Review of Main Projects, Characteristics and Challenges in Flexibility Markets for Services Addressed to Electricity Distribution Network ^†

Abstract

The expansion of distributed renewable resources, together with increased demand from the electrification of transport and heating sectors, impacts distribution networks significantly. Additionally, the emergence of non-programmable and intermittent generators is set to diminish the dominance of traditional rotating and programmable generation, thereby affecting the overall stability of the system. Nevertheless, the flexibility offered by distributed resources has the potential to alleviate the necessity for network reinforcement and contribute to system stability at competitive costs. Local flexibility procurement should be rooted in local markets, serving as mechanisms to address distribution congestion and coordinate the provision of flexibility for transmission network services. The multitude of existing systems and the interdependence of flexibility services have given rise to diverse solutions, still undergoing experimentation in various countries. This paper aims to scrutinize key projects that have established local flexibility markets, delineating their fundamental characteristics, the most common solutions, identifying prevalent barriers and suggesting potential future improvements. The investigation focuses on the most uncertain aspects of local markets: possible TSO-DSO coordination schemes, the time horizon for the acquisition of services and the baseline definition methodologies.

1. Introduction

The European Green Deal, together with the Fit-for-55 package, aims to make Europe the first carbon-neutral continent. Energy uses are estimated to be responsible for 77% of greenhouse gas emissions among all other uses [1]. Therefore, to reach the final emissions targets, a fast increase in renewable energy production and electrification of transport and other consumptions is ongoing. It can be observed that small renewable plants placed near consumption points help reach the final target, and a large growth of distributed generation is foreseen. New non-programmable and intermittent generators will also impact the stability of the whole system, reducing the share of traditional rotating and programmable generation. Indeed, the opening to the services market at the national and European level launched in 2012 [2] requires a market reform able to manage the large increase in volatility and uncertainty in the system’s stability. Two parallel strands overlap in this context: (1) Distribution System Operators (DSOs) must be revealed to be able to sustain the drastic changes that their grids are undergoing, and (2) DSOs must cooperate with Transmission System Operators (TSOs) to guarantee distribution generation availability and usability.

This article focuses on the first step, introducing the concept of flexibility of resources that can play a key role in overcoming difficulties and uncertainty caused by the increase in renewable resources and increase in consumption. Indeed, it is difficult to imagine that the current fit&forget approach, according to which every capacity problem is solved by reinforcing or building new lines, will be cost-efficient in a framework where production and consumption profiles are mostly volatile. Therefore, flexibility services at the local level are envisioned as a substitute for grid reinforcement; they are to be introduced both in long- and short-term operation procedures. Among the different procurement strategies aligned with recommendations of international/national regulators [1,3], the acquisition of local flexibility should be based on dedicated markets, which are designed to address local violations and, furthermore, to aggregate flexibility for transmission network services. A proficient local flexibility market should align with the current planning and operational procedures of the DSO, as delineated in [4]. This can be achieved via the adoption of a standardized approach that effectively aids the planning stage and enhances operations by facilitating the procurement of congestion management services. While regulatory guidelines [5] have outlined some fundamental aspects and major network associations [6] have proposed a technical strategy for acquiring flexibility services, the multitude of existing systems and the mutual dependency on flexibility services result in a wide variety of solutions that are yet to be fully determined.

On the basis of the highlighted necessities, many research and experimental projects, together with the scientific community, are currently investigating local market applications. All initiatives feature different characteristics, adapting the proposed implementation to the identified necessities. The aim of the paper consists of providing references to these projects and to extract and comment on the characteristics of the most relevant ones according to the authors’ opinion.

This paper extends the work presented in [7] by considering and describing additional projects and review papers. This extension reports a more detailed and comprehensive discussion of identified barriers and proposes possible solutions. The analysis of some of the less-defined aspects of local markets is added, such as the time horizon for the acquisition of services and the baseline definition methodologies. The examination considers the main projects that have been established and/or implemented regarding local flexibility markets, particularly focusing on European initiatives addressing local congestion management at the distribution level. In Section 2, a literature analysis is carried out, many projects and platforms for local service procurement are studied and, on their basis, local market integration is analyzed. Specifically, TSO-DSO coordination schemes investigated in previous projects are reviewed (Section 2.1), and methodologies to define market timeframe and baseline profiles are confronted and critically observed (respectively Section 2.2 and Section 2.3) without neglecting the main markets and bid characteristics (Section 2.4). In Section 3, given the acquired experience, possible alternatives to integrate local markets in the current regulatory framework are provided as a function of system needs and characteristics. In Section 4, barriers to the development of markets are analyzed and categorized according to three different typologies (technical, economic and regulatory), and, in the Conclusions Section, key aspects are considered for sustaining the future evolution of the power system.

2. Literature Analysis

The literature review analyzed more than seventy references, primarily concentrating on European projects featuring demonstration pilots on local markets (

Table 1. List of considered projects on the exploitation of local flexibility for distribution services (square brackets indicate references, [-] indicates missing references and Others actors include balancing responsible parties and balancing service providers).

