of competition in the provision of reactive power services. Table 1 describes the procurement of two ancillary services in Belgium, GB, Germany and Nordic countries[footnoteRef:15]. For the rest of ancillary services see [47]. [14: Even though the existence of a market-based mechanisms for reactive power procurement, a new framework proposed by Elia is in evaluation [46]. ] [15: The Nordic countries operate the Nordic power market: “The world’s oldest and most successful international power market” [48], p. 5. ] Table 1: Frequency Containment Reserve and Reactive Power Procurement in selected countries We observe that the Nordic countries promote a pure market-based approach in the procurement of Frequency Containment Reserve (FCR) (also known as primary reserve)[footnoteRef:16]. In the Nord Pool system there is an obligation to maintain reserves and TSOs are also responsible for maintaining reserves in their own control area (island operation)[footnoteRef:17]. In fact, Nordic countries are the ones with the maximum time resolution (hours rather than months) [footnoteRef:18] in the procurement of FCR and the majority of them use a marginal pricing (pay-as-clear) approach[footnoteRef:19]. The Nordic model looks superior (in different aspects) to similar models in the rest of Europe [27]. The country with the lowest time resolution is GB (monthly auctions) however trials (Frequency Response Auctions trials) are currently in progress to reduce the resolution to weekly auctions. On the other hand, reactive power in all these countries (including Germany) is still provided under a mandatory approach, with some exceptions (in GB with ERPS). It is also observed that in Sweden and Finland there is no financial compensation for reactive power (similar to CAISO) and that only in Sweden the distribution system operator may act as a provider of reactive power [47]. [16: There are different types of reserves, including active Frequency Restoration Reserve (aFRR), manual FRR (mFRR) and replacement reserves. The names and size of these markets may vary across TSOs. ] [17: Country specific reserve obligations have been set between the Nordic TSOs.] [18: Here the maximum time resolution is better and closer to real time (hours rather than months). ] [19: Pay-as-clear (rather than pay-as-bid) is the method suggested in “The settlement of balancing energy for standard for balancing products and specific balancing products”, Art. 6 (4) of Internal market for electricity regulation (recast, June 2019). ] At the distribution level, there is no currently (March 2020) reported procurement of reactive power services using competitive mechanisms - as business as usual rather than in a trial- for use locally or by the transmission system. However, this could change in the near future when a more active role for the DSOs and where more coordination between DSOs and TSOs is expected in the procurement of non-frequency ancillary services (e.g. voltage control, black start). Reactive power at the distribution level is currently managed via connection agreements by limiting the values of power factors (PFs) in agreement with the national or state regulation on Network Codes (i.e. Distribution Code in GB, Interconnection Handbook in California) and also through financial incentives. In the USA, the IEEE 1547 standard rules the Interconnection and Interoperability of DER with electric power systems. Further details are provided in section 2.3. 2.2 The Economics of Ancillary Services This section explores the economics of ancillary services in Nordic countries, GB and Belgium by looking at two things: the trend of ancillary services costs and the size of ancillary services market versus energy market, over the period 2011-2018. The most comprehensive information – for all the countries we look at - is available on frequency response and reserve ancillary services. A cost breakdown for non-frequency ancillary services was not possible in all the countries, but in GB, Belgium and Denmark[footnoteRef:20]. In the identification of frequency ancillary services (from the annual reports) it was noted that even in the Nordic countries the names of these types of services differ. This required careful selection and evaluation when examining their costs[footnoteRef:21]. Figure 1 depicts the trend of frequency ancillary services costs (except Denmark which considers both[footnoteRef:22]) over time. [20: Separated reactive power and black start costs are available only for GB and Belgium. In the case of Denmark an aggregated figure is provided for non-ancillary service costs (reactive power, black start, others) from 2013 onwards only. Finland and Sweden do not compensate for reactive power. ] [21: Based on the annual reports we select in Denmark: cost of AS, purchase of regulating power, and non-frequency ancillary services costs (the last of these is only available from 2013); in Norway: primary, secondary and tertiary reserves, net regulating power, and special adjustments; in Sweden: primary regulation, power reserve, disruption reserve; in Finland: cost of reserves; in Belgium: FCR (R1 Contracting), aFRR (R2 Contracting), mFRR (R3 standard contracting, R3 Flex ICH contracting and R3 Flex contracting), reactive power, black start. In GB from monthly balancing services summary (Auto summary): STOR, fast start, fast reserve, frequency response, reactive power, black start. ] [22: In order to keep the same base year for the index (2011=100) we decided to include the costs of frequency and non-frequency ancillary services in the case of Denmark. A cost breakdown is only provided since 2013. ] Figure 1: Trend of frequency ancillary services costs in selected countries In general, it is observed a downward trend until 2017, with some exceptions in Finland and Belgium. For instance in Finland this was driven by the increase in the purchase of frequency control reserves in the hourly market. An important drop is observed in Belgium after 2014 due to the introduction of short term auctions for the combined reservation of FCR and aFRR [50]. On the other hand, the downward trend of total costs in other countries may be due to the strengthening of regional cooperation on sharing capacity reserve [49]. Norway is the exception here with an increase of total cost in 2013 and 2014 due to the launch of the secondary reserve market and long term disconnections [51, 52]. A slight decrease is also observed in GB, with a reduction of 18% in 2017 compared with 2011 prices. The higher costs in 2018 may be explained by the important increase in the cost of reserve products. For instance in Belgium this upward cost was primarily attributable to an increase in the price of mFRR (from €3.4/MW/h in 2017 to €9.9/MW/h in 2018) [53]. In Norway, weather issues increased the purchase of reserves during 2018 [54]. In Denmark higher prices for frequency reserve were behind this rise [49]. In the comparison of ancillary services versus energy markets, the size of the energy market cost was estimated by the product of day-ahead prices (spot prices) and the final electricity consumption[footnoteRef:23] in each country[footnoteRef:24]. Table 2 illustrates the share of ancillary services over total energy costs. [23: Using electricity consumption overestimates the true number as we are ignoring system losses which must be covered by energy market purchases, but this should not have much impact on the trend shares in the table.] [24: Final electricity consumption and day-ahead prices refer to average annual figures. In order to mitigate important annual fluctuations of spot prices, average national power cost over the whole timespan (period 2011-2018) was used instead. Regarding Nordic countries with more than one bidding areas, the average day-ahead prices were computed. There are six areas in Norway, four zones in Sweden, two zones in Denmark and one zone in Finland.] Table 2: Share of ancillary services costs over total energy costs in selected countries It is observed that the size of the ancillary services market is small compared with the energy market. Denmark is the one with the highest average costs of frequency ancillary services during the period 2011-2018 representing 7.2% of total energy costs. On the other hand, Norway is the one with the lowest share, around 1.1% of total energy costs. Even though the share of intermittent renewables in total production in Denmark increased sharply between 2011 (27.8%) and 2018 (50%), the share of ancillary service costs over energy costs did not. One explanation can be the downward trend of ancillary services costs over time (with an average reduction of around 20% compared to 2011), see Figure 1. This reduction may be explained as a result of economies obtained through the strengthening of the regional cooperation on sharing reserve capacity, joint procurement with Sweden, among others, [49, 55, 56]. It is also observed, as expected, that most of the costs of ancillary services are concentrated on frequency ancillary services. In terms of the impact that reactive power and black start can have on total ancillary services costs, we note that in Denmark this is 2.5% of total energy costs[footnoteRef:25] while in GB and Belgium they are 0.7% and 0.4% of total energy costs on average respectively. From this, we notice that ancillary services market is not likely to be large relative to the energy market, especially the non-frequency markets (i.e. reactive power). [25: A disaggregation is not possible in Denmark, it may include not only reactive power and black start but other non-frequency ancillary services. ] 2.3 Adaptation of Codes and Standards that favour Reactive Power Support Network Codes for grid connected generators are evolving in line with the integration of more renewable generation in the system and the need to maintain system security and stability. An example of this is the Network Codes that are being updated by different EU member countries as part of the implementation of the Third Package. Among these codes we have the Requirements for Generators (RfG) connection code applicable only for new generating units (with 0.8kW and up) called Power Generating Modules (PGMs) [57]. The aim of the RfG code is to harmonise standards for generating units connected to the grid, improving the integration of internal and cross-border electricity markets. ENTSO-E develops the codes and is also responsible for monitoring their national implementation in EU member countries. PGMs can be classified as Synchronous Power Generating Modules (SPGMs) or Power Park Modules (PPMs)[footnoteRef:26]. However, not all the types of PGMs need to comply with the RfG requirements. Among