Assessment of Italian Distribution Grids and Implications for Energy Communities’ Integration: A Focus on Reverse Power Flow and Energy Balance

Abstract

This study evaluates the potential impact of new energy communities (ECs) on the electric infrastructure within the Italian regulatory framework using publicly available information on reverse power flow metrics in high-voltage (HV)/medium-voltage (MV) interfaces and calculating the municipal energy balance. The current legislation is incentivizing EC configurations where members connected to the same HV/MV interface are sharing energy, predominantly produced by new-generation units. To identify critical territories, primary substation service areas are overlapped with reverse flow occurrences, focusing on cases that exceed 5% of the year. The output is utilized to indicate the municipalities that fall within these areas. The municipalities deemed critical are further evaluated, defining a Key Performance Index (KPI) as the ratio of local production capacity to consumption, with generation data procured by the national database on production units and load estimates derived from provincial cumulative data, adjusted using census information on population and employment with a municipal resolution. A piecewise linearization approach is employed to examine the cumulative distribution function (CDF) of the KPI, enabling a traffic light-like criticality classification. The results provide a relative assessment and highlight municipalities with a higher risk of detrimental impact of EC adoption within the current framework. The outcome is presented as a national georeferenced map illustrating the municipal criticality. This emphasizes the need for revising the regulative framework, potentially enabling the utilization of existing generators in critical areas and leveraging load flexibility and increased local energy sharing to procure benefits from EC adoption.

1. Introduction

The establishment of Energy Communities (ECs) in Europe, particularly in the case of Italy, is closely aligned with the guidelines and objectives for energy and climate defined by the European Union (EU). The regulations put forward prioritize sustainability, energy efficiency, and most importantly, the promotion of renewable energy, as outlined by the EU’s treaties and legislation [1]. Considering such commitments, in 2010 the European 2020 strategy projected specific targets for these improvements known as the “20-20-20” targets. This strategy was followed by the European Energy Union Strategy in 2015 [2], aiming to enhance security and sustainability and promote integration across the member states. An important outcome of this strategy was the “Clean Energy for All Europeans Package” [3], a set of proposals launched in 2016 aimed at reforming the EU’s energy policy framework. In terms of EC adoption, the most significant legislative act arising from this package was the Renewable Energy Directive RED II (2018/2001), which focuses specifically on promoting the use of renewable energy sources (RES), as well as the sole concept of an EC for the first time in the European legislative framework.

Nevertheless, the implementation of EU-wide directives and the framework defining how ECs can be created and the geographical constraints for potential members have significantly varied across member states. For example, in the Netherlands, a virtual model has been adopted, with a subsidy scheme supporting RES-based ECs in a specific locality, incentivizing the energy generated by the community up to a predefined ceiling [4]. A different approach is adopted in Portugal [5], where the incentive schemes are based on a conventional grid model and encourage the aggregation of members within 2 km on the low-voltage (LV) network and 4 km on the medium-voltage (MV) network. In Spain, an administrative division-based approach is adopted, where rather than applying distance constraints, the EC members must belong to the same municipality [6]. In Switzerland, on top of belonging to the same municipality, members must also be supplied by the same Distribution System Operator (DSO). In Germany and the northern countries, the members simply must be located in the same building, in which case the electric grid connecting the members is considered a private network [7], whereas in Austria a database has been created to associate each user to the MV/LV substation of origin [8], effectively identifying users who could join the same EC.

In Italy, the RED II directive was incorporated into national legislation, defining ECs as cooperative entities that enable local consumers and producers to share energy virtually using the existing public infrastructure. The added benefit compared to a physical model of energy sharing using infrastructure that is isolated from the public infrastructure is that there is no technical constraint for complete energy independence or self-sufficiency. Rather, the focus of an EC is on increasing self-consumption, incentivizing members to maximize the use of locally produced renewable energy.

The national law governing the adoption and incentivization of ECs in Italy is Resolution 727/2022/R/eel [9], published on 27 December 2022, which governs the development of energy communities in Italy. According to this legislation, the portion of the energy shared within the EC that is incentivized must be produced and consumed under the same high-voltage (HV)/MV substation, hereinafter referred to as the primary substation. In this way, the incentives designed for promoting EC adoption aim to valorize avoided energy losses and reduction in the electric infrastructure’s loading by increasing the local energy shared on the MV grid. Moreover, the framework imposes additional constrictions on the energy produced and the production units themselves that are published in [10]. A key sustainable factor is that no more than 5% of the energy produced within an EC can be generated by non-RES production plants. Therefore, in addition to fostering socio-economic benefits for the citizens and encouraging the informed utilization of electricity, the widespread adoption of ECs in Italy aims to support the decarbonization of the national energy sector.