Project	Ref.	Actors			Active	Review
		TSO	DSO	Others		[8]	[9]	[10]	[11]	[12]	[13]	[14]	[15]	[16]	[17]	[18]	[19]	[20]
ALPGRID	[21]		x	x					x
ATLAS	[22]			x					x
BeFlexible	[23]	x	x	x
COORDINET	[24]	x	x	x			x	x	x	x	x			x	x	x	x
Cornwall Local Energy Market	[25]	x	x							x
CROSSBOW	[26]	x														x	x
DA/RE	[27]	x	x		x	x		x								x
De-Flex-Market	[28]	x			x				x		x
DELTA	[29]	x	x	x
DOMINOES	[30]	x	x											x
DRES2Market	[31]	x	x							x
EcoGrid 2.0	[32]	x	x	x					x		x
EDGE	[33]		x	x	x
E-LAND	[34]	x	x
EMPOWER	[-]	x	x						x		x
Enera	[35]	x	x	x			x		x	x	x	x		x	x	x	x	x
ENKO Flexibility Platform	[36]	x	x		x											x
Equigy Crowd Balancing Platform	[37]	x			x	x										x
ESIOS	[38]	x			x											x
EUniversal	[39]	x	x	x												x
EU-SysFlex	[40]	x	x						x	x	x	x		x			x
EvolvDSO	[41]	x	x
FEVER	[42]	x	x	x										x
FLEXCoop	[43]											x
Flex-DLM	[44]		x						x		x
FLEXGRID	[45]	x	x
Enedis	[46]		x		x													x
flexiblepower	[47]		x		x
FLEXICIENCY	[48]		x						x		x		x
FLEXITRANSTORE	[49]			x												x	x
FlexMart	[-]		x						x		x
FutureFlow	[50]	x												x
GIFT	[51]		x
GOFLEX	[52]		x							x				x	x
GOPACS	[53]	x	x		x	x	x		x	x	x			x		x	x	x
IFLEX	[54]			x
InteGridy	[55]		x	x
INTERFLEX	[56]	x	x			x			x	x	x	x		x			x
InteGrid	[-]	x	x						x	x	x						x
Internet of Energy project (IO.E) in Belgium	[57]			x
INTERPLAN	[58]	x	x
INTERRFACE	[59]	x	x				x	x	x	x	x					x	x
INVADE	[60]	x	x	x
iPOWER	[61]		x	x
IREMEL	[-]	x	x						x	x	x
Magnitude	[62]			x								x
MERLON	[63]	x	x	x
MiNDFlex	[64]		x	x	x
MUSE GRIDS	[65]			x
NODES-INTRAFLEX	[66]	x	x	x	x	x	x	x	x	x	x			x	x	x	x
NORFLEX	[67]			x				x								x
OneNet	[68]	x	x	x
OSMOSE	[69]	x	x														x
PICLO FLEX	[70]		x		x	x	x		x	x	x			x	x	x	x	x
PLATONE	[71]	x	x	x					x							x	x
Power Potential	[72]	x	x
Project LEO	[73]	x	x
Redispatch 2.0	[74]	x	x		x
REFLEX	[75]			x
RomeFlex	[76]	x	x	x	x
SENSIBLE	[-]	x	x						x		x		x
SmarterEMC2	[77]	x	x
SmartNet	[78]	x	x			x							x		x		x
SMILE	[79]			x
Soteria	[80]			x
Store & Go	[81]			x
TDX-ASSIST	[82]	x	x														x
TenneT Cooperation Project Grid Stabilisation—Vehicle 2 Grid	[83]			x
UPGRID	[84]	x	x
USEF	[85]	x	x	x					x		x
WiseGRID	[86]	x	x

). Additionally, thirteen papers were identified that detail and compare a subset of projects [8,9,10,11,12,13,14,15,16,17,18,19,20]. However, in addition to considering a larger number of projects, the objective of this paper is not to undertake an exhaustive analysis and comparison of each project, which is deeply completed in [10,11,12], but rather to delineate the main characteristics of the adopted markets understanding that there is not a single best solution, but the best choice depends on the application context. From the literature review, several general traits of proposed local flexibility markets have been identified, also quantifying the number of projects that adopt a specific solution [13].

The main services procured via local flexibility markets include resolving current congestion and voltage violations on the distribution network managed by the DSO. Although a few projects consider additional services such as black-start and island operation, they are not considered in the present analysis as they are adopted by only a minority of projects.

Ref. [19] proposes the definition of harmonized products, considering the possibility of comparing relevant attributes and their respective numerical values. In this context, it is also observed how harmonized products could help in reducing effort in TSO-DSO coordination. Indeed, in certain instances, the local platform is also utilized by the TSO to secure flexibility for balancing the system and contain current and voltage violations on the transmission network. In this regard, the interaction between the procurement of local and global services is one of the key aspects shaping the operational dynamics of the new local market in alignment with pre-existing ancillary services markets. This interaction is significant as the outcomes of local markets can impact dispatchment in the global system and vice versa.

Each of the following paragraphs focuses on a single market aspect, each of them selected according to the authors’ opinions developed from the literature review. Different TSO-DSO coordination schemes are presented in Section 2.1, and implications for the choice of the scheme are then described in Section 2.2, where the importance of the timeframe for acquiring various services is discussed. For instance, if local flexibility is planned during the day-ahead market, local activation could be seamlessly integrated into the day-ahead bids of aggregators. Another crucial aspect explored in the literature is the definition of the baseline profiles of flexible resources (Section 2.3), which directly affects the quantification of activated flexibility. Finally, the main alternatives for market implementation are presented (Section 2.4), forming the groundwork for the classification described in Section 3.

2.1. TSO-DSO Coordination Schemes

In the review, seven distinct coordination schemes can be delineated, characterized by an increasing level of integration between local and global markets. These coordination schemes will be briefly outlined, with particular attention to the procurement priority of distributed energy resources (DER), following the terminology and the schemes of [13] and adding the new coordination scheme M0. In this coordination scheme (M0—Figure 1), local and global markets are entirely segregated, and the agents are responsible for proposing bids in separated markets. Participants are responsible for offering bids to one market without determining the imbalance in the other market. This coordination is recommended when there is a significant difference in the timeframe of services acquisition (e.g., local services are acquired during the day-ahead market, while global services are acquired near real-time) or when the expected activation of local power is small, thus exerting negligible impact on the overall system balance.

The second coordination scheme (M1) does not foresee the local market: the sole responsibility of the DSO consists of pre-qualifying the resources for the TSO market (Figure 2). This market structure is implemented in scenarios where the is no requirement to procure local services, but the utilization of local resources for global needs could determine issues in the distribution network.

In the third market scheme (M2), precedence is given to the local market over the global one (Figure 3). The DSO is responsible for selecting resources necessary to address local congestion, with any remaining flexibility offers being directed toward the TSO. Furthermore, the DSO may validate the offers chosen by the TSO to ensure their activation does not impact the secure operation of the distribution network. Under this scheme, the local market may also aggregate local resources to offer them to the TSO. Consequently, bids submitted to the TSO would inherently adhere to the constraints of the distribution network. This arrangement implies that the local market can coordinate resources from various aggregators, a circumstance that warrants careful consideration from a regulatory standpoint to prevent discrimination.

In the fourth scheme (M3), the two markets establish predefined profiles that the DSO must adhere to. In this scenario, the DSO can use the local flexibility to address local congestion as well (Figure 4).

In the fifth scheme (M2/3), only the DSO can use local resources. The overall imbalance resulting from resource activation is then communicated to the TSO (Figure 5), which manages any possible imbalances within its market.