the exceptions are those classified as “emerging technologies” [58] and also storage devices except for pump-storage [57]. [26: PPM refer to asynchronous generation and to a unit or aggregated units generating electricity. ] Under the RfG PGMs are subject to specific technical requirements arranged in four bands (Types A-D) based on the connection voltage (up to 110kV from Types A-C and over 110 kV for Type D) and capacity (thresholds are proposed by national TSOs, ratified via industry consultation and approved by the regulatory authority)[footnoteRef:27]. Different sizes of thresholds have been adopted by EU member countries, in line with the recommendations provided by ENTSO-E. Table 3 compares the ones adopted by the transmission system operators from GB (NGESO), Northern Ireland (SONI) and Belgium (Elia). [27: RfG entered into force as European law on the 17 May 2016. In GB compliance with the code started on 17 May 2019. In the case of distribution, Engineering Recommendations (ER) G98 and G99 apply (called originally G83 and G59 respectively).] Table 3: Capacity Thresholds for selected countries Reactive power capabilities are required for Types B-D and different types of reactive power control modes may be required too. Reactive power capability for Type C and D is defined in terms of voltage against Q/Pmax for lead and lag modes, instead of the traditional power factor. There are specific reactive power capabilities for different types of PGM, such as SPGM, PPM and Offshore PPM (OPPM). In GB voltage control (type of reactive power control) is usually the preferred choice for both transmission and distribution connected generation. Table A2 from the Appendix shows the reactive power capabilities for the different bands and type of PGM agreed in GB in line with the new EU connection code requirements for RfG added to GB Codes[footnoteRef:28]. [28: A similar approach has been proposed to the Distribution Code, in agreement with the new G99 and G98. ] In terms of the trend in size of generating units, we observe that the number of small-scale distribution-connected generators is increasing over time, which is also in line with DER expansion. For instance, in GB, it is observed that most of the distribution-connected capacity is placed in three bands (Types B-D) representing 88% of the total (by Nov. 2015). Figure 2 illustrates the generator size band (and associated capacity by Nov. 2015) proposed in GB. However, in the future a reduction of this share is expected. By the end of 2021 most of new distribution connected capacity will be categorised as Type A and B representing around 66% of the total. From this figure, Type A is the one with the largest share, 46%, in comparison with its respective share in Nov. 2015, with only 12% [60]. Figure 2: Generation by band in GB Due to this fact, the extension of reactive capability requirements to Type A generators, could be feasible in the future, in line with the reactive capability requirements set in the revised IEEE 1547 Standard. The revised IEEE 1547 Standard[footnoteRef:29], released in 2018, set new requirements for reactive power capability (absorption and injection) applied to DER (including control functionalities). A progressive revision of the original/first standard (IEEE 1547 Standard 2003) started years ago driven by the increase in DER penetration. The revised IEEE 1547 Standard established different points of evaluation of the reactive power requirements which vary depending on the size of DER and average load demand type of DER (retail DER or wholesale DER)[footnoteRef:30]. The revised standard identifies two categories for reactive power and control requirements, Category A and Category B. Category A covers the minimum requirements applicable to all the state-of-the art DER. Category B covers in addition to those applicable to Category A, additional requirements. Different kinds of control models are also proposed, some of them mandatory and others optional. For both categories, the reactive power capability is required for an actual power output equal or higher than 20% of nameplate active power rating [61]. [29: See: https://standards.ieee.org/standard/1547-2018.html] [30: Wholesale DER are those with aggregated DER nameplate rating of at least 500 kW and with an average load demand equals or less than 10% of the DER nameplate rating [61]. ] In contrast with the new RfG connection code in Europe, the IEEE 1547 is not legally mandated but a set of best practices and guidance for connecting DER. However, in the USA most states have adopted the earlier IEEE 1547 Standard and amendments to it (which is reflected in their interconnection rules), then the adaptation of the 2018 IEEE 1547 Standard is something expected too. 2.4 The Role of DER in Reactive Power Procurement Unlike transmission connected generators, DER in general is not necessarily required to provide reactive power support to control local voltage levels. In fact, DER and its participation in organised wholesale markets has been concentrated on the provision of demand side response services (mainly from industrial and commercial businesses and most recently from residential customers) in energy, capacity and related ancillary services markets (e.g. Demand Response Auction Mechanism - DRAM in California, Demand-Side Ancillary Service Program - DSASP and Demand-Ahead Demand Response Program-DADRP in New York, Power Responsive in GB) using both auctions and non-competitive mechanisms. As such, DRAM, a pay-as-bid auction, represents one of the pioneers in the use of DER to aggregate demand response directly to the CAISO market using a market-based approach[footnoteRef:31]. In contrast with the traditional demand response practices, DRAM is procured by the electricity distribution utility (instead of the system operator). Through DRAM, Investor Owned Utilities (IOUs) make a capacity (also known as resource adequacy - RA) payment to demand response (DR) aggregators. IOUs acquire this capacity only and do not dispatch the resources. The IOUs are not allowed to claim revenues that can be received by the bidders (technically offerors) from the energy market. [31: The implementation of DRAM has been via trials. Four DRAM pilots have been concluded (DRAM 1-4). The California Public Utilities Commission [69] has authorised the extension of DRAM. Improvements in the program design and oversight process to monitor the progress are among the conditions imposed by CPUC for its extension. ] Reactive power support from DER is expected to take a more active role in this in the future. This is reflected by the introduction of new requirements such as those specified in Network Codes and standards, the use of advanced technologies such as smart-inverters in DER and the upward trend in DER capacity. According to [13] inverter efficiency is the most important parameter in the provision of reactive power by DER. The increase of DER capacity and the decline of centralised generation imply that the use of DER capabilities will be important to support both transmission and distribution system reliability [62]. However, the level of participation of DER in different balancing markets is not the same across jurisdictions. For instance, Belgium, France, Ireland and Switzerland have the most advanced balancing markets for DER and demand response, while GB is close behind due to the complexity of ancillary service products and market structure [63]. Increasing levels of DER participation in the wholesale market, also imply the need for greater DER visibility by system operators (at the transmission level) and additional coordination between transmission system operators and electricity distribution firms [64]. Different proposals are currently being evaluated in GB [65] and Australia [66]. In addition, in regions with a high level of DER penetration, electricity distribution firms can become a source rather than sink of reactive power [67]. DER can also introduce additional system complexity. This implies that “trials” are required to measure and evaluate the effectiveness of DER in providing reactive power and voltage support [68]. 3. Reactive Power Procurement: The International Experience This section evaluates the international experience in reactive power procurement by selected system operators: Australian Energy Market Operator (AEMO), NGESO from GB and four ISOs from the USA (CAISO, PJM, NYISO, ISO-NE). 3.1 Australia AEMO identifies three major categories of ancillary services: Frequency Control Ancillary Services (FCAS) Markets, Network Support and Control Ancillary Services (NSCAS) and System Restart Ancillary Services (SRAS). Voltage Control Ancillary Services (VCAS) is a subcategory of the NSCAS that relates to reactive power services. NSCAS is classified as a “non-market service”, which means that these are not acquired by AEMO as part of the spot market [70], in contrast with other ancillary services such as FCAS. Based on the Rule 2011 No.2 [71], Transmission Network Service Providers (TNSP) have the primary responsibility for meeting the NSCAS needs in the National Electricity Market (NEM) starting on April 2012. If this gap[footnoteRef:32] remains unmet by the TNSP, AEMO will seek tenders for NSCAS providers under ancillary services agreements. AEMO acts as NSCAS procurer of Last-Resort and will acquire NSCAS only to ensure power system security and reliability of supply on the transmission networks. TNSP may acquire NSCAS under connection agreements or network support agreements [60], however they aim first to make maximum use of the existing reactive resources. [32: Represented by the difference between the NSCAS needs of the NEM power system (arising within a 5-year horizon) and the NSCAS that the TNSPs predict to be procured. ] Procurement of NSCAS under network support agreements is an option for reactive support beyond the performance standards [72]. The trend costs of ancillary services for the period 2012-2018 is illustrated in Figure 3. Figure 3: Trend of Ancillary Service costs in Australia A significant reduction of reactive power costs can be observed, in line with the National Electricity Amendment, Rule 2011. By the end of 2018 reactive power costs represented only around 5% of the total ancillary costs procured by AEMO. The section below discusses the tender mechanism applied by AEMO for the procurement of NSCAS with a focus on reactive power services (VCAS). VCAS Tenders AEMO distinguishes mainly two kinds of VCAS modes of operation in the tender process for acquiring NSCAS: VCAS Generation Mode and VCAS Synchronous Condensor Mode. In Generation mode, VCAS represents the amount of reactive