With that being said, the restriction that is the main motivation behind the research in this paper is that the portion of the energy shared that is eligible for incentivization must be produced by power plants that are either new or have had their capacity enhanced, in which case they would only be incentivized for the enhanced portion of the production. This implies that only capacity commissioned following the implementation of the regulative framework is eligible for incentives. Considering the promotion of increased generation capacity and the fact that ECs rely on the public infrastructure for energy sharing, the careful evaluation of their impact on the national grid is crucial, both from a wide-scale and local perspective. The reason behind this need is that even though the creation of the EC will contribute to an increase in the energy shared within such EC, it may not be the case that the same will be true for the distribution network to which the members are connected. Given that the current legislative framework does not consider in any way the current status of the distribution network in terms of energy balance and existing production, promoting new generators as a requirement for EC adoption in networks saturated with local production would have a marginal impact on self-consumption. Instead, it would lead to overproduction and as a potential consequence, reverse power flow, which is a phenomenon where electricity flows in the opposite direction of its traditional path. It worsens the operating conditions of distribution networks not only due to increased energy losses, high voltages, and reverse power flow but is also the cause of further technical issues such as protection and voltage regulation coordination [11].

This paper proposes a procedure aimed to address the following research question: “What is the potential impact of EC adoption on the electric grid within the current legislative framework and how to guide policymakers in ensuring future grid sustainability?”. The assessment is conducted through a nationwide investigation of the Italian scenario, initially focusing on identifying territories that experience reverse power flow phenomena. Then, further analysis is performed with a municipal resolution, estimating the energy balance by calculating the ratio between installed distributed generation on the MV grid and annual electricity consumption. This ratio is promoted as a KPI to evaluate the potential criticality of new EC adoption on the electric infrastructure. Then, the municipalities are categorized into three levels of potential criticality using a clustering approach based on a linear approximation of the KPI’s CDF. The methodology proposed aims to identify and analyze publicly available datasets, facilitating the outcome to support informed decision-making and policy development.

The outcome of the assessment is presented in the form of a national map that aims to highlight regions characterized by a relatively high share of RES and limited energy demand, alongside territories with a low RES share and high energy consumption. facilitating the interpretation of the information even for non-technical members of national legislative bodies. Finally, the utility of this paper is not solely to propose an evaluation and categorization of the national territory, but to also propose a revision of the regulatory framework focused on mitigating the impact on the electric grid, proposing category-specific rules on incentivizing new production units for the purpose of EC adoption. The outcome can serve as a first-level decision support tool for policymakers, facilitating the effective development of energy policies. Additionally, it can aid citizens in improving their understanding regarding the status of the electric grid and identify the most suitable energy-related actions for the area in which they reside.

Limited by the availability of national-scale data, the first simplifying hypothesis is that the status of the energy balance on the electric grid can be directly evaluated using the annual consumption patterns and amount of production installed, whereas a time-series analysis would lead to a more accurate estimation, as further differentiation of different technologies could be considered. Still, this is partially remedied through the fact that a significant portion of Italy’s production is from a single source, as detailed in the following sections. Next, another simplifying assumption is that the energy balance itself is the sole index that could assess the potential impact; however, it is well known that the state and the topology of the local distribution grid can be equally important. However, the unavailability and sensitivity of data related to the electric networks is one of the main motivations facilitating the approach proposed.

The remaining part of this paper is structured as follows: Section 2 provides an overview of the state of the art in the relevant field, highlighting key studies and various approaches proposed in the literature. In Section 3, the methodology employed in this research is detailed, outlining the approach and techniques used to achieve this study’s objectives, as well as detailing the information utilized and the scope of the analysis. The results of the analysis are presented in Section 4, followed by a discussion in Section 5 that interprets the findings and discusses potential regulatory revisions based on the outcome obtained. Finally, Section 6 concludes the paper, summarizing the key aspects of the approach and the outcomes.