In the sixth scheme (M4), a unified market exists for both DSO and TSO to acquire the needed services (Figure 6). In this case, both the DSO and TSO share equal priority in procuring local resources, and their selection process jointly takes into account constraints at both distribution and transmission network levels.

The last scheme (M5) represents the most inclusive market environment, as it allows not only the DSO and TSO to procure services but also permits responsible parties to acquire flexibility to balance their own resource portfolios (Figure 7). While in other schemes, the balancing responsible parties can also procure balancing resources (e.g., in the global market), in M5, they directly compete with the network operators. Consequently, the DSO must offer higher prices relative to other market participants in order to prioritize the use of resources.

The choice of coordination scheme relies on several factors. Typically, the necessity for coordination increases with the volume of local flexibility because of its impact on system balancing. Additionally, as local resources become more involved in supporting the transmission network operations and as the DSO requires resources to address local congestion, the need for coordination intensifies. During the initial experimental implementations, where the demand for local flexibility and the availability of resources were limited, the simplest coordination schemes were validated with success by testing their technical efficacy.

2.2. Service Acquisition Timeframe

Another important aspect of implementing local markets is establishing the timeframe for acquiring services. While not thoroughly investigated in all the analyzed projects, this factor significantly influences the selection of the coordination scheme. Typically, the timeframe is closely related to the ability of DSO and TSO to forecast violations, which depends on the variability of user exchange profiles and on the forecasting methodology employed. Usually, the DSO forecasts are the most critical since they consider smaller aggregation perimeters than TSO, which determines an intrinsic higher uncertainty due to the lower number of users.

The timeframe of the acquisition of local services contributes to defining how to integrate the acquisition of local services within the workflow of already existing markets. For instance, in scenarios where violations can be detected prior to the opening of the global market, they may be addressed independently without relying on the TSO market. This is the case of some currently active platforms (e.g., [47]), where the activation of resources for certain services is scheduled days in advance. However, in these cases, to consider possible errors in the forecast calculations, the DSO could acquire more resources than needed, increasing the cost of local flexibility. In the opposite case, when the congestion can be detected only close to real time, a strong integration with the global market is needed to compensate for unforeseen imbalances. For these reasons, the timeframe is defined by a trade-off between the over-acquisition of services from DSO and the imbalances in the TSO.

2.3. Baseline

The timeframe also influences the identification of the baseline profile of resources [12], which is essential for quantifying the delivered service. In the literature it is not found a general review of the baseline calculation methodologies and their impact on the local market design. For this reason, on the basis of the analyzed literature (Table 2), the main methods used to define the baseline are identified:

(a)

The DSO (or the local market operator) calculates the baseline on the basis of historical patterns, possibly considering meteorological forecast and real-time data [70,87,88,89,90];

(b)

The aggregator computes and provides the baseline profile to the DSO [32,45,91,92];

(c)

The baseline is established from the results of the previous (energy) market sessions [32,93];

(d)

The baseline profile is extrapolated from the measured power exchange before and after the service delivery [94].

The definition of the baseline is particularly relevant when the DSO and TSO share the same market: Since the remuneration of resources should consider the activation of the two markets, the baseline computation must be aligned. In this case, there is also the difficulty that the aggregation of perimeters for the global market can be different with respect to the aggregation perimeters for the local market. For example, a perimeter for global services could contain many perimeters for local services, and the baseline computation of the larger perimeter should consider the baseline calculation of all the smaller perimeters. A possible solution is to design the local and global markets in a way to that allows leaving the aggregators the responsibility of presenting coherent baselines. For example, if the aggregator knows that in the local market the baseline is computed by historical value, it can take this information into account when defining the baseline for participating in global markets. Instead, if the markets are separated, the baseline computation can be different.

The baseline calculation is particularly important for resources that cannot be programmed with high precision, like residential users or photovoltaic generation. The forecast and baseline calculation can be affected by a large error due to the high variability of such users. Given that the accuracy of power profiles impacts the uncertainty of flexibility margins and that non-programmable resources often have limited control margins, they reveal to be more suitable for providing services that do not require high precision, like in some local markets. For example, if the flexibility resources reduce the power more than the quantity required by the DSO, there are usually no issues for the operation of the network (the same does not apply in the case of balancing services). However, the uncertainties in determining and communicating baselines should be considered in the planning phase. In such cases, operational limits are kept conservative, and penalties apply for partial or missed activations with no remuneration for activations exceeding requests (Figure 8).

Furthermore, the challenges associated with the uncertainty of resource activation include those related to the definition of baseline. Since the compensation for resources depends on the difference between activation and the baseline profile, uncertainties in these two components combine, affecting the estimation of the service activation level. Therefore, the aggregator must account for these uncertainties in defining offers, and, subsequently, the market operator must ensure that the aggregates adhere to the baseline profile and activation, allowing for a margin of error on both. For all the methods to define the baseline profile, its precision decreases as the size of the aggregation perimeter decreases. For these reasons, the more granular the market, the more attention must be paid to baseline profile definition, and various methodologies can be employed, with varying effectiveness depending on the types of users considered. For instance, if a network node serves a controllable generator, the baseline profile may align with the production schedule defined by the aggregator within the day-ahead market. If a node predominantly serves residential users, statistical methods must be employed to predict its profile and to consider a defined confidence margin during the verification of the service delivery.

Table 2. Literature on baseline calculation methods.