power capability (generation or absorption) by the NSCAS equipment in excess of the performance standard for reactive power for the NSCAS equipment supplied up to the connection point to the transmission network. In Synchronous Condensor Mode, refers to the reactive power capability (generation or absorption) when the generating unit is not producing active energy. The types of product are not limited by the operation modes of unused reactive power capacities of the generating units previously described. Other types of reactive plants can also compete such as capacitors and reactors[footnoteRef:33], SVC, STATCOMs, HVDC/HVAC transmission lines, etc. [73]. Participants are free to offer their best solution. However, some of the tender requirements may be more specific than others. [33: After the contracting period this equipment may be included in the TNSP’s regulatory asset base (RAB) with a zero capital value in the RAB [75]. ] The request for reactive power service may be for different term lengths [73]: a. short term: up to 12 months with the option to extend the service for 12 additional months, usually for existing facilities; b. long term: for a period of 5 year or longer, installations of new or utilisation of existing reactive plants; c. a combination of both, short term with existing installations until the construction of long-term reactive power equipment. Depending on the mode, participants (with winning selected offers) are subject to specific payments, see Table 4. There are two different payment structures based on the type of the generating unit operation mode. A compensation payment applies only when the generating unit is constrained off to generate or absorb reactive power during a trading interval. A Testing charge refers to the cost of specific tests that will be paid by AEMO. For additional details about the type of payments and their calculation see [74]. Table 4: Payment Structure In the evaluation of the tenders, AEMO assess the optimal combination of reactive power services taking into consideration the locational effectiveness of each VCAS equipment (depicted in a map provided by AEMO) at the least cost possible. AEMO does not provide details of the evaluation criteria (quantitative or qualitative) but only a general list of the criteria to be considered. 3.2 Great Britain In GB, NGESO procures a larger number of ancillary services (in comparison with the SOs from the USA and Australia). As of as January 2018, there were over 20 ancillary services with different procurement methods (e.g. tenders, mandatory and bilateral agreements). According to [63], one of the reasons behind this large number is the introduction of products by NGESO whose requirements are tailored to different technologies (instead of keeping a technology neutral approach). The acquisition of reactive power ancillary services (leading or lagging) is based on three mechanisms: Obligatory Reactive Power Service (ORPS), Enhanced Reactive Power Service (ERPS) and Transmission Constraint Management (TCM)[footnoteRef:34]. Reactive power can be provided by synchronous and non-synchronous generators which are subject to different PF range requirements, 0.85 (lag)/0.95 (lead) and 0.95(lag/lead) respectively. Figure 4 depicts the trend of balancing services costs in GB. [34: In some cases, a fourth method applies. This happens when NGESO procures reactive power via the balancing mechanism. In this case, generators are compensated for changing their real power positions to meet reactive power requirements. ] There is an upward trend in total balancing services costs. Reactive power ancillary service costs, represented mainly by those incurred under ORPS, are around £80m per year and represents circa 10% of total balancing services costs (average annual figures, 2013/14 – 2018/19)[footnoteRef:35]. It is also noticed that requirements for reactive power absorption (leading) during this period is the one with the largest share (around 85%). A description of the reactive power ancillary services is provided below. [35: Reactive power costs do not include those for managing voltage constraints, currently grouped in the ‘constraints’ cost category which covers both: thermal (active power) and voltage constraints. Reactive power costs due to voltage constraints are around £50m per year [42]. ] Figure 4: Trend of Balancing Services Costs and reactive power utilisation and costs in GB ORPS relates to the capacity for absorbing or generating reactive power to manage system voltages. This is a mandatory service for transmission connected power stations (generally over 50 MW) that are subject to the Grid Code (CC 6.3.2). In agreement with majority of European countries [47], this is the most common approach for procuring reactive power services in GB. Under ORPS, service providers receive a default payment for utilisation (£/Mvarh) that is updated monthly in agreement with market indicators (Schedule 3 of the CUSC). The default payment rate amounts to £3.27/Mvarh[footnoteRef:36]. A mandatory service agreement (MSA) is required to be signed by service providers. ORPS is the most common way to acquire reactive power services by National Grid. Over the last ten years, the requirement for reactive power absorption has increased (due to the downward trend in the demand for active power) and this trend is expected to continue [42]. [36: Average figure for the period Jan. – Dec. 2018, authors’ own estimations (2018 prices) and up to £4.34/Mvarh (average) in the Power Potential area with combining utilisation costs and repositioning [79]. ] ERPS is procured via tenders and applies to power stations whose reactive capability exceeds the minimum technical requirements of ORPS (set out in the Grid Code) and for those that wished to be paid at a rate different from ORPS. Tenders are held every six months and the delivery period is for a minimum of 12 months and thereafter in 6-month increments. The evaluation criteria for the selection of offers are set in the CUSC and considers economics (market price versus default price), intrinsic capability value (tendered reactive service versus alternative of National Grid reactive assets), among other things. Those with winning offers receive the following payments: a capability price (£/Mvar/h), and/or a synchronised capability price (£/Mvar/h), and/or a utilisation price (£/Mvar/h)[footnoteRef:37]. In contrast with ORPS which guarantees a default payment set through a formula, this mechanism has not been successful in the last years. No generator has provided reactive power under a Market contract since 2009. The percentage of total Mvar lagging capability with Market Agreements has been reduced from 70% (highest peak in Oct. 2000) to 0% in Oct. 2009 [76]. According to National Grid, one of the reasons is that ERPS competes with ORPS. The other could be the cap applied to the total funding for reactive power provision [77]. [37: See: https://www.nationalgrid.com/uk/electricity/balancing-services/reactive-power-services/enhanced-reactive-power-service] National Grid has released the Reactive Power Roadmap [42], and proposes the rationalisation (Stage 1) and simplification (Stage 2) of the current balancing services, including reactive power. TCM provides an ad-hoc solution. A transmission constraint can arise for different reasons (related to voltage or thermal constraints). There are two mechanisms for contracting voltage support, via bilateral agreements (between NGESO and individual providers) or via tenders. However, constraint management tenders are a way to procure it if there is enough competition[footnoteRef:38]. [38: See: https://www.nationalgrid.com/uk/electricity/balancing-services/system-security-services/transmission-constraint-management] 3.3 United States System operators from the USA procure different kinds of ancillary services, however those that are subject to competitive mechanisms are those grouped under the regulation and reserve categories, see Table A2 for further details. Many of these services are co-optimised with energy (i.e. when bids/offers are submitted for energy and ancillary services to the sport market, and are cleared using optimisation models to maximise social welfare [78] ). Reactive power is usually managed using the mandatory approach, which means that generators are required to provide reactive power support within their mandatory PF range specified in the interconnection requirement. PF requirements vary across ISOs and depend also on the type of generating facility. In CAISO and PJM synchronous large (>20 MW) and small (<=20 MW) generators are required to adjust their PF within the range of 0.90 lead and 0.95 lag [32,79]. In NYISO and ISO-NE the same PF requirement (0.95 lag/lead) is required for large/small synchronous [80,81]. Across all the ISOs, the same PF requirement applies for non-synchronous generating facilities. Based on FERC Order 827, new non-synchronous generation are required to provide dynamic reactive power within the range of 0.95 lead/lag at the high-side at the generator substation[footnoteRef:39]. [39: See: https://www.ferc.gov/whats-new/comm-meet/2016/061616/E-1.pdf] In terms of payments, a multi-part payment (composed of capability, opportunity costs and other payments) applies for reactive power service provision, in agreement with the Schedule 2 of the ISOs’ tariffs. However, generators are not always compensated in the same way. CAISO, the California System Operator, is the only one that does not compensate for the installation of reactive power capability (even for non-synchronous generators) because capability payments are not applied[footnoteRef:40]. This is supported by the fact that in California there are no centralised capacity markets but bilateral contracts for capacity (i.e. Resource Adequacy). This allows generators to reflect in their costs associated with energy, capacity and ancillary services. Then, providing capability payments would result in double payment for reactive power and hence double charging for reactive power to load serving entities [82]. The rest of ISOs compensate generators using different methods, such as a fixed rate (NYISO, ISO-NE) and the American Electric Power - AEP methodology (supported by FERC) used by PJM. AEP is a fixed-cost recovery approach and generators are required to submit the appropriate fillings to FERC for the evaluation of the cost-based revenue requirement for supplying reactive power. Based on this, NYISO and ISO-NE have set the compensation payments at US$ 2,747.61/Mvar/year [83] and US$ 1,188.33/Mvar/year [84] respectively for 2018. Figure 5 depicts the trend in reactive power costs in PJM. There