2. Energy Communities’ Impact on the Electric Infrastructure: State of the Art

Assessing the impact of ECs on the electric grid is a complex endeavor due to the necessity to evaluate various non-standard configurations and the sensitivity of the data relevant to electric distribution grids, which are not publicly available. Despite this, ECs are designed to promote local self-consumption within the distribution grid, thereby having a potentially positive effect on the infrastructure. However, in practice, operational challenges may arise, impacting technical factors such as voltage profiles, line loading, and short-circuit currents. Additionally, integrating more power generation into distribution networks (DNs) can influence both the quantity and pricing of electricity in the market, which in turn affects the operations of both, transmission and distribution networks.

While there is a substantial body of literature on energy communities [12], only a limited number of studies have evaluated their impact on the electric grid. The European Commission funded a very generalized project on a wide-scale assessment of ECs’ impact on grids, focusing on increasing the grid’s flexibility and the flourishment of the electricity markets. However, in this case, the analysis was general and failed to focus on a specific regulatory framework and a specific case [13]. On the opposite, a more focused study [14] assessed the potential impact of renewable energy communities (REC) on distribution grids, employing linear programming optimization to size photovoltaics and energy storage systems. However, this analysis relied on a deterministic approach using a simplified CIGRE distribution grid model. Similarly, the authors of [15] conducted a case study of an all-electric community in Denver using a physics-based urban energy modeling platform to assess the influence of various technologies on energy usage, carbon emissions, and peak demand. Nevertheless, the grid impact evaluation was primarily limited to energy balance goals, such as reducing the power peaks injected and absorbed by the grid. In [16], the focus is on detailed appliance modeling for specific houses, but grid impact was only estimated based on transformer loading. Large-scale studies, such as [17], which explore the impact of medium-scale EC development across Europe, typically concentrate on capacity expansion in cross-border transmission and national generation and storage. However, these studies often lack detailed grid modeling and instead assess net transfer capacity between virtual nodes representing specific areas, usually at the national level. In [18], an investigation was proposed into the interplay between energy storage ownership, electricity tariff design, and energy sharing within community energy systems. This study aimed to explore the impact of different storage systems, tariffs, and energy sharing on peak power exchanges and the autonomy of local energy communities.

Similarly, the study presented in [19] provides a detailed techno-economical evaluation based on high-resolution measurement-based datasets deriving from 3594 households. The study aimed to evaluate the correlation between the configuration and the performance of a solar-based EC. However, the impact on the electric infrastructure was limited to the estimation of the maximum import and export power flowing through the transformer. In [20], an approach towards facilitating an online peer-to-peer (P2P) energy marketplace, using initial statistics illustrated, based on simulated transactions facilitated through the platform, that the physical impact of the energy transactions on the power grid is important and cannot be neglected. On a parallel but slightly different path, the authors of [21] propose different methodologies to evaluate the hosting capacity of a distribution network, which within the Italian framework is correlated to the potential impact of EC on the electric grid as a result of the requirement for the installation of new production units. The study proposed a novel model considering grid-related uncertainties and multi-generator connections, enabling a fairly accurate estimation despite the absence of specific data on the electric network. The findings indicate that, in the existing literature, grid impact is generally evaluated in terms of the overall reduction in energy demand and peak load reduction in the specific EC investigated, rather than with a large-scale approach considering the distribution network.

A more grid-oriented approach is proposed in [22], where the goal is to assess how different energy communities could affect the operation of the electric grid. Real-life MV distribution grids relevant to the Italian context were considered, and the results showed that energy communities in rural areas could significantly influence the distribution grids’ losses and the loading factors of branches. The authors performed an annual load flow analysis with an hourly resolution, and the results pointed out that one of the main factors determining the impact of ECs on the electric grid is the share of already installed generators. Specifically, passive grids tend to show higher benefits in hosting ECs, compared to grids that already have a significant share of generators in place, as hypothesized by this paper. On a similar note, the work proposed in [23] aims to assess the grid impact of ECs but focuses on the LV grid. The authors simulated the formation of new energy communities over a wider area and found a significant impact on LV networks, potentially creating unforeseen contingencies and reverse power flow, concluding that grid enhancements will be necessary. Moreover, the authors of [24] investigated the implementation of a single EC and the relevant impact on the distribution transformer and found that even though the configuration led to an improvement in local self-consumption, it was also the cause of a reverse power flow during peak production hours. As a result, the authors proposed a demand-side management scheme in order to alleviate this negative impact on the electric infrastructure, which was indeed not completely able to solve the issue. On top of it, in practice, this type of demand-side scheme may lead to an increase in operational costs. These grid-centric studies were done within the Italian context, and they are relevant to the hypotheses made in this paper, as the impact evaluation is strongly connected to the legislative framework and the promotion of new RES-based generation capacities on the distribution network. However, the approach proposed in this paper aims to provide a more generalized, scalable, and public data-driven approach to assessing the potential impact of EC adoption, rather than analysis of a specific distribution network that necessitates sensitive data.