Ref.	Baseline Calculation
[32]	The baseline profile is an estimate of the consumption of a set/group of loads, assuming that they are neither under external control nor show effects of past or future external control (external control refers to the actions that an aggregator would perform to activate flexibility). The load schedule is assumed to be equal to the expected consumption declared by the related aggregator on the day before. In cases where there are several metering points with the same aggregator, the aggregator will predict the entire consumption. If the aggregator does not respond to the day-before signals, the load schedule and baseline are the same.
[87]	The baseline profile is agreed upon by the parties on the basis of historical behavior of the involved energy resources, taking into account any changes in the year or weather forecasts.
[88]	The techniques for calculating the baseline profile vary depending on the availability of measurements. When a history of power exchange profiles is available, statistical regression methods use the collected data to calculate load forecasts. If the information is not available, the consumption of a few users of the same typology can be measured and then used as a reference for all other users of the same type.
[89]	The baseline profile is calculated using historical data by means of simple and well-defined methodologies. In particular, the methodology of similar days [95], which is widely accepted and used by the New England balancing operator, results in being the best among those analyzed [96].
[90]	The baseline profile is calculated as the average of the power exchange of the time frames similar to the reference time frame for the specific market session.
[91]	The actor in charge of calculating the basic profile may vary from case to case. It can be calculated by an independent external entity or agreed upon by the different market players. Since aggregators optimize the exchange profiles of all controlled resources to minimize the cost of procurement for each of them, this schedule can be used as a baseline profile. The DSO, also based on the basic profile communicated by the aggregators, predicts the evolution of the loads and generators that are used to determine the flexibility needs.
[92]	Aggregators optimize their portfolio of assets (e.g., in the day-ahead market) and communicate the resulting profile to the local flexibility market.
[93]	The results of previous markets (energy and/or ancillary services) can be used to determine the baseline profile with which flexibility needs can be defined.
[94]	There are two options for calculating the basic profile: Market participants must declare the ex-ante baseline profile; The market buyer (TSO, DSO, etc.) calculates the ex-post base profile using meter measurements.
[97]	The calculation of the baseline profile and the modulation of the load due to the activation of the service is carried out via the measurements of the meters of the resources participating in the market. The meter data must therefore be accessible to the market operator.
[98]	The day-ahead market results define the baseline profile.
[99]	The baseline can be calculated from historical data, and different methods of calculations are illustrated in [99]. The survey shows that the most promising methods are characterized by calculation procedures that are easy to implement, but at the same time accurate to avoid opportunistic behavior.
[100]	The baseline profile can consist of a forecast made by the aggregator. In other cases, the profile may be calculated by the market operator and communicated to the aggregator. In the first case, the profile takes on two roles: Communicate the planned production/consumption profile so that the DSO can include it in a network security analysis; Define the reference for flexibility exchanges between the aggregator and the DSO.

Table 2. Literature on baseline calculation methods.

Ref.	Baseline Calculation
[32]	The baseline profile is an estimate of the consumption of a set/group of loads, assuming that they are neither under external control nor show effects of past or future external control (external control refers to the actions that an aggregator would perform to activate flexibility). The load schedule is assumed to be equal to the expected consumption declared by the related aggregator on the day before. In cases where there are several metering points with the same aggregator, the aggregator will predict the entire consumption. If the aggregator does not respond to the day-before signals, the load schedule and baseline are the same.
[87]	The baseline profile is agreed upon by the parties on the basis of historical behavior of the involved energy resources, taking into account any changes in the year or weather forecasts.
[88]	The techniques for calculating the baseline profile vary depending on the availability of measurements. When a history of power exchange profiles is available, statistical regression methods use the collected data to calculate load forecasts. If the information is not available, the consumption of a few users of the same typology can be measured and then used as a reference for all other users of the same type.
[89]	The baseline profile is calculated using historical data by means of simple and well-defined methodologies. In particular, the methodology of similar days [95], which is widely accepted and used by the New England balancing operator, results in being the best among those analyzed [96].
[90]	The baseline profile is calculated as the average of the power exchange of the time frames similar to the reference time frame for the specific market session.
[91]	The actor in charge of calculating the basic profile may vary from case to case. It can be calculated by an independent external entity or agreed upon by the different market players. Since aggregators optimize the exchange profiles of all controlled resources to minimize the cost of procurement for each of them, this schedule can be used as a baseline profile. The DSO, also based on the basic profile communicated by the aggregators, predicts the evolution of the loads and generators that are used to determine the flexibility needs.
[92]	Aggregators optimize their portfolio of assets (e.g., in the day-ahead market) and communicate the resulting profile to the local flexibility market.
[93]	The results of previous markets (energy and/or ancillary services) can be used to determine the baseline profile with which flexibility needs can be defined.
[94]	There are two options for calculating the basic profile: Market participants must declare the ex-ante baseline profile; The market buyer (TSO, DSO, etc.) calculates the ex-post base profile using meter measurements.
[97]	The calculation of the baseline profile and the modulation of the load due to the activation of the service is carried out via the measurements of the meters of the resources participating in the market. The meter data must therefore be accessible to the market operator.
[98]	The day-ahead market results define the baseline profile.
[99]	The baseline can be calculated from historical data, and different methods of calculations are illustrated in [99]. The survey shows that the most promising methods are characterized by calculation procedures that are easy to implement, but at the same time accurate to avoid opportunistic behavior.
[100]	The baseline profile can consist of a forecast made by the aggregator. In other cases, the profile may be calculated by the market operator and communicated to the aggregator. In the first case, the profile takes on two roles: Communicate the planned production/consumption profile so that the DSO can include it in a network security analysis; Define the reference for flexibility exchanges between the aggregator and the DSO.

2.4. Market Schemes and Bid Characteristics

Two types of procurement schemes are proposed in the literature for acquiring local flexibility services. The first is similar to a capacity market, where the availability of resource modulation is acquired, and the related flexibility is activated on demand by the DSO based on their merit order list and the forecasted violations. This alternative is referred to as long-term procurement and, generally, concerns an availability price (EUR/kW) and an activation price (EUR/kWh). The other alternative, also called short-term procurement, consists of a spot market, which consists of immediately settling the full payment in exchange for the service delivery. The two schemes can be merged in a single procurement strategy: first, the DSO acquires the estimated necessary flexibility in advance, and then the activation is determined by running a spot market near real time where also additional resources can participate. Furthermore, there are two main acquisition mechanisms of the bids: continuous trading and periodic auctions. The latter is most used since it allows to consider in a single optimization process all the bids and constraints. Once the bids are collected, the optimum solution can be determined by two possible clearing mechanisms: pay-as-bid or pay-as-clear. Finally, the interaction schemes between the market participants can be one-sided if the only buyer of the resources is the DSO (or the TSO) or two-sided if also BRP can buy flexibility.

The characteristics of the market, the acquired services and the bid structure determine the model and the algorithm used for clearing the market [9,101]. Their complexity increases with the intricacy of the services and bids (

Table 3. Main possible characteristics of flexibility bids [7].