is a slight decrease in reactive power costs ($/MWh) in the period 2012-2018. Reactive power unit costs represent an important share in total ancillary services[footnoteRef:41] unit costs and also in terms of total costs, representing in 2018 52% of ancillary services costs. [40: The other ISO that does not compensate for reactive capability is SPP, however this is not part of this study. ] [41: Composed of: reactive power, black start, regulation, synchronised reserves, non-synchronised reserves and day ahead scheduling reserve. ] Figure 5: Trend of ancillary services and reactive power costs in PJM There are many criticisms of the way reactive power capability payments are currently made in PJM market. According to [86], separated reactive power capability payments (OATT Schedule 2) should be eliminated with reactive power costs recovered in the capacity market instead. However, if this is not possible reactive power capability should be measured based on tested Mvar rating rather than nameplate Mvar rating. The latter are usually higher. Degradation is one of the issues but also the difference between theoretical versus actual Mvar ratings (i.e. difference between pre installation nameplate ratings and post installation tested capability). Opportunity cost is the other type of payment which is only applicable when generating resources operate outside their mandatory range. The four ISOs make this kind of compensation. NYISO defines this type of costs as: “The Lost Opportunity Cost payment shall be calculated as the maximum of zero or the difference between: (i) the product of: (a) the appropriate MW of output reduction and (b) the Real-Time Locational Base Marginal Price (LBMP) at the Generator bus; and (ii) the Generator’s Energy Bid for the reduced output of the Generator multiplied by the time duration of reduction in hours or fractions thereof” [80] p.745. The two other payments are related to the cost of energy consumed or produced by generating units and non-generator reactive resources for providing reactive power services. Table 5 shows the different procurement methods and type of remuneration that reactive power providers can get for offering their services in selected system operators’ jurisdictions. Table 5: Reactive Power Procuring and Payment Methods: A Comparison 4. The Power Potential Initiative: A New Market for Reactive Power The Power Potential project is a customer funded initiative that proposes the creation of a reactive power market using DER and additional capacity in the South East Region of the UK. The transmission network has reached its capacity in this area (limited by dynamic voltage stability and thermal capacity). The Power Potential project will alleviate the problem by procuring resources reactive power and active power services from different kinds of DER, using a market-based mechanism. Savings to energy customers has been estimated over £412m by 2050 (based on its potential implementation at 59 sites across GB) and up to 3.7 GW of additional connected capacity in this region [88]. 4.1 What is auctioned? The Power Potential project is soliciting offers from DER to provide reactive power services (dynamic voltage support) and active power support (for constraint management and system balancing) in the South-East of England. National Grid has identified four Grid Supply Points (GSPs) and their respective served areas where both reactive power and active power services are required. There is no limitation in the size of the resource but the qualifying DER are those with 1 MW or over, for both portfolio resources (aggregation) and directly contracted resources [89]. DER are expected to be connected at 11 KV or above for most effectiveness. A capability to provide 0.95 PF lagging or leading (equivalent to 32% of the maximum export capacity) is required, specific control modes, among other technical characteristics [90]. 4.2 How DER can participate DER participants (including aggregators) are subject to specific pre-qualification and testing (Mandatory Technical Trials and Optional Technical Trials)[footnoteRef:42] before taking part of the trial (only for reactive power service). After being selected, DER will subject to two different stages (Wave 1 and Wave 2). The third stage (Wave 3) is when DER and traditional transmission connected generators compete from a BAU approach. Wave 1 and 2 are funded from Power Potential budget and Wave 3 from NGESO budget[footnoteRef:43]. Details about Wave 1 and Wave 2 are provided below, for further details see [90, 91]. Currently, DER commissioning is ongoing. Wave 1 is expected to start in April 2020 and Wave 2 in July. [42: Mandatory Technical Trials are required for both, reactive power services and active power services. Optional Technical Trials are only for reactive power services and take place after the Mandatory Technical Trials. ] [43: The implementation of Wave 3 is subject to the project’s Steering Committee decision.] Wave 1 aims to demonstrate the technical solution. In this wave, DER participate on a non-competitive basis. Wave 1 involves simulating and measuring DER speed of response to voltage change and measuring effectiveness of DER delivery at each GSP. In order to encourage the participation of DER in the trial, they will receive a fixed fee for fixed number of hours for their participation in this first stage regardless of the size of DER. For the reactive power service, DER that participate in the Optional Technical Trials will be entitled up to £45,000 per site if this is available for 1,850 hours and over of the Wave 1 window during the trial year (with a minimum of 700 hours for receiving payment)[footnoteRef:44]. For the active power service, DER will receive an administered utilisation price set to £150/MWh. These payments would help to reduce the net investment that DER may be required to do for the acquisition of control and communication equipment, in order to participate in the trial. Some equipment will be provided by the project, but others need to be acquired by DER (i.e. communication and control capital costs per DER). [44: The maximum number of hours that DER can be entitled in Wave 1 is 2,520 hours that correspond to 15 weeks with 24-hour availability windows. ] Wave 2 aims to evaluate the financial viability of Power Potential. In this wave DER that provide reactive power services compete with each other and the market is run for at least 1,800 hours during the trial period which will be achieved by running auctions across a period made up several weeks. DER (with winning offers) will receive two kinds of payments: a secured availability payment (£/Mvar/h) and a potential utilisation payment (£/Mvarh). There is no cap set for either payments, however bids will be assessed against the alternative cost of reinforcing the network (i.e. the counterfactual being explored in wave 2 is whether DER can provide reactive support more efficiently than the installation of reactive equipment on the transmission network). It has been decided that DER will compete in day-ahead auctions, though they can choose not to alter their offers daily. For active power services, availability payments are not applicable and offers will be evaluated considering their costs in comparison with other options available to NGESO. Offers from DER can be received from different types of resources that are able to provide reactive and active power services in the area covered by at least one of the GSP specified by National Grid. Heatmaps are available to inform participants of the location of the GSP that would be more suitable for them. Offers can be made only for one GSP at the same time. Simultaneous offers to different GSPs are not allowed. The submission of offers, will be via a UK Power Networks platform (web portal). A participation of at least 5 DER with at least 40 Mvar of reactive power availability across the 4 GSPs is expected to ensure sufficient learning. Bids for reactive and active power services are expected to be submitted in service windows of 4 hourly EFA (Electricity Forward Agreement) blocks (in line with those set for other ancillary services), which means a total of 6 period per day. 4.3 About the evaluation criteria and eligibility Reactive and active power services will be procured through a market mechanism in day-ahead auctions and real time respectively using the pay-as-bid methodology. The idea is to select the offers based on a combination of both lowest costs and highest effectiveness with respect to GSP but limited to the current budget (around £0.6m for both Wave 1 and Wave 2). NG will forecast the reactive and active power services to be procured for each GSP and will instruct the Distributed Energy Resources Management System (DERMS)[footnoteRef:45] about this. The DERMS will evaluate the resources available (free capacity) at the lowest cost based as NGESO’s instructions. DER will then be instructed by the DERMS about the services to be provided at set points. Non-cost variables have not been considered in the evaluation criteria. [45: The DERMS run by UK Power Networks, is the Power Potential platform that facilitates the communication between National Grid and DER connected to UK Power Networks. It was developed by ZIV Automation. ] 5. Discussion and Lessons Learned 5.1 Procurement of reactive power and the need for market-based mechanisms The use of market-based mechanisms in the procurement of reactive power is practically non-existent globally. This is in contrast with other ancillary services such as frequency regulation and capacity reserves that are usually co-optimised with energy in day-ahead and/or real time market. Provision of reactive power by third parties (mainly generators) is generally managed under a mandatory approach (within a specific range of PFs that can vary depending on the type of generator). Generators are usually compensated using a fixed methodology (e.g. a flat compensation rate or a cost-based methodology) without any market determined prices. Other compensation schemes such the payment for opportunity cost, which include local marginal prices, happen under “very infrequent circumstances” [92] p.5. This means that in CAISO (where capability payments do not exist when generators operate within the PF range), generators are basically not compensated at all. There is also a risk of over-compensation or under-compensation when a fixed methodology is applied. For instance, even though a competitive market exists in GB, transmission connected generators prefer to get the default payment under ORPS rather than participate in the ERPS market. In other markets where a flat rate applies (NYISO