The literature review reveals a gap in the investigation of the impact of energy communities on the electric grid and broader energy infrastructure, highlighting a significant shortcoming in the planning of energy communities. Specifically, studies were found to either completely neglect the aspect of the impact of ECs on the electric infrastructure or fail to analyze the wide-scale aspect of it. It is evident that, especially on a national scale, the concept of grid-positive communities requires a more comprehensive investigation, as the current understanding remains vague and overly general. Moreover, given that specific distribution grid parameters such as energy fluxes, grid topology, and the capacities of lines and transformers are not publicly available, determining which EC configurations could be grid-positive becomes a challenging task.

In this paper, it is consequently assumed that a primary benchmark for classifying grid impact is correlated with the ratio between the share of distributed generators in place and the local consumption. However, this approach is not straightforward to implement in real-world conditions, as it is essential to first identify which distribution grid serves a given territory. Similarly, gathering data relevant to the distributed generators active in that area can be challenging. The most complex issue, however, lies in evaluating the energy demand of the area itself, i.e., the amount of energy consumed by the loads connected to the distribution grid under investigation.

The scenario is coherent with the target and the evolution proposed in the latest Renewable Energy Directive RED III [25], in particular, article 20a, which focuses on energy system integration, where the goal is to “facilitate energy system integration based on renewable electricity, and to ensure that the electricity system allows for a higher share of renewable electricity in a cost-optimal manner”. According to the document, this requires the definition of obligations related to data access, specifically requiring the Transmission System Operator (TSO) and if possible, DSOs, to make available information on the share of renewable energy and the greenhouse gas emissions content of the supplied electricity in their territory. The main target is to increase transparency and provide more information to electricity market players, aggregators, consumers, and end-users, including electric vehicle users.

3. Materials and Methods

This section details the pathway toward a country-scale assessment with a municipal resolution while utilizing data sources that are both distribution grid oriented and administrative division based. This approach allows for a complete overview of the problem using all the available resources to evaluate the potential issues of further integration of renewables on the distribution grid for the purpose of EC adoption by focusing the analysis on known reverse power flow issues and analyzing the current energy balance in order to obtain real information regarding the saturation of a territory with local production relative to the consumption. Therefore, the assessment procedure is conducted in two main steps, with the complete procedure being detailed in Figure 1.

The first step in addressing the research project was to identify the municipalities that exhibit local overproduction, which requires an effective and scalable method for representing the electric grid and power fluxes. This is a non-trivial challenge due to the unavailability of comprehensive country-wide data on electric distribution grids. The approach adopted for obtaining general information on the energy balance of the individual MV distribution networks is based on the previously mentioned Italian Resolution 727/2022/R/eel [9], which regulates the establishment of ECs in Italy based on the delineation of the territory supplied by each primary substation, i.e., conventional areas (CAs). In practice, the delineation of these areas is not straightforward, as complex configurations often exist. Simplifications were therefore necessary, and the resulting geographical areas were classified as “conventional areas” to emphasize that these are intended to provide a conventional clustering of the national territory. The goal was to create an approximated, yet easily interpretable and usable representation of the areas served by each primary substation.

The most widely adopted method for calculating these conventional areas employed by approximately 90% of Italian DSOs is the approach developed by the research team from Politecnico di Milano and authors of this paper, outlined under Italian Patent n. 102022000023970 [26]. As a result, a total of 2107 primary substations were identified across Italy, each with a geographically delineated territory. These data were aggregated and published by GSE, the Italian public company responsible for managing and promoting the country’s energy system, including renewable energy incentives, energy efficiency programs, and oversight of the electricity grid and energy markets. A country-wide map is now publicly available, delineating the service areas of each primary substation using polygons. This is illustrated in Figure 2, where different colors aim to distinguish the territories served by different DSOs.

With respect to the Italian regulatory framework, ECs that aggregate production and consumption resources geographically located within the same conventional areas are expected to contribute to local energy self-consumption, supported by an incentive scheme designed to promote more effective energy behavior, i.e., a renewable evolution coordinated with local energy needs.