Field	Description
Time window	The time window in which flexibility must/can be provided
Modulation power	Active or reactive power
Alert time	Time in advance communicated the request for actual activation
Congestion/location area	Geographic region within which resources must be located to provide the requested service
Quantity	Amount of flexibility required, typically a difference from an agreed base profile
Maximum duration of service	Maximum duration of service provision
Minimum duration of service	Minimum duration of service provision
Capacity remuneration	Price that the aggregator (or resource) offers for resource availability
Volume	Price for energy modulation
remuneration
Maximum number of activations	Number of activations that can be requested in a given time interval
Recovery time	Minimum time interval between two successive activations
Penalties	Penalties in case of non-compliance with dispatching orders
Minimum quantity	Minimum quantity to offer
Maximum quantity	Maximum quantity offered
Ramp constraint	Maximum variation to rise or fall in power
Full activation time	Time needed for the resource to reach the proposed power delivery
Divisibility	Possibility of partial acceptance of the offer
Resource Type	Source type
Rebound effect	Describes the characteristics of the possible rebound effect due to resource activation

, defined on the basis of [11,19,92,102,103,104]). For example, if only the congestion management on the distribution network is considered then the market could be cleared with a linear programming optimization (i.e., DC–Optimal Power Flow). Instead, if the market is also used to solve voltage violations and the bids are not divisible (i.e., a bid has to be completely accepted or not), it must be based on a non-linear and/or mixed-integer problems (i.e., AC model of the network, which is much more complex than a DC–Optimal Power Flow). A list of techniques for the optimization of local markets, together with the mathematical modeling and simulation tools, are reported in [105,106,107] with particular reference to the interaction between DSO and TSO.

3. Characterization of Proposed Markets

Starting from the description of local markets, the number of proposed markets implementing each feature is evaluated to give an idea of which are the most implemented solutions (

Table 4. Amount of investigated local markets for each particular feature [7]. The acronyms of coordination schemes are presented in Section 2.1.

Coordination Scheme between DSO and TSO

M0

M1

M2

M2/3

M3

M4

M5

No coordination scheme

Centralized market

Separated local and global market

Coordinate balancing

Shared responsibility

Common local and global market

Integrated market

8

11

20

1

8

5

5

Proposed services

DSO congestion

DSO voltage

Balancing

TSO congestion

TSO voltage

Island operation

40

26

24

25

20

4

Buyers/sellers coordination

One-sided

Two-sided

45

9

Negotiations

Periodic auctions

Continuous trading

33

10

Price formation

Pay as bid

Pay as clear

34

6

Type of participants

Any

Loads

Storages

Generation

44

3

4

3

Aggregators

Allowed The DSO usually defines the perimeter of aggregation.

Nodal The aggregation is possible, but only at the nodal level

DSO The DSO is responsible for the aggregation

37

7

2

Time horizon

Short period Market closure hours before real time

Long period Market closure Month or years before the real time

35

10

). It is important to notice that the characterization is merely indicative: each proposed market possesses unique attributes that may not align with any scheme. Furthermore, not all aspects of the market are consistently detailed.

One of the fundamental aspects is the coordination between DSO and TSO, as local services must be managed to avoid impacting the operation of the transmission network. While the number of projects does not necessarily indicate the effectiveness of a scheme, it does offer an indication of the most promising or readily implementable option. The M2 scheme emerges as the most frequently proposed, likely due to its balanced interaction between DSO and TSO: the DSO primarily validates bids for the TSO, maintaining a straightforward interaction. Conversely, the M2/3 scheme is the least used, possibly because it closely resembles M2 but lacks the capability to transfer unused local offers to the TSO. With the exception of the M2/3 scheme, all adoptions are investigated in various projects. However, the prevalent scheme in actual platforms utilized by network operators mainly falls within the first three categories: M0, M1 and M2. In these schemes, the role of each actor is well defined since each service is managed by a single operator in a specific market (e.g., congestion for distribution network in local markets). In this way it is also simpler to associate the power modulations of balance responsible parties to the specific network operator request.

The necessity for coordination with the TSO also influences the services considered by the local market. While resolving DSO congestion remains the most investigated service, as it aligns with the main objective of local markets, services directed to the TSO (balancing, congestion, etc.) are commonly integrated into the market. This integration is crucial because the activation of distribution resources has the potential to exert significant effects on the transmission network operation. However, as explained in the previous paragraph, the provision of local and global services is not usually considered in a single market, but the global services are negotiated only after the local market is cleared.

The majority of proposed markets follow a one-sided model, being characterized by a single buyer of the service (either DSO or TSO) and multiple sellers. However, in certain markets that incorporate balancing services, a two-sided mechanism is proposed, allowing balance responsible parties to procure flexibility for their own interest. The one-sided model has the advantage of associating each resource activation with the corresponding provided service. It allows for a more straightforward implementation of long-term contracts since each network operator can independently define and acquire its flexibility needs. The two-sided mechanism allows for clearing in a single market many flexibility needs, optimizing the network in a single run and avoiding the need for multiple markets. However, this requires a clear definition of the role of each actor in the common market and a more complex division of cost and benefit since a single power modulation can solve multiple services.

Periodic auctions are the favored method of negotiations, likely attributed to the necessity of incorporating network constraints into the market optimization processes. In particular, grid constraints introduce complexity and time constraints in the optimization algorithm of the market, making the implementation of continuous-trading markets less feasible. In addition, periodic auctions are more easily integrated into the already existing energy and flexibility markets, which have a well-defined time schedule.

The complexity of the market, network constraints, potential liquidity issues and the presence of only one buyer (the DSO) likely contribute to the preference for pay-as-bid price remuneration. Furthermore, in most cases, the market only determines the selection of resources rather than their activation. Activation signals are dispatched only upon identification of constraint violations based on the merit order list of received bids.

While most projects permit the participation of resources that comply with defined market rules, there are exceptions where implementations focus on specific resources (e.g., storage systems), although such cases are limited. This second choice, even if not technologically neutral, is driven by the issues of integrating a generic market resource that has specific characteristics (e.g., the capacity limitation of the storage system). In cases where there is an abundance of a specific asset, it is easier to implement a market that is designed accordingly to the resource specification. This issue is usually present when there are resources with a capacity constraint (e.g., storage system), and it is necessary to define if this constraint is integrated directly into the market algorithm that autonomously determines the best use of the resource or the responsibility of presenting offers that respect the capacity constraint is left to the balance responsible parties. Usually, when multiple resources are present, this second choice is the most common since it makes easier and more transparent the interpretation of the market results since all the resources, both with and without capacity constraints, are treated equally by the optimization algorithm.