and ISO-NE) and all generators receive the same compensation rate regardless of the technology and associated costs, there are also concerns that those markets undercompensate certain generators for the provision of reactive power [92], especially in comparison with the AEP methodology[footnoteRef:46]. [46: As an example of the inequality of these two methodologies: A wind generator that provides 200Mvar (lagging and leading) of reactive power capability from a 200 MW plant (with 0.9 PF) would be compensated for this capability differently. It would receive per year US$0 (CAISO), US$225k (ISO-NE), US$525k (NYISO), and US$ 1.9m (PJM). In the first one reactive capability is not compensated at all [95]. ] The AEP methodology (a FERC approved cost-based revenue methodology) based their estimations on the generators’ investments for reactive capability and the nameplate capacity is used for measuring this. A key observation on this methodology is the potential degradation of the nameplate output overtime, which translates to lower output in reactive power support [86]. According to [92], as the cost of equipment (which sets the cost-based rates) does not vary regardless of the reduction of the nameplate output due to degradation, then it is not appropriate to use the degradation in production to vary the reactive compensation received by suppliers. A different point is made by FERC by suggesting asking for testing in order to verify any potential degradation in the reactive power output of the generators. FERC has found that this happens in fact [93]. The other observation is that AEP was initially developed for synchronous generators however its application has been expanded for all type of generators, including non-synchronous (i.e. wind turbines). Investment for reactive power for wind generators can be higher in non-synchronous generators than in synchronous generators with the same nameplate capacity, especially in terms of more turbines and generator/exciters [94]. However, the installation of this additional equipment brings a higher level of reliability to provide reactive supply [95]. This suggests that the AEP methodology, if remains, should be able to allow the recovery of all the costs related to the provision of reactive power regardless of the type of resource. The risk of over-compensation or under-compensation for reactive power capability using a fixed methodology can be mitigated by introducing more market-based solutions. The proliferation of more DER can help to deal with the poor locational effectiveness that is observed when the resource is placed far from the point where reactive power services are required (Vars do not travel well). DER reactive power capabilities will also improve, in line with the upgrade of Network Codes and standards (i.e. IEEE 1547). Then, DER may constitute an important source of reactive power support for the power grid. Procuring reactive power from DER will also require greater interaction between DER, electricity distribution utilities and TSOs. The use of a market-based approach using DER for reactive power services represents one more channel to procure this type of ancillary service. The use of tenders as a unique mechanism for procuring reactive power from transmission connected generators (like in Belgium with Elia), it is not exactly the best scenario either especially due to the potential lack of competition which is accentuated by the local nature of reactive power. In Belgium, the centralised mechanism for procuring reactive power via tenders has been overturned in the last couple of years and replaced by fixed prices (imposed via a royal decree) due to the unreasonable nature of the tender prices, as judged by the regulator [46]. The use of fixed methodologies for compensating reactive power capability will continue in organised wholesale markets. However, it is expected that an enhanced methodology reflects the costs of providing reactive power services taking into consideration the nature of generators and appropriate measurements for estimating Mvar ratings and then reactive power capability. 5.2 A Market-Based Approach for Reactive Power: The Conceptual Auction Design Figure 6 illustrates the architecture of the Power Potential trial including main parties, data flow, connections and a summary of considerations for auction design applicable to the procurement of reactive power, further details are provided in the paragraphs below. The centrality in auction design of encouraging new entry and more participants in the auction. According to [96] a good auction should aim to attract entry, prevent collusion and predatory behaviour. New initiatives in auction design for reactive power procurement encourage new entrants (i.e. DER) aggregated or individual as we can see from Figure 6. This leads to more market participants in the supply of reactive power services (DER plus transmission connected resources). However, in the future the participation of new entrants should depend on whether it can compete (in terms of prices) with the traditional transmission connected resources or other future options. It is also important for repeated procurement auctions to be designed in a way that incumbent suppliers do not start by offering predatory (below cost) prices in the early auctions in order to deter entrants and / or then begin coordinating their offers with each other in order to raise outturn prices in the longer run. The importance of enhancing competition between the reactive power suppliers (i.e. DER) across the different supply sites (i.e. GSPs) via a package auction design. For instance, from Figure 6, it would be possible for DER1 to supply both GSP 1 and GSP 2, subject to technical capabilities. [97] suggest that auctions should aim to enhance substitution if multiple objects are for sale, encourage price discovery and induce truthful bidding. A joint auction allows a higher combination of products enhancing competition via substitution between reactive power suppliers. This would be a more complex auction design than for example the Power Potential initiative, however total procurement cost could be lower by selecting the combinations that maximise social welfare (discussed further below). Figure 6: Architecture of Power Potential trial Consideration of pay-as-clear price determination format and the incorporation of a quality dimension. In general, auction theory suggests that second price auctions are better than pay-as-bid as a way of determining prices [98]. Pay-as-bid is an approach that is well-known by system operators (NGESO, CAISO) and market participants. While pay-as-bid can promote generally lower prices in the short run [99], its lack of transparency on true costs may reduce dynamic efficiency relative to pay-as-clear (used by most Nordic countries in the procurement of FCR). It can also bias the equilibrium price and risk inefficiency. In economic theory, a second-price auction would work better for true price discovery with higher dynamic efficiency in comparison with pay-as-bid. In addition, the objective of a procurement auction is not only to minimise the price paid but to maximise economic welfare. The consideration of quality dimensions in the procurement process (represented by the locational effectiveness of reactive power) should be a part of good auction design. Consideration should be given to the penalty scheme and the pricing format for reactive power supply (availability + utilisation). It could be the case that non-delivery penalty affects the availability payment only. However due to the new requirements for DER (Grid Code GC0100), reactive capability is going to be compulsory. This implies only a utilisation payment is necessary. This is something that would need to be taken into account when contemplating the large-scale of market-based trials such as the one discussed in this study. On the other hand, penalties could be mitigated by reducing the risks of having DER with poor delivery. The use of non-cost variables in the evaluation of DER can help with this. The scoring matrix from DRAM sets a good reference for the identification of these variables. The frequency and periodicity of the auction and the cost benefit of nearer to real time procurement and co-optimisation. More frequent auctions allow both parties (suppliers and the system operator) to adjust the reactive power offers and demand in nearly real time. According to [16], reactive power can be implemented under different time frames but getting closer to real time consumption might create local market power issues which would need to be managed. Shorter trading periods can help to reduce ancillary services costs by allowing similar trading periods for each ancillary service (reactive power in this case) and the energy market. This is the case of other ancillary services such as frequency regulation and reserves which are procured with energy (day ahead and real time) in specific jurisdictions. This practice is referred to as co-optimisation and may result in important system costing savings. The careful specification of the counterfactual against which the auction results are to be evaluated. Reactive power can be acquired through auctions but also via transmission or distribution reactive equipment or through other future options (identifying and despatching of a specific DER using a similar approach to the current mandatory mechanism, offering a fixed price to the DER for reactive power). Running a reactive power auction mechanism for a small number of supply sites (i.e. GSPs) could be costly. The design of the contract between the DSO and TSO to incentivise optimal risk sharing. With the implementation of market-based initiatives for reactive power procurement using DER, the DSO assumes a new role that may expose it to a significant energy price risk (unlike now). Proper contractual agreements are required to incentivise DSOs to optimise their provision of reactive power and other ancillary services [100]. About the size of the reactive power market revenues. From Section 2.2. we observe that the market for ancillary services is still very small compared with the wholesale market (energy), and that the size of this varies by jurisdiction. This is especially true if reactive power markets are local and associated with a single network node, rather than across the whole market. Small size has implications in terms of whether the market facilitation fixed costs can be covered. Expanding the pool of participants in the presence of small amounts of total revenue implies that the absolute revenue expected to accrue to an individual market participant is small. This raises the issue as to whether participant’s internal fixed costs (both in terms of on-site hardware and management time) are likely to be covered even in the medium to long-run. Convincing DER to participate in the market would seem to be a significant issue, as evidenced by the reported delay to the Power Potential project trail itself. For instance by June 2019 around 60% of the participant contracts had been signed [101]. Set against this falling market participation costs, due to the falling costs of digital trading platforms designed around small market participants – and able to bid automatically on their behalf - could act to encourage more participants [102]. 