Furthermore, since 2008, Article ARG/elt 99/08 [28] has been requiring distributors to publish information on HV/MV interfaces with a reverse power flow for at least 1% of the hours of the year and those with at least 5%. In Italy, there are a total of 124 distribution system operators, 25 of which are managing at least one HV/MV interface and sharing this information publicly on their website, such as the case of the largest distributor in Italy, e-distribuzione [29]. As a matter of fact, the definition of the conventional areas has added a significant technical significance to this information, as now the two can be merged in order to obtain a national-scale map of territories with diverse reverse power criticality, which is exactly the outcome of the first step of the procedure proposed. In cases where a primary substation contains more busbars with different levels of criticality, the worst one is considered.

However, this level of discretization is by far not sufficient for a proper evaluation, which is why the second step of the procedure leverages further administrative division-based information to enhance the outcome. In this case, the focus will be on the most critical group that exhibits a reverse power flow of more than 5% of the year. Given that the objective of this paper is to provide actionable information to policymakers to identify the most effective strategies for promoting energy communities in distinct regions, rather than fragmentation based on the structure of the electric grid, an administrative division of the Italian territory has been utilized to analyze and display the final outcome. Specifically, data calculated for each conventional PS area have been disaggregated to correspond to individual municipal territories.

First, the administrative division of Italy is made available by the National Institute of Statistics in a geospatial format [30], which allows the transitioning of the previous reverse power flow output to the municipal level. This is conducted by a set of geospatial operations including reprojection of the two geospatial datasets into a local coordinate system, following an intersection between the municipalities and the areas that exhibit a more critical reverse power flow that finds the shared territory among the two and provides information regarding the segment of each municipality that falls within an area with a critical reverse power flow. Subsequently, for each municipality the share of surface area within the critical areas is calculated. Finally, the initial criticality categorization of the municipalities is obtained with a pre-defined threshold ($A_{TH}$) regarding the share that should be within the critical areas previously defined in order to be considered as critical. This filter is necessary to avoid situations where municipalities are classified as potentially critical in cases where a marginal share of their area is within a critical conventional area.

The second step of the procedure is the further evaluation of the critical municipalities obtained utilizing open-source datasets and focusing on estimations regarding local production and consumption, where the aim is to identify, among municipalities with a reverse power flow higher than 5%, the ones that are the most critical in terms of high existing production relative to the consumption. First, the datasets identified as useful for this step of the procedure are summarized in

Table 1. Summary of the input data utilized in the KPI evaluation.

Data	Variable	Spatial Resolution	Dataset, Source
Number and size of generators	$P_i$	Single element (i)	AtlaImpianti [31]
Population distribution	${pop}_m$	Municipal (m)	Census [32]
Distribution of employees	${lav}_m$	Municipal (m)	Census [32]
Aggregated consumption of the domestic and tertiary sector	$E_{dom,p},E_{ser,p}$	Provincial (p)	Annual report [33]

, alongside the codification of the variables in the further steps.

The KPI defined to drive the municipality criticality assessment is the ratio between the installed generation and estimated consumption, weighted against the 8760 h of the year to arrive at more digestible values, calculated for each municipality (m) as in (1).

$${KPI}_m= {\displaystyle \frac{P_m \left[\mathrm{k}\mathrm{W}\right]}{E_m \left[\mathrm{k}\mathrm{W}\mathrm{h}\right]}}\ast 8760$$

(1)

The formulation proposed follows the logic that the risk of overproduction is proportional to the amount of installed production and inversely proportional to the electric load. The national data on generators installed, which include information about the municipality to which each generator is connected [31], were processed to determine the total installed generation on a municipal level $\left(P_m\right)$, while applying a filter $\left(P_{TH}\right)$ to amount only for the generators eligible for connection to the distribution network, as detailed in (2). According to the Italian technical connection norms for MV users (CEI 0-16) [34], production units of up to 10 MW can be installed on the MV grid, with the ones above 2 MW subject to specific feasibility analysis. This can guide the selection of the threshold $P_{TH}$, on top of the option for further sensitivity analysis as it is subject to case-specific uncertainty.

$$P_m=\sum P_i ; \forall { P}_i\in m \wedge { P}_i\le P_{TH}$$

(2)