Aggregation is typically permitted, with the DSO defining the perimeter of aggregation based on required services. In certain instances, aggregation is nodal (e.g., medium voltage node) to circumvent the need for defining perimeters, or it may be altogether unfeasible. In exceptional cases, the DSO is directly responsible for the aggregation of resources, in particular in the market scheme M2, where the DSO is responsible for submitting unused bids to the global market. The perimeter of aggregation strictly depends on the requested services: in case of congestion, only the resources that contribute to solving the issue can be aggregated (e.g., in case of overcurrent in an MV/LV transformer, only the underlying resources can be aggregated).

Most markets operate on a relatively short time horizon, with gate closures typically set at day-ahead or intraday intervals immediately preceding or following energy market closures. While some markets extend over longer time horizons (e.g., annually), various activation process alternatives exist beyond gate closure timing, such as sending activation signals immediately post-market closure or only in the event of a violation. The time horizon depends on the capacity of operators to foresee the service needs and on the time schedule of already existing markets. However, for brevity, these alternatives are not reported.

From Table 4, the characteristics of the most prevalent market can be outlined. The primary buyer of services is the DSO, tasked with defining service parameters (e.g., aggregation scope, time horizon) and forecasting required quantities. The local market operator collects flexibility requests from the DSO and offers from local resources, typically aggregated. This process considers periodic auctions with a daily or longer time horizon if flexibility procurement is needed further in advance. After the gate closure, a merit-order list of available resources is compiled based on bid prices. Subsequently, the local market operator, in collaboration with the DSO, reserves necessary bids for distribution operations and forward the remaining bids to the TSO. Additionally, the DSO communicates to the TSO that resources cannot be utilized for global services. As real time approaches (time horizon contingent on forecast quality and TSO market coordination), the DSO activates required resources based on anticipated violations and communicates activation levels to the TSO, possibly via the local market operator. Resources are then compensated based on their activation and bid prices, with additional compensation possibly allocated for resource availability in long-term procurement mechanisms.

The presented coordination schemes mainly consider how DSO and TSO can interact during the procurement of services in short-term markets without considering how this coordination can be affected by the presence of long-term contractualization. Indeed, in the literature [108], it is pointed out that voluntary markets (such as short-term procurement solutions) often do not represent the best alternative. In support of this statement, the INC-DEC strategy is explained: when service providers can anticipate market prices, they are inclined to bid strategically. Giving an example of bidding strategy: in a region characterized by low electricity supply, producers will bid at high prices to maximize their profits in the following re-dispatch market. Such behavior can aggravate network congestion, increase electricity prices and distort financial markets and price signals used to define where to build new plants. Along these lines, the same behavior could be expected during the procurement of local flexibility services. To prevent such phenomena, long-term contracts are a viable alternative. Indeed, when long-term procurement is deployed, availability and activation prices are established a priori; thus, the possibility of gaming is averted. However, when long- and short-term procurements are used together, the short-term market competitiveness can be compromised because of the effects of activation prices defined in the long-term procurement.

4. Identified Barriers

On the basis of the described review of established local markets for procuring ancillary services at the distribution level, numerous significant barriers have been identified during the testing phase. These barriers can be categorized into three distinct groups: technical, economic and regulatory.

Technical barriers consider all aspects related to available technologies and infrastructure, as well as constraints imposed by current market design and the need to define new technical parameters. Economic barriers, on the other hand, concern monetary considerations associated with the development and deployment of these technologies, the emergence of new costs linked to procuring such services and the necessity to establish new economic relationships among market participants. Lastly, regulatory barriers primarily relate to the lack of defined roles, coordination procedures and priorities, as well as the necessity to reform electricity markets to incorporate the new services offered by distributed flexibility resources. The analysis of these barriers is based on the interpretation of the authors, starting from the consultation of the literature reported in the following subsections. Particularly relevant for the scope of the analysis are observations collected in [94,109], from which most of the barriers have been extracted and integrated with the experience gathered from other projects.

4.1. Technical Barriers

Technical barriers pose a significant challenge to the implementation of local flexibility services, particularly concerning technology and telemetry prerequisites (

Table 5. Main technical barriers limiting the adoption of local markets.

The difficulty for some types of users, especially smaller user loads, to offer flexibility with sufficient performance is due, for example, to the following: Impossibility of modulating the power exchanged in some types of converters (e.g., older photovoltaic systems); The difficulty of calculating the baseline; The scarcity of the quantities offered; The uncertainty of correct activation.

The characteristics of bids for balancing markets (high minimum size…) exclude the participation of smaller resources and cannot therefore be replicated in local markets. It is then necessary the harmonization between offers presented in local markets and those presented in global markets.

Insufficient measurement and communication standards of smart meters (number of measurements, precision, update period, communication channels).

Poor experience of aggregators in providing services at the distribution level and DSO in quantifying and resolving violations via ancillary services.

Lack of standards for the following: Data model; Data exchange (protocols…); Functionality of the markets; Characteristics of the offers; Methods of control of distributed resources; Prequalification procedures.

Uncertainty that the local market can guarantee sufficient flexibility over a long time horizon.

Difficulty of DSO to forecast network violations both for long (years) and short (hours) time horizons due to the uncertainty of scenarios and the small aggregation perimeters of resources.

Definition of the baseline profiles of resources participating in the market and coordination with the baseline of other markets.

Training of employees, both DSOs and aggregators.

Difficulty in defining flexibility services for low voltage components, which are often poorly monitored, and also due to the low number of users involved, which makes efficient acquisition of regulation resources difficult.

). Strict market access requirements, established in previous years looking at traditional generation characteristics, introduce significant limitations to the integration of distributed resources. For instance, specific constraints for providing balancing services, such as minimum bid offers and deviation penalty designs [94], may not align with demand-side management participation, excluding them from the provision of balancing services. Therefore, bearing in mind that some services, such as balancing ones, are essential for ensuring system stability, access barriers for the provision of different services should be smoothened in order to foster the engagement of different resources. For example, looking at congestion management in the distribution network, fewer technical requirements exist. Resources can provide additional services beyond what is needed, and the unrequested modulation is typically not remunerated and does not determine issues. Given the varied technical characteristics of potentially beneficial activations (e.g., fewer resources and lower modulable power), it is important to establish dedicated requirements for services procured in local ancillary services markets to avoid the exclusion of significant resources. Basically, the idea is to go from technology-specific requirements to service-specific requirements, allowing a larger pool of resources to provide services in aggregated configuration or individually.