5.3 Observations on the Power Potential Initiative Power Potential is an opportunity to trial the technical solution (DERMS implementation, DER local effectiveness, DER response time to voltage change), the commercial solution (represented by the avoided cost of transmission reinforcement for reactive power compensation), new roles (DNO as a facilitator for the procurement of reactive power and its transition to a DSO, the TSO as a contractor of reactive power services from distribution companies) and new interactions (between DNOs and the TSO and the need to coordinate in order to capture whole system benefits). Power Potential could also help to identify any regulatory barrier that may limit the value of the competitively procured reactive power from DER and its large-scale implementation. A centralised market for reactive power is not currently an option due to the poor capability of reactive power to travel long distances. For this reason, the organisation of a large-scale reactive power market does not seem to be feasible [44]. Then, a regional or more decentralised market, rather than the centralised one, would be much more viable especially due to the upward trend of DER. In the context of Power Potential these regional markets can be represented by a group of grid supply points (in this case four GSPs). 6. Conclusions This review suggests that there is a lack of competitive mechanisms for the procurement of reactive power at both the transmission and the distribution level. Competition in the USA is null and in Europe is very limited, with some few exceptions. In Australia, competition is also limited because the primary responsibility is given to Transmission Network Service Operators instead. This is in contrast to the competitive procurement of other ancillary services such as frequency regulation and capacity reserves. Instead, reactive power tends to be paid for via administratively determined pricing methodologies (involving fixed rates or cost-recovery) for procuring reactive power. This means that reactive power suppliers are likely to be over or under compensated. The introduction of more market oriented mechanisms and resources (such as DER) for acquiring reactive and active power services by the system operator opens new opportunities and new ways to deal with voltage stability issues, and to mitigate the risk of over or under compensation. This represents one more channel to procure ancillary services contributing to the diversification of procurement methods. At the same time, this imposes new challenges such as the implementation of new types of agreements (apart from the traditional ones) between DER/system operator/electricity distribution firm and the use of new platforms to manage reactive power. Power Potential is a first of its kind in seeking to competitively procure reactive power from DER. It offers the opportunity to trial not only the DER performance in the provision of reactive and active power but also an innovative procurement mechanism design. This paper provides key recommendations for such a design drawing on general lessons from auction theory and practice. Our discussion of the principles of mechanism design would suggest that attention is given to the following. First, the frequency of the auction and its price determination mechanism offers significant scope for learning what sort of price resolution might be necessary/desirable or possible. Second, consideration of the use of pay-as-clear (rather than pay-as-bid) helps to reveal information about underlying costs and to experiment with a different (and arguably superior) payment rule. Third, more consideration of how to enhance substitutability of products within the trail area should be given, particularly by integrating the procurement across a regional market that can be represented by a group of reactive power multiple grid supply points (4 in the case of Power Potential). Acknowledgments The authors would like to acknowledge the support from AEMO, CAISO, ISO-NE, NYISO, PJM, SCE, UK Power Networks and National Grid in the provision of relevant information and clarifications. The authors also acknowledge the financial support of National Grid via the Electricity Network Innovation Competition (NIC) Power Potential Project. The views expressed herein are those of the authors and do not reflect the views of the EPRG or any other organisation that is also involved in the Power Potential project. List of Acronyms AEMO Australian Energy Market Operator AEP American Electric Power aFRR automatic frequency restoration reserve CUSC connection and use of system code DAM day ahead market DER distributed energy resources DERMS distributed energy resources management system DSO distribution system operator ERPS enhanced reactive power service FCAS frequency response ancillary services FCR frequency containment reserve ISO independent system operator GSP grid supply point LBMP locational base marginal price ISO-NE ISO New England mFRR manual frequency restoration reserve NGESO National Grid Electricity System Operator NSCAS network support and control ancillary services NYISO New York ISO ORPS obligatory reactive power service PGM power generating modules PPM power park modules RfG requirements for generators RTM real time market STOR short term operating reserve TCM transmission constraint management TNSP transmission network service providers TSO transmission system operator VCAS voltage control ancillary services References [1] Dutra, J., Barbalho, A. The convergence of business models and long-term financing in the energy transition. Competition and Regulation in Network Industries 2018;18(3-4): 256-270. [2] Creti, A. and Fontini, F. Economics of Electricity: Markets, Competition and Rules, Cambridge: Cambridge University Press, 2019. [3] Kristiansen, T. The Nordic approach to market-based provision of ancillary services. Energy Policy 2007; 35: 3681-3700 [4] Kundur, P. Power System Stability and Control. The EPRI Power System Engineering Series, McGraw-Hill. [5] NG. Monthly balancing services summaries (MBSS) August 2019. National Grid Electricity System Operators. [6] Statnett-Fingrid-Energinet-Svenska Kraftnat (2018), The way forward – Solutions for a changing Nordic power system. Mar. 2018. [7] DENA. Ancillary Services Dena Innovation Report: Actual achievements and actions needed at present for stable operation of the electric power system up to 2030. German Energy Agency, Dec. 2018. [8] Dierkes, S., Bennewitz, F., Maercks, M, Verheggen, L., Moser, A. Impact of distributed reactive power control of renewable energy sources in smart grids on voltage stability of the power system. 2014 Electric Power Quality and Supply Reliability Conference (PQ), IEEE, pp. 119–126. [9] Simshauser, P. The Hidden Costs of Wind Generation in a Thermal Power System: What Cost?, The Australian Economic Review 44(3): 269-292. [10] NYISO. Distributed Energy Resources Market Design Concept Proposal. A Report by the New York Independent System Operator, Dec. 2017. [11] Pathak, A.K., Sharma, M.P., Bundele, M. A critical review of voltage and reactive power management of wind farms. Renewable and Sustainable Energy Reviews 2015;51: 460-471. [12] Saquib, M.A., Saleem, A.Z. Power-quality issues and the need for reactive-power compensation in the grid integration of wind power. Renewable and Sustainable Energy Reviews 2015;43: 51-64. [13] Gandhi, O., Rodriguez-Gallegos, C.D., Zhang, W., Srinivasan, D. Economic and technical analysis of reactive power provision from distributed energy resources in microgrids. Applied Energy 2018;210: 827-841. [14] Bai, L, Jiang, T., Li, F., Chen, H., Li, X. Distributed energy storage planning in soft open point based active distribution networks incorporating network reconfiguration and DG reactive power capability. Applied Energy 2018;210:1082-1091. [15] Ahmadi. H., Foroud, A.A. Improvement of the simultaneous active and reactive power markets pricing and structure. IET Gener. Transm. Distrib. 2016;10(1): 81-92. [16] Ahmadimanesh, A., Kalantar, M. Modification of the mandatory generation region of producers in the reactive power market by considering reactive power losses. Energy & Environment 2017;28(7): 744-762. [17] El-Samahy, I., Bhattacharya, K., Canizares, C., Anjos, M.F., Pan, J. A Procurement Market Model for Reactive Power Services Considering System Security. IEEE Trans. Power Syst. 2008;23(1):137-149. [18] Frias, P., Gomez, O., Soler, D. A Reactive Power Capacity Market Using Annual Auctions. IEEE Trans. Power Syst. 2008;23(3): 1458-1468. [19] Molina-Garcia, A., Mastromauro, R.A., Garcia-Sanchez, T., Pugliese, S., Liserre, M., Stasi, S. Reactive Power Flow Control for PV Inverters Voltage Support in LV Distribution Networks. IEEE Trans. Smart Grid 2017; 8(1): 447-456. [20] Gandhi, O., Zhang, C.D., Rodriguez-Gallegos, Bieri, M., Reindl, Sriinivasan, D. Analytical Approach to Reactive Power Dispatch and Energy Arbitrage in Distribution Systems with DERs. IEEE Trans. Power Syst. 2018 pp. 1-1. [21] Zhang, W., Gandhi, O., Quan, H., Rodriguez-Gallegos, C., Srinivasan, D. A multi-agent based integrated volt-var optimization engine for fast vehicle-to-grid reactive power dispatch and electric vehicle coordination. Applied Energy 2018;229(1):96-110. [22] Zubo, R., Mokryani, G., Abd-Alhameed, R. Optimal operation of distribution networks with high penetration of wind and solar power within a joint active and reactive distribution market environment. Applied Energy 2018;22:713-722. [23] Bayat, A., Baghari, A. Optimal active and reactive power allocation in distribution networks using a novel heuristic approach. Applied Energy 2019;(233-234):71-85. [24] Rebours, Y.G., Kirshen, D.S., Trotignon, M., Rossignol, S. A Survey of Frequency and Voltage Control Ancillary Services – Part II: Economic Features. IEEE Trans. Power Syst. 2007;22(1):358-366 [25] Zhou, Z., Levin, T., Conzelmann, G. Survey of U.S. Ancillary Services Markets. ANL/ESD-16/1. Energy Systems Division, Argonne National Laboratory, 2016. [26] Pollitt, M, Anaya, K. Competition in Markets for Ancillary Services? The implications of rising distributed generation, EPRG WP 1928, 2019. [27] Svenska Kraftnat. System Development Plan 2018-2027. Towards a flexible power system in a changing world. Nov. 2017. [28] CAISO. 