Then, the municipal consumption ($E_m$) is estimated using the national annual report made available by the transmission system operator [33], which provides provincial consumption patterns based on the categorized utilization type. The procedure proposed uses the population distribution as proxy for domestic consumption and the employment as a proxy for the tertiary services consumption. Both the population and employment municipal distributions are made available by the Italian National Institute of Statistics owing to the results of the 2021 census efforts. The industrial consumption is disregarded at this step as it is generally connected to the HV network. The consumption data were first resampled on the municipal level by applying the spatial correlation municipality-province and using the population and employment ratios between each municipality and the province it belongs to. Equation (3) shows the estimation of the municipal domestic consumption, (4) is the tertiary services consumption, and, finally, (5) is the summation of the two as the final value for the municipality.

$$E_{dom,m}= E_{dom,p}\ast {\displaystyle \frac{{pop}_m}{{\sum }_{m\in p}{pop}_m}}$$

(3)

$$E_{ser,m}= E_{ser,p}\ast {\displaystyle \frac{{lav}_m}{{\sum }_{m\in p}{lav}_m}}$$

(4)

$$E_m= E_{ser,m}+E_{dom,m}$$

(5)

Finally, in the absence of a direct criteria for the KPI that can be used to perform the classification, a last step is introduced to perform a traffic light-like categorization of the municipalities’ perceived criticality. In this case, the indication provided by this classification is only relative, i.e., it differentiates between the municipalities already classified as critical. This classification is obtained with a two-step process. First, the cumulative distribution function (CDF) of the KPI is obtained for the critical municipalities. Then, the non-linear CDF is approximated with three linear segments using piecewise linearization, each defined as in (6) and representing one class in the criticality categorization.

$$f\left(KPI\right)=a_{k}x+b_k;x\in [x_k, x_{k+1}]$$

(6)

where the vector $x$ represents the break points that define the linear segments. The optimal break points are obtained with an optimization procedure based on Sequential Least Squares Programming [35] that minimizes the objective function defined as the root mean square error (RMSE) between the linear approximation and the original CDF, defined in (7). The values of the break points directly determine the boundaries of each criticality category within the classification of municipalities.

$$RMSE=\sqrt{{\displaystyle \frac{1}{N_M}}\sum _{m \in M}{[{CDF}_m-{f\left(KPI\right)}_m]}^2}$$

(7)

where $M$ is the vector of municipalities and $N_M$ is the number of municipalities.

4. Numerical Outcome on the Italian Scenario

The first outcome of this study is the areas that suffer from criticality in terms of reverse power flow in primary substations. Utilizing the geospatial representation of the primary substation service areas, the Italian reverse flow map is presented in Figure 3.

The resulting output, derived through the proposed method by cross-referencing the data from the map of Conventional Areas [27] with the data published annually by each individual DSO on reverse power flow in HV/MV interfaces, in accordance with Resolution ARG/elt 99/08 [28], provides an effective geospatial representation of the national territory, highlighting areas where the distributed generators already in place exceed the local energy demand for more than 5% of the hours in a year.

The next step was the evaluation of critical municipalities that lie within the most critical service areas depicted in Figure 3, or those exhibiting reverse flow more than 5% of the year (in orange). Following the intersection between the critical CAs and the municipalities described in Section 3, it was discovered that 24.5% of the municipalities lie entirely within the critical CAs, 50% are completely outside, and the rest fall partially within, with varying proportions.

For the remaining municipalities, the selection of the filtering parameter $A_{TH}$ impacts the selection of the municipalities for further analysis. This parameter was set to 20%, to exclude only municipalities that have their boundaries marginally intersecting the critical CAs. The map of the shares of municipalities within the critical areas is presented in Figure 4, whereas the map of 3523 municipalities included in the further evaluation considering the selection of the parameter $A_{TH}$ are displayed in Figure 5.

Next, the municipalities depicted in Figure 5 were subjected to analysis of the production and consumption detailed in Section 3. According to the annual consumption report published in [33], the national domestic consumption amounted to 64.525 TWh, with the tertiary services consumption equal to 94.697 TWh, distributed across the provinces. Following the scaling proposed by Equations (3)–(5), the electric consumption of the municipalities subject to assessment was estimated and is depicted in Figure 6 in a geospatial representation.