Another technical barrier is identified in the determination of the baseline. The assessment of the baseline is a highly discussed topic when defining parameters, and ensuring a certain level of accuracy is mandatory as settlement, and consequently, service remuneration, depending on the baseline definition. In this regard, there are two potential approaches: service providers may declare the baseline ex-ante, or the market operator could calculate it using metering data and utilizing knowledge of reserved energy for the service. While both strategies are functional, neither guarantees a precise evaluation of the energy provided for the service. For example, oscillations in generation/consumption profiles due to external factors could be inaccurately assessed, being considered as a provided service even if not voluntarily activated. Moreover, it is essential for the baseline to be included in data exchange between the TSO and DSO. A dual and separate evaluation could lead to discrepancies and necessitate additional coordination efforts. This particular aspect should be incorporated into the interactions between system operators, which are mainly associated with regulatory barriers. Lastly, standardizing procurement timing and the amount of capacity allocated to a particular service is crucial to prevent distortions caused by local market activities on the global market.

The requirements must be taken into account in the establishment of the prequalification processes which are typically technology-specific rather than being service- or product-specific [109]. Furthermore, prequalification is presently defined at national level and separately for different technologies, which increases the necessary effort and introduces some discrepancies among national regulations.

Establishing a common set of attributes for new flexibility services and standardized prequalification procedures could facilitate the development of new local markets, introducing minimal standardization at the European level. This collection of parameters should also include requirements for verifying provided services, including measurement frequency, accuracy and communication channels. From the perspective of network operators, there are additional technical challenges that remain unresolved. Implementing a local market necessitates the capability of the network operator to identify and quantify forecasted violations. To achieve this objective, it is important for the DSO to develop both long- and short-term network observability tools for identifying congestion [110]: long-term tools assess the benefits of local markets compared to network reinforcements, while short-term tools are essential for accurately identifying flexibility needs and subsequently acquiring services within the market. A crucial component of the observability tool is the digitization of network assets, enabling the accurate modeling of violations and resources eligible to participate in the market.

Finally, another critical aspect consists of the necessity of training employees, both within DSOs and aggregators, to effectively manage and participate in the local market.

4.2. Economic Barriers

The establishment of local flexibility markets requires significant investment across the entire system, impacting the business cases of involved actors (

Table 6. Main economic barriers limiting the adoption of local markets.

Lack of experience and data on the possible profitability of local markets that allows aggregators to develop business cases.

Difficulty for users, especially industrial ones, to change their business model.

Possible benefits for small and medium users are often limited.

Uncertainties about possible profits, but also about possible penalties in case of failure to provide the service.

High initial costs to enable resource flexibility.

Uncertainty for DSOs when comparing reinforcement and flexibility costs.

Lack of criteria to remunerate investments in information and telecommunications technologies by DSOs.

). Indeed, the lack of expertise and historical data are always characteristics of first-of-a-kind mechanisms, such as the procurement of flexibility resources via local markets. The profitability of local markets is a subject of ongoing discussion and is not easily predictable due to various factors. These considerations pose challenges for balancing service providers in formulating business cases and for potential end users in adjusting their usual business practices.

Ensuring the qualification of resources is essential to guarantee efficient service provision. However, market requirements must also be flexible enough to encourage an acceptable number of resources to participate. Achieving a satisfactory number of involved resources becomes crucial when the overall qualified capacity is significant, particularly to achieve a significant contribution from the demand-side response. This highlights the importance of industrial consumers consistently participating in local flexibility markets. Moreover, accurately evaluating and comparing the costs associated with grid investments versus those related to procuring flexibility services is complex for DSOs. The latter costs should encompass considerations such as payment for imbalances on the central market (resulting from local activations), investment in communication and information infrastructure, costs associated with unplanned unavailability of flexibility resources and others.

Furthermore, the potential for the exercise of market power should be considered, and mechanisms to counteract opportunistic behavior should be established. For example, while market-based procurements are typically preferred, in regions characterized by resource scarcity, must-run dispatches should be considered as a potential alternative.

4.3. Regulatory Barriers

Regulatory barriers primarily consist of gaps in defining the roles and responsibilities of market participants, such as aggregators and network or market operators, as well as the products and services for the local flexibility market (

Table 7. Main regulatory barriers limiting the adoption of local markets.

Lack of definition of possible local services (resolution of voltage violations via reactive power…) and mechanisms for providing local flexibility.

Absence of definition of roles, functions and responsibilities of the participating actors.

Lack of rules for the coordination between TSOs and DSOs and, more generally, between local services and system services.

Asymmetry of treatment between OPEX and CAPEX in the remuneration of DSOs: Absence of local flexibility regulation. DSOs are currently remunerated only on historical spending (especially for OPEX). There is a lack of remuneration schemes based on long-term development plans. Lack of incentives to extend the useful life of components.

Reduced possibility to implement sandboxes by DSO and TSO to implement local flexibility markets that extend the present regulation.

Lack of transparency on the residual capacity of the network to install new loads and generators and on the network reinforcement plans, which allows for identifying the areas where flexibility would be necessary and how much it could be remunerated.

Lack of guarantees of independence between the various market players (DSO, aggregators, market operators…).

Lack of connection costs and network tariff structure linked to connection capacity instead of energy consumption.

Lack of privacy protection of network data and users participating in the market.

Lack of regulation to avoid the risk of opportunistic behavior, for example, due to scarce liquidity in local markets.

). Resource aggregation is fundamental for the management of very small resources; thus, it is essential to understand how aggregators should delineate the characteristics of a particular portfolio. In this context, challenges arise due to a lack of historical data, particularly when portfolios comprise different flexibility technologies with varying technical characteristics. Furthermore, network operators (i.e., DSO) must guarantee their independence from market participants, avoiding conflicts of interest during market procedures. This also requires the definition of open and transparent market-based procedures, highlighting technological neutrality.

Furthermore, a lack of definition regarding the services acquired in the local market remains. To provide support to regulatory development, network operators should conduct a comprehensive analysis focusing on system needs and what services could address them. This analysis should encompass issues experienced on their infrastructure, including the type of violation, frequency of occurrence and required power modulation, among others. Additionally, this analysis serves as the basis for defining products, which should also consider the technical input from aggregators and final users. For instance, exploring the feasibility of internal resource balancing within the same aggregator deserves further examination.