2017 Annual Report of Markets Issues and Performance. California Independent System Operator Corporation, June 2018. [29] Enel. Reglamento de Servicios Complementarios. June 2017. [30] Fraunhofer ISE. Recent Facts about Photovoltaics in Germany. May 2019. [31] GE Energy. PJM Renewable Integration Study. Task Report: Review of industry practice and experience in the integration of wind and solar generation. Prepared for PJM Interconnection, LLC. Nov. 2012. [32] PJM, PJM Open Access Transmission Tariff, PJM. Effective date 23/04/2018-Docket#ER18-893-000. [33] EPRI, Wholesale Electricity Market Design Initiatives in the United States: Survey and Research Needs. 2016 Technical Update. Electric Power Research Institute, Nov. 2016. [34] AER. Report into market ancillary service prices above $500/MW. South Australia, Australian Energy Regulator, Jan. 2018. [35] Read, E. Grant. Co-Optimization of Energy and Ancillary Service Markets. In: Pardalos P., Rebennack, S., Pereira M., Iliadis N. (eds) Handbook of Power Systems I, Energy Systems. Springer, Berlin, Heidelberg, 2010. [36] EC. Commission Regulation (EU) 2017/2195 of 23 November 2017 establishing a guideline on electricity balancing. European Commission, Nov. 2017. [37] Anaya, K., Pollitt, M. Regulating the Electricity System Operator: Lessons for Great Britain from around the world. EPRG WP 1718 University of Cambridge, Dec. 2017. [38] ISO-NE. Reserve Products & Market Overview. Introduction to Wholesale Electricity Markets (WEM 101), March 2018. [39] ISO-NE. Regulation Market. Manual M-REG. Revision 2. ISO New England, Oct. 2018. [40] NG. System Needs and Product Strategy, Warwick: National Grid, Jun. 2017. [41] NG. Product Roadmap for Frequency Response and Reserve. National Grid, Dec. 2017. [42] NG. Product Roadmap for Reactive Power. National Grid, May 2018. [43] NG. Demand Turn Up service review. National Grid ESO, Feb. 2019. [44] Potomac Economics. 2017 State of the Market Report for the ERCOT Electricity Markets. Independent Market Monitor for ERCOT, May 2018. [45] PJM. PJM Manual 11: Energy & Ancillary Services Market Operations. Revision: 104, Feb. 2019. [46] Elia. Study on the future design of the ancillary service of voltage and reactive power control. Elia, Oct. 2018. [47] ENTSO-E. Survey on Ancillary Services Procurement, Balancing Market Design 2018. March 2019. [48] Pöyry. Nordic Market Design Forum – Feasibility study. Final Report. Pöyry, Sep. 2017. [49] Energinet. Security of Electricity Supply Report 2019, Doc. 18/03904-143. [50] CREG. Study on the functioning and price evolution of the Belgian wholesale electricity market-monitoring report 2017. Commission de Regulation de l’Electricite et du Gaz, June 2018. [51] Statenett (2015), Annual Report 2014. [52] Statenett (2017), Annual Report 2016. [53] CREG. Study on the functioning and price evolution of the Belgian wholesale electricity market-monitoring report 2018. Commission de Regulation de l’Electricite et du Gaz, Sep. 2019. [54] Statenett (2019), Annual Report 2018. [55] Energinet. Annual Report 2015, Mar. 2016. [56] Energinet. Annual Report 2016, Mar. 2017 [57] EC. Commission Regulation (EU) 2016/631 of 14 April 2016 establishing a network code on requirements for grid connection of generators. European Commission, Apr. 2016. [58] OFGEM. Requirement for generators – ‘emerging technology’ decision document. Final decision. Office of Gas and Electricity Markets, May 2017. [59] SONI-NIE. SONI and NIE Networks’ proposal for the general application of technical requirements in accordance with Articles 13-28 of the Commission Regulation (EU) 2016/631 establishing a network code on requirements for grid connection of generators. SONI and Northern Ireland Electricity Networks, May 2018. [60] NG. GC0100 EU Connection Codes GB Implementation – Mod 1. Stage 06: Final Code Modification Report. Grid Code. National Grid, Feb. 2018. [61] IEEE. IEEE 1547: Standard for Interconnection and Interoperability of Distributed Energy Resources with Associated Electric Power System Interfaces. Vermont Chapter, Feb. 2017. [62] NERC. Distribution Energy Resources. Connection Modelling and Reliability Considerations. North American Electric Reliability Corporation, NERC, Feb. 2017. [63] SmarteEn. The smarten Map. European Balancing Markets Edition. Smart Energy Europe. [64] More than Smart. Coordination of Transmission and Distribution Operations in a High Distributed Energy Resource Electric Grid. Prepared by staff from CAISO, PG&E, SCE, SDG&E with support from More Than Smart, Jun. 2017. [65] ENA. Open Networks Future Worlds. Developing change options to facilitate energy decarbonisation, digitisation and decentralization, 2018. [66] AEMO – ENA. Open Energy Networks. Consultation on how best to transition to a two-way grid that allows better integration of DER for the benefit of all customers. Australian Energy Market Operator and Energy Networks Australia, June 2018. [67] AEMO. Visibility of Distributed Energy Resources. Future Power System Security Program. Australian Energy Market Operator, Jan. 2017. [68] Exelon Companies. Comment and Policy Statement of the Exelon Companies. In the matter of the Value of Distributed Energy Resources. Case No. 15-E-0751. Exelon Companies, Apr. 2016. [69] CPUC. Energy Division’s Evaluation of Demand Response Auction Mechanism. Final Report (Public Version – Redacted). California Public Utilities Commission, Jan. 2019. [70] AEMC. National Electricity Rules. Version 106. Australian Energy Market Commission, 2018. [71] AEMC. Rule Determination. National Electricity Amendment (Network Support and Control Ancillary Services) Rule 2011. Australian Energy Market Commission, Apr. 2011. [72] Transgrid. Transmission Annual Planning Report 2017. TransGrid, Jun. 2017. [73] AEMO. Request for Expression of Interest Network Support and Control Ancillary Services. Australian Energy Market Operator, July 2012. [74] AEMO. Network Support and Control Ancillary Services Agreement. Generic Proforma. Australian Energy Market Operator, Dec. 2017. [75] AEMO. TransGrid Transmission Determination 2018-23 Network Support and Control Ancillary Services. Australian Energy Market Operator, Jan. 2018. [76] NG. Reactive Market Report. Thirty Eight Tender Round for Obligatory and Enhanced Reactive Power Services. National Grid, July 2017 [77] Energy UK. Ancillary services report 2017. Energy UK, Apr. 2017. [78] Yang, Y., Bao, M., Ding, Y., Song, Y., Lin, Z., Shao, Ch. Review of Information Disclosure in Different Electricity Markets. Energies 2018 ; 11(12), 3424. [79] CAISO. Fifth Replacement FERC Electric Tariff, California Independent System Operator Corporation, May 2017. [80] NYISO. NYISO Tariffs. New York Independent System Operator, Oct. 2017. [81] ISO-NE. Schedule 22 Large Generator Interconnection Procedures. Independ System Operator New England, May 2017. [82] PG&E. Stakeholder Comments. Reactive Power and Financial Compensation. Pacific Gas & Electric, Jun. 2015. [83] NYISO, 2018 NYISO Voltage Support Service Rates, New York Independent System Operator, 2018. [84] ISO-NE. Section II. Open Access Transmission Tariff (OATT). ISO New England, 2018. [85] Monitoring Analytics. State of the Market Report for PJM, annual reports (2012-2018). [86] Monitoring Analytics. State of the Market Report for PJM. January through September. Aug. 2018. [87] NYISO. NYISO Tariffs. Market Administration and Control Area Services Tariff (MST) 15 MST Rate Schedules - Payments for Supplying Voltage Su. New York Independent System Operator, Nov. 2014. [88] NG-UK Power Networks. Transmission and Distribution Interface 2.0 (TDI). Bid Document to OFGEM, Aug. 2016 [89] NG – UK Power Networks. The Power Potential Project. Head of Terms – Subject to contract. National Grid and UK Power Networks, Jan. 2018. [90] NG – UK Power Networks. Power Potential DER Technical Requirements v2.5.3. National Grid and UK Power Networks, Mar. 2019. [91] NG – UK Power Networks. Power Potential Market Procedures v5. National Grid and UK Power Networks, Mar. 2019 [92] EPSA. Comments of the Electric Power Supply Association. Reactive Supply Compensation in Markets Operated by Regional Transmission Organizations and Independent System Operators. Docket No. AD16-17-000, Jun. 2016. [93] FERC. Order accepting proposed rate schedule and establishing hearing and settlement judge procedures. Docket No. NJ18-9-000. Federal Energy Regulatory Commission, Feb. 2018. [94] FERC. Order accepting and suspending proposed rate schedule and establishing hearing and settlement judge procedures. Docket No. ER18-642-000. Federal Energy Regulatory Commission, Mar. 2018. [95] EDF Renewables. Introductory Remarks, Panel 3 addressing “Cost Incurred by Non-Synchronous Generators for Reactive Supply”. Reactive Supply Compensation in Markets Operated by Regional Transmission Organizations and Independent System Operators. Docket No. AD16-17-000, Jun. 2016. [96] Klemperer, P. What Really Matters in Auction Design, Journal of Economic Perspectives 2002;16(1): 169-189. [97] Ausubel, L. and Cramton, P. Auction design for wind rights. Report to Bureau of Ocean Energy Management, Regulation and Enforcement, 2011. [98] Krishna, V. Auction Theory, San Diego: Academic Press, 2009. [99] Holt, C.A. Competitive Bidding for Contracts under Alternative Auction Procedures, Journal of Political Economy, 88 (3): 433-445. [100] Kim, S-W, Pollitt, M, Jin, Y.G., Kim, J. and Yoon, Y.T. Contractual Framework for the Devolution of System Balancing Responsibility from Transmission System Operator to Distribution System Operators, EPRG WP 1715, University of Cambridge, 2017. [101] NG – UK Power Networks. Power Potential Regional Market Advisory Panel, Outcomes, 26th June 2019. [102] Küfeoğlu, S., Gaomin, L., Anaya, K., Pollitt, M. Digitalisation and New Business Models in Energy Sector. EPRG WP 1920, University of Cambridge, 2019. [103] NG. GC0101 EU Connection Codes GB Implementation – Mod 2. Stage 06: Final Code Modification Report. Grid Code. National Grid, Feb. 2018. [104] NG, GC0100, GC0101 and GC0102 Complete European Connection Conditions Legal Text. European Connection Conditions. Issue 5, Revision 21, 2018. Appendix Table A1: A Comparison of Current Ancillary Services in the USA and GB (selected services) Table A2 : Reactive Capability Specifications for Power Generating Modules in Great Britain (EU Connection Code requirements) 1 image1.emf BelgiumGBGermanyDenmarkFinlandNorwaySweden Frequency Containment Reserve - FCR Procurement schememarkethybridmarketmarketmarkethybridmarket Product maximum resolution (time)week(s)month(s)week(s)hour(s)hour(s)hour(s)hour(s) Settlement rulePABPABPABMPMPMPPAB Type of providerG+L+B+PSG+L+B+PSG+L+B+PSG+L+BG+L+BGG Minimum size (MW) 111 0.3 (FCR-N), 1 (FCR-D) 0.1 (FCR-N), 1 (FCR-D)11 MonitoringEPChybridRTEPChybridhybridhybrid Voltage control - Reactive power Procurement scheme tenders (>5Mvar) Mandatory (>50MW) & tendersMandatoryMandatoryMandatory Mandatory (> 1Mvar)Mandatory Paid by the TSO/ESOyesyespartlypartlynoyesno Settlement rulefreePABfreehybridN/ARPN/A DSOs as providersnonononononoyes MonitoringhybridRTEPChybridhybridRTRT G: Generator, L: Load, B: Batteries, PS: Pump storage, EPC (ex post check): monitoring carried out after 24h of the delivery period, PAB: paid as bid, MP:marginal pricing, RP: regulated price, RT:real time, FCR-N: FCR for normal operation, FCR-D: FCR for disturbances, N/A: inf. no provided by the TSO Source: [46-47], NGESO, ENERGINET and FINGRID websites. image2.emf Notes: (1) all figures adjusted to 2018 prices, (2) Energinet figures include both AS (frequency and non-frequency AS). Source: TSOs annual reports (2011-2018), NGESO monthly Balancing Services Summary (2011-2017), National Statistics Agencies, [49]. 