In terms of the generation evaluation, exploring the national database of generators available in [31] resulted in a total of 906,013 generators, only 590 of which are not eligible for a connection to the distribution as a result of their nominal power being over the 10 MW limit, leaving 26.46 GW of production to be potentially connected to the distribution networks. Interestingly, 903,400 of these generators represent around 99.8% in terms of number, but only 39.25% of the power will certainly be connected to the distribution networks around Italy. Everything that is between is a question mark and has a potential impact on the outcome. Figure 7 provides an overview of the different shares of energy sources in the overall mix, considering different thresholds for the power connected to the distribution network $P_{TH}$. It becomes clear that as the threshold decreases, the share of solar photovoltaics increases. This is well expected, as most of the small-scale production is indeed solar. Further analysis of this trend shows that solar PV production amounts to 90.6% of the production below 500 kW, 94.8% below 200 kW, and 96% below 100 kW, which is the standard limit for LV connections. This is useful to demonstrate that the decision on the $P_{TH}$ has a minor effect on the most dominant energy source, but instead mostly affects the production coming from hydro and conventional energy sources. Moreover, the fact that there is a dominant source of production connected to the distribution network indicates similar hourly production profiles and reduces the uncertainty introduced by considering solely the overall production patterns. Due to the different production profiles of different energy sources and the filter excluding the non-dominant resources, it can be observed that the overall peak production in most municipalities will not be significantly impacted by the selection of the filter $P_{TH}$. With that being said, for the purpose of this study, a value of $P_{TH}=6 \mathrm{M}\mathrm{W}$ was selected.

Then, this generation was distributed among municipalities, with the map presented in Figure 8 showing the spatial distribution of 24.43 GW of power installed on DNs across Italy. There are 515 (6.3%) municipalities with installed power greater than 10 MW, 97 (1.2%) with power greater than 25 MW, and only 3 with power greater than 100 MW, with Rome having the highest one, equal to 168 MW.

Having the production capacity installed within each municipality (Figure 8) and its estimated annual consumption (Figure 6), as well as the critical municipalities that require further investigation displayed in Figure 5, the KPI defined in Equation (1) is calculated. The CDF of this parameter is provided in Figure 9, whereas the geospatial distribution of the municipalities’ KPI value is provided in Figure 10. An interesting example is the case with the highest KPI value, which is the small municipality of Zerba in the province of Piacenza, which has a population of 92 inhabitants alongside 34 registered employees. However, according to the production data, a total of 4002 MW of power is installed within the municipality boundaries, which rightfully leads to a high value for the KPI.

The final step was the categorization of the critical municipalities based on the KPI assessment into three distinct clusters, and the approach proposed in Section 3 showed effectiveness, as traditional clustering approaches may not work because due to the opposing effects of the two clustering features, which are the production and consumption, the concept of closeness does not have the same meaning. First, Figure 11 represents a scatter plot of the critical municipalities, alongside the conceptualization behind their classification and the different criticality zones. In this case, points that are very far may actually exhibit similar behavior, for example, a large municipality with high production and a small municipality with limited generation. Municipalities with high consumption relative to production belong in the green zone, as in that case, and further production would be preferable. On the other hand, cases with high power installed but relatively low consumption will belong in the red zone, as further installations of generation could be detrimental to the electric grid. The definition of the criticality zones was performed with the categorization procedure detailed in Section 3, with the three-piece linear approximation of the KPI’s CDF. The outcome of this linearization procedure, as well as the optimal break points that determine the boundary between the different criticality zones, or the slopes of the linear curves that are dividing the zones on the scatter plot in Figure 11, is presented in Figure 12.

The output of the categorization procedure puts all critical municipalities with a KPI higher than 7.78 as very critical, whereas the low criticality is assigned to those with a value lower than 1.29; everything in between is classified as medium criticality. According to this clustering, 1423 municipalities are assigned low criticality, 1597 medium, and 503 high criticality, with 4569 being classified as non-critical as an outcome of the analysis of conventional areas. The final municipal criticality map is shown in Figure 13.

5. Practical Implications and Legislative Guidance

The outcome of this analysis is highly valuable to initiate discussions among national regulatory authorities regarding potential revisions in the regulatory framework revolving around ECs and their incentivization to avoid detrimental impacts on the electric infrastructure, which should be of utmost importance. The main aspect of energy communities is for citizens to become more aware of their energy consumption and change their habits in order to increase energy-sharing among them, something that is eventually rewarded with financial benefits. Of course, the benefits of increased local energy-sharing are many-fold, some of which reduce the utilization of electric grids, decreasing energy losses and reducing the need for further upgrades. However, the outcome of this paper shows that this concept is not so straightforward and that it should not be taken for granted that promoting an EC will certainly be grid-positive, i.e., lead to an increase in the local self-consumption because Italy already has an abundance of distributed generation and, in some cases, the effect can be the exact opposite. This is why a case-based regulatory differentiation is a must to ensure that the effects of newly formed ECs are positive across the country. Having proposed an outcome with a municipal resolution only aids the relevant policymaker in introducing the aforementioned differentiation. The main focus should be on the incentives and on the requirement to install new production in order to join an energy community.