Furthermore in a system where flexibility services become a fundamental aspect, the potential and integration of these services should be fully exploited. Taking into account that both TSO and DSO will benefit from flexibility services offered by the same pool of resources, coordination becomes essential. Indeed, a well-designed regulatory framework should include a coordination scheme between TSO and DSO: this scheme should outline priorities between different network operators for reserving flexibility resources, services prioritization, data exchange protocols, common prequalification procedures and other relevant aspects.

Moreover, the regulatory framework can serve as a tool for overcoming economic barriers imposed by remuneration mechanisms currently applied. In this regard, incentivizing mechanisms for TSO and DSO primarily focus on capital expenditure (CAPEX), considering the required investment. While remuneration mechanisms targeting capital investments facilitate the deployment of flexibility resources and support infrastructure and telemetry installation, incentives should also consider operational expenditure (OPEX) associated with developing an additional market. Furthermore, DSOs are currently remunerated based on OPEX evaluated from historical expenditures. Evaluating these costs should be reviewed to introduce incentives that support DSOs in managing and operating local flexibility markets, which could help remove economic barriers. To accurately determine when the local market is more effective compared to traditional network reinforcement, these incentives should be incorporated into network planning procedures, which ought to consider both OPEX and CAPEX costs [111]. Additionally, this integration enables DSOs to account for the management of their assets, such as network reconfiguration and the use of voltage-regulating transformers.

For the same reason, the potential contribution of distributed resources to flexibility provision should be considered in the TSO planning process, necessitating better coordination with the DSO planning procedures. An illustrative example of this coordination phase can be found in the FlexPlan project, where a grid expansion planning tool considers flexibility assets alongside conventional grid expansion. This facilitates correlation between TSO and DSO network expansion via a method known as T&D decomposition [112]. For instance, it would be appropriate to include considerations for costs arising from the need to balance positions on the global market when a resource is activated by local ancillary services. Moreover, there is a lack of transparency in grid expansion planning, with data usage and sharing restrictions imposed by privacy policies, which is an obstacle to the development of flexibility resources in critical and strategic locations. Recently, steps forward have been taken by the European Commission, with DSOs now required to share their grid expansion planning [1]. Additionally, new guidelines for the utilization of demand response in local markets have been issued by ACER [113].

Additional barriers are identified in the definition of regulated tariffs. The current design and definition of tariffs are not always favorable for potential flexibility providers and often fail to support the deployment of distributed flexibility resources. For example, existing tariffs for on-site exchange, which remunerate the injection of internally produced energy in terms of EUR/MWh, typically feature fixed prices that may not align with market trends. This reduces the benefits of resources being able to adjust their power output on the basis of network requirements.

Overall, following EU regulatory guidelines, flexibility resources should compete in local markets for the provision of services, similarly to what is developed in global markets. However, considering a transition period where there is no assurance that enough flexibility will be available, permissions are granted when the amount of flexible capacity is not sufficient to reach a satisfactory level of liquidity. In case of scarce liquidity, different strategies can be deployed, such as the use of long-term availability/options contracts, mitigating the risk of market power exertion from strategic, flexible entities.

5. Conclusions and Proposals

The manuscript aims to report an exhaustive list of references to the current and recently concluded initiatives on local flexibility markets. From the analysis of their documentation, relevant projects have been identified, also considering a wide spectrum of accessible and promising implementations. Indeed, experiences with various alternatives allowed us to investigate distinct market models and TSO-DSO coordination schemes. Finding the optimal architecture has proven to be challenging as each approach carries its own set of advantages and disadvantages, depending on existing markets, such as the global ancillary service market. From the analysis of the investigated projects, the authors summarize a list of the most compelling barriers and briefly report them in the following paragraphs.

Common barriers emerge from current regulations, encompassing technical requirements and economic parameters. One significant challenge lies in ensuring that flexibility resources can participate in local markets at a level playing field. Existing qualification procedures and technical requirements must be tailored to suit distribution system needs, which differ from those of transmission networks. Consequently, parameters like the minimum required bid should be adjusted to account for the typically smaller sizes of demand-side management and other distributed resources. Moreover, there is a need to reevaluate prequalification procedures, shifting from technology-specific assessments to service-specific evaluations. This change could facilitate access to both local and global flexibility markets, encompassing resources capable of providing a limited range of services. Additionally, adopting a broader perspective beyond national boundaries may facilitate the establishment of standards for distributed resource participation in local markets. Given the similarities in distribution network needs across different countries, a common definition of services and resource requirements could be advantageous.

To date, DSOs have predominantly applied the Fit&Forget approach. However, with the definition of local markets, this grid expansion strategy becomes inadequate. It is necessary for DSOs to acquire expertise in long-term planning procedures to identify and assess potential grid violations. In this context, coordination between TSO and DSO can be proven to be beneficial. DSOs could leverage the TSO experience to implement long-term network observability tools. Additionally, transparency should be prioritized, especially for nodes where system flexibility is crucial, to guide service providers in investing in strategic locations.

Regarding TSO-DSO coordination, efforts should be directed toward defining interactions between local and global markets. Ideally, resources should be permitted to offer services in both markets, with final selection and activation aimed at minimizing overall system costs. However, such a solution would drive significant computational effort and complexity. Therefore, coordination for selecting shared resources should be identified to facilitate the development of local flexibility markets. Furthermore, common requirements for shared services should be defined to eliminate the need for additional testing procedures once a resource is enabled in the local market, thereby simplifying the process.

There are numerous evident gaps in current regulations, particularly in the absence of clear definitions of roles and responsibilities. Regulatory updates are necessary to deepen the concept of resource aggregation and to promote a transparent market that integrates distributed resources on a level playing field, whether aggregated or not. Criteria for prioritizing DSO and TSO needs should be established to encourage the development of impartial processes for resource selection. These criteria should aim to minimize overall system costs while considering social welfare aspects such as health and established common objectives.

Furthermore, incentivization mechanisms should be formulated to encourage the development and deployment of flexibility resources. These mechanisms should also address the additional costs incurred by system operators for managing local markets and the required infrastructure. A market reform could facilitate the development of strategies for procuring flexibility services: while local markets should always be the preferred solution, in cases where liquidity is limited due to resource scarcity, flexibility contracts should be initiated to leverage the potential of strategic resources.