05010015020025020112012201320142015201620172018Index (2011=100)DKNOFISEBEGB image3.emf Year DK FINO SE GBBE DK (*) GB (*) BE (*) 2011 na0.9%1.36%2.0%2.4%2.1%10.7%3.0%2.5% 2012 na1.3%1.03%1.5%2.5%2.6%11.7%3.1%3.1% 2013 7.4%2.0%1.27%1.8%2.5%3.0%9.5%3.1%3.4% 2014 7.7%2.0%1.45%1.8%2.4%3.2%10.4%3.1%3.7% 2015 7.4%1.8%1.11%1.3%2.2%2.1%9.7%2.9%2.5% 2016 8.1%1.6%0.93%1.6%2.1%1.9%9.1%3.0%2.2% 2017 5.7%1.6%0.79%1.6%2.1%1.9%7.2%3.1%2.2% 2018 7.0%1.7%0.99%nana3.4%9.2%na3.6% average7.2%1.6%1.1%1.7%2.3%2.5%9.7%3.0%2.9% Notes (1): estimations made using adjusted figures (2018 prices), (2) na: data no available or no provided in the format required; (3): figure refers mainly to frequency related AS, (4) *: includes frequency and non-frequency AS (e.g. reactive power, black start, others) in selected countries, (5): % Ren. refers to the share of gross electricity production from intermittent resources over total electricity production Source: TSOs annual reports (2011-2018), NGESO monthly Balancing Services Summary (2011-2017), Eurostat & Nord Pool database, national statistics agencies, Ofgem, [49], [53]. image4.emf Countries&TSOsType AType BType CType D GB (NGESO)0.8kW ≤P≤ 0.99MW1MW ≤P≤ 9.99MW10MW ≤P≤ 49.99MW50MW≤P Northern Ireland (SONI)0.8kW ≤P≤ 0.09MW0.1MW ≤P≤ 4.9MW5MW ≤P≤ 9.9MW10MW≤P Belgium (Elia)0.8kW ≤ P < 1MW1MW ≤ P < 25MW25MW ≤ P< 75MW75MW≤P Voltage conditionor 110kV>=Vcp Vcp: voltage at connection point, P: Max. capacity. Source: [46], [59,60]. and Vcp<110kV (Type A, B, C) image5.emf DNO ED1: refers to generators connected to the distribution network. TEC: refers to generators with transmission entry connection. Embedded: refers to generators connected to the distribution network with access rights to the transmission network. Figures from Nov. 2015. (TEC, Embedded), week 24 2015 (DNO). Source: [60] p. 176. 98%2%1MW-9.99MW (5.3 GW)DNO ED1TECEmbedded79%11%10%10MW-49.99MW (11.9 GW)DNO ED1TECEmbedded10%90%0%>50MW (76.1 GW)DNO ED1TECEmbedded100%0.8kW-0.99MW (2.9 GW)DNO ED1TECEmbeddedType AType BType CType D image6.emf Annual figures: Jan.-Dec., FC: Frequency Control, Reg: Regulation, Cont: Contingency, RP: Reactive Power. Source: AEMO AS Payments Summary - Annual Reports (2012-2018), Australian Taxation Offfice. 0%5%10%15%20%25%30%35%40% - 50 100 150 200 250 300 2012 2013 2014 2015 2016 20172018AUS$ million (2018 prices)FC (Reg.)FC (Cont.)RPRestart% RP over total image7.emf ModeGeneration ModeSynchronous Condensor Mode Price and payments Availability charge - RP generation  Availability charge - RP absorption  Enabling charge  Compensation payment  Testing charge  Source: [74] image8.emf Source: NG monthly Balancing Services Summary (Apr. 2013-Jan. 2019), NG reactive utilisation dataset (Jan. 2019), Office for National Statistics. - 500,000 1,000,000 1,500,000 2,000,000 2,500,000 3,000,000 3,500,000020406080100120140160Apr-13Jul-13Oct-13Jan-14Apr-14Jul-14Oct-14Jan-15Apr-15Jul-15Oct-15Jan-16Apr-16Jul-16Oct-16Jan-17Apr-17Jul-17Oct-17Jan-18Apr-18Jul-18Oct-18Jan-19Mvarh£ million (2019 prices)RP leadRP lagtotal balancing costsRP costsLinear (total balancing costs) image9.emf Source: [85], U.S. Bureau of Labor Statistics 01002003004005006007000.00.20.40.60.81.01.21.41.62012201320142015201620172018millions $ (2018 prices)$/MWh (2018 prices)ancillary services ($/MWh)Reactive ($/MWh)Reactive total costsLinear (Reactive ($/MWh)) image10.emf CountrySOPeriodicity Compulsory /MandatoryTendersCapabilityAvailabilityEnabling Utilisation Opportunity costsOthers USACAISO  variable NYISO  variable PJM  variable ISO-NE  variable AustraliaAEMO (GM)  variable AEMO (SCM)  variable GBNG (ORPS)  variable NG(ERPS)  every six months, with term contract minimum 1 year and then in six-month increments GM: generation mode, SCM: synchronous condensor mode. Others include: testing charges, cost of energy used to energise equipment that provides voltage support. Source: [74], [79], [80], [84], [86], [87], NGESO Reactive Power Service Guides. Procurement methodType of payment image11.emf Aggregator DER1 DER2 GSP 1 DER3 DER4 GSP 2 DER5 DER6 GSP 3 LEGEND DER7 data flow (request for service, dispatch orders) DER8 connection (existing/new/in progress interfaces) GSP 4 RTU: remote terminal unit GSP: grid supply point Users (DER Controller) transmission (NGESO)distribution (UKPN)Customer premises Source: [90], adapted NationalGridDERMS (distributed energy resources management system)RTURTUWEBAPIWANMarketdesign considerations -New participants(DER) for RP procurement supported by a well-designed auction mechanism -Possibility of joint auctions (DER exposes to more than 1 GSP)-Pay-as-clear over pay-as-bid for true price discovery-Consideration of price formation and penalties scheme to mitigate risk-Preference for more frequent/periodical auctions (the closer to real time the better)-Evaluation of alternative options (counterfactuals vs RP procurement from DER)-Optimal risk sharing (DSO in its new role)-RP market still small and new for DER, which may reduce participation, technological advances can help image12.emf GB Ancillary service markets and names CAISOISO-NEMISOPJM (1)SPPNYISOERCOT (2)NG (3) Regulation RTDA,RTRT DA,RT Regulation UpDA,RTDA,RTDA Regulation DownDA,RTDA,RTDA Regulation (performance) RTNA Regulation Up MileageDA,RTDA,RT Regulation Down MileageDA,RTDA,RT Regulation ServiceRT Regulation movementDA,RT Regulating MileageDA,RT Frequency response Mandatory frequency response Firm Frequency Response (dynamic) monthly tenders Firm Frequency Response (static) monthly tenders Spinning reserve DA,RTDA,RTDA,RTDA,RT Ten-minute spinning reserveRT, FR Synchronised reserve (within 10min.)RT Responsive reserve DA,RT Non-spinning reserve DA,RTDA,RTDA,RT Ten-minute non-spinning reserveRT, FR Non-synchronised/Quick start (within 10 min.)RT Thirty-minute operating reserve RT, FR Supplemental reserve (4)DA,RTRT (5)DA,RT Ramp reserves (6) RT, FMDA,RT Reserve BM startup Demand turn up (7) tender Fast reserve monthly tenders Optional Reserve Services Short term operating reserve (Committed) 3 tenders/y Short term operating reserve (Flexible) 3 tenders/y Short term operating reserve (Premium Flexible) 3 tenders/y Replacement reserve (to go live by Dec. 2019) bids Reactive power (voltage support) Mandatory reactive power service Enhanced reactive power service (8) biannual tenders Black start NA market-based mechanisms (tenders) other (cost-based, lost opportunity cost, revenue-based, mandatory) Markets: DA: Day Ahead, RT: Real Time, FR: Forward reserve(pre-DA), FM: Fifteen Minutes,: NA: No available (1): Regulation in PJM is provided by a combination of resources following 2 signals: RegA (slow response) and RegD (quick response). (2) In ERCOT, regulation up and down may also be procured in the supplementary ancillary services market (SASM) in case of replacement of AS capacity or limitations at a generating unit. (3) Simplified list of AS as of Apr. 2019. (4): Provided by online or off-line resources in MISO/PJM. (5): PJM uses a day-ahead scheduling reserve in addition to the RT for supplemental reserve (30min). (6): Ramp product: Up and Down Ramp Capability (MISO) in addition to the RT for supplemental reserve (30min); Flexible Ramping (CAISO). (7): Not to be procured in 2019 due to a decline in volume procured and number of participants over the last three years. (8): Not currently active for procurement. Source: [37-45]. USA Regulation/FrequencyResponseReservesOthers image13.emf ConceptType BType C/D RP capabilityBased on PF U-Q/Pmax requirement(rectangular shape) (normal operation)0.95 PF (lag), 0.95 PF (lead)V (p.u)max: 1.05, V (p.u)min: 0.95 Q/Pmax (lead): -0.426 (PF=0.92), Q/Pmax (lag): 0.426 (PF=0.92) RP capability (below max. cap.) In agreement with the generator reactive performance chartIn agreement with the generator reactive performance chart ConceptType BType C/D RP capabilityBased on PF U-Q/Pmax requirement (non-rectangular shape) (normal operation)0.95 PF (lag), 0.95 PF (lead)Connection point above 33kV: V(p.u) range = 0.225, V (p.u)max: 1.1, V (p.u) min:0.875 Q/Pmax (lead): -0.33 (PF=0.95), Q/Pmax (lag): 0.33 (PF=0.95) Connection point below 33 kV: V (p.u) max: 1.05, V (p.u) min: 0.95 Q/Pmax (lead); -0.33 (PF=0.95), Q/Pmax (lag): 0.33 (PF=0.95) RP capability P-Q/Pmax requirement (non-rectangular shape) (below max. cap.)P (p.u)max: 1, P (p.u)min: 0.2 Q/Pmax range=0.66 In agreement with the generator reactive performance chart RP capability Configuration 1: Q/Pmax range=0 (PF=1), Configuration 2: Q/Pmax range=0.33 (PF=0.986) Configuration 1&2: V(p.u) range = 0.225 RP capability (below max. cap.)Configuration 1: Q/Pmax range=0 (PF=1), Configuration 2: Q/Pmax range=0.33 (PF=0.986) Configuration 1&2: P (p.u)max: 1, P (p.u)min: 0.2 Notes: (1) For PPGM RP capability (below max. cap) there is no obligation to provide this capability for V(p.u)<0.2 in agreement with the current reactive power capabilities for non-synchronous generators. The same applies for OPPM. (2) For Type B, 0.95 should be adopted unless a different value is agreed with the GB System Operator or the DNO. Source: [103, 104]. U-Q/Pmax profile (non-rectangular shape) P-Q/Pmax profile (non-rectangular shape) In agreement with the generator reactive performance chart Synchronous Power Modules (SPM) Power Park Generation Modules (PPGM) Offshore Power Park Modules (OPPM)