For example, in the non-critical white areas, incentives may be higher and the requirement for new production units should be held, as it would certainly bring local benefits. As we move upwards on the curve of criticality, incentives can be reduced and the requirement for new-generation units should decrease. Arriving at the highest criticality (red municipalities), the requirement for new production units should certainly be omitted as it is clear that more than enough local production is already in place. Moreover, in this case, the production is much higher than the local load, leading to the notion that it is difficult to have a severe improvement in the local self-consumption profile even with the formation of energy communities, which is why economically speaking, in this case, the incentives should be lower due to the lack of real benefits. With that being said, the question of incentive differentiation is not solely technical in nature, but deeply rooted in socio-political considerations, requiring thorough investigation from a policy and governance perspective.

It is essential to emphasize that the proposed KPI and the categorized map of criticality assessment aim to propose a viable approach to identify municipalities where the implementation of an energy community could have a significant impact on the electric grid, potentially leading to increased voltage violations, overloading, and, ultimately, higher energy losses. Clearly, the proposed KPI should be approached as an estimation of the risk associated with these drawbacks, not as a technically accurate measure of the problem. Such an analysis could be performed only by the DSO. The final goal is to assist policymakers in properly addressing the issue and, if necessary, improving the regulatory framework. Similarly, the same information could support citizens and private companies in understanding how to improve the energy ecosystem in their area.

6. Conclusions

The study presented in this paper conducts a national-scale evaluation of the current state of reverse power flow occurrences and energy balance, with the goal of assessing the potential impact of wide-scale EC adoption on the distribution grids within the Italian framework, with a municipal resolution. Rather than focusing on the potential impact of ECs on specific distribution networks as studies performed in the literature [22,23], the aim of this paper is to perform a large-scale assessment in order to provide appropriate legislative guidance. However, it should be noted that the notion hypothesized in this paper, that a high local production-to-consumption ratio leads to potential issues with the EC adoption from the aspect of the electric grid is a concept concluded by the relevant literature [16,20,22,24]. Interestingly, the comparative analysis shown in [22] evaluated two distribution grids with significantly different local production levels. The findings revealed that further EC adoption with a high amount of local production leads to worsening operating conditions, demonstrating the relevance of the analysis and approach proposed in this paper. Due to the socio-political nature of energy communities and the aim to provide regulatory guidance, the analysis was conducted on the municipal scale. First, the approach proposed utilizes information on primary substation service areas and a discrete categorization of the reverse power flow in terms of hours in the year in order to provide an assessment of municipalities that should be considered critical. Considering municipalities whose territory falls at least 20% within a critical service area, 3523 municipalities were selected for further analysis, whereas the remaining 4569 were deemed non-critical. Approaching open-source datasets made available by the National Institute of Statistics and the transmission system operator, the evaluation of the municipal annual consumption was performed alongside analysis of the current installation of generators with a rated power equal to or lower than 6 MW. These two datasets were the source to evaluate a KPI defined as the ratio between municipal production and consumption, hypothesizing that this has a direct effect on how preferential the installation of new production units is locally for the adoption within energy communities. Finally, a categorization was performed using a three-piece linear approximation of the CDF curve of the KPI. Following the selection of the optimal breaking points, the three linear segments were created, assigning the traffic-light-based categorization to the municipalities based on the KPI value. The outcome was that 40.4% of the critical municipalities were categorized as relatively low, 45.3% as medium, and 14.3% as high criticality. It should be noted that this categorization is only relative and cannot provide information on the absolute criticality, as in this case there are other factors in play that cannot be grasped with the limited open-source data investigated within this paper. With that being said, it is clear that this analysis aims to utilize open-source information as much as possible and provide a first reasonable estimate of the current energy balance on the distribution grids with a municipal resolution. Interestingly, the main limitation of the paper proposed, which is its reliance on generalized publicly available data and the difficulty in validating the results, is what makes it valuable and intriguing in the first place, as it raises the issue with the current legislative framework and proposes pathways on performing analysis that could enhance it. This would lead to an increase in the country-scale benefits of local new EC adoption. Certainly, given that production and consumption patterns are seen as sensitive data, a more accurate analysis could be performed only in a centralized manner with the relevant national stakeholders.