1. Introduction
The decarbonization of the building stock is an important step on the path towards a climate-neutral society. According to the International Energy Agency [
1], buildings are responsible for 28% of global CO
2 emissions for energy generation, and, in Europe, residential buildings represent the second-most energy-intensive sector with a 26% share of final energy consumption. The main driver for the reduction of buildings CO
2 emissions is the combination of energy efficiency measures with the electrification of building energy consumption [
2,
3,
4].
The widely regarded most effective approach to pursue the reduction of CO
2 emissions by electrifying the energy demand of buildings is to combine heat pump-driven systems with photovoltaic (PV) panels [
5]. However, there is often a time mismatch between power generation and consumption that could lead to an overload of the electric grid, with consequences on the stability and quality of the service. Moreover, the time mismatch could result in buildings still importing the largest fraction of their energy demand from the grid, with no guarantee of it being renewable. To overcome these issues, strategies must be implemented to maximize the self-consumption (
SC) of onsite generated energy. According to Luthander et al. [
6], there are two ways to increase SC: energy storage (e.g., batteries or water tanks) and demand-side management (DSM) (e.g., peak shaving and appliance scheduling). Energy storage makes part of the unconsumed generation available when there is no generation or during peak demand. On the other hand, the idea behind DSM is to adapt the load to the supply. Both strategies lead to valuable improvements in SC, with DSM being less effective, as reported by J. Widén (2014) [
7].
Lopes et al. [
8] have suggested also including the sharing of the produced energy among neighboring buildings in the SC maximization strategies. A neighborhood in which the locally generated energy is shared is known as an energy community. The energy community concept is not new and there are several examples in Europe [
9], some of which can be dated back to the beginning of the previous century in response to electricity poverty of unindustrialized areas, such as mountain areas [
10]. Energy communities could lead to the achievement of net-zero energy neighborhoods rather than buildings by increasing the self-consumption. Marique et al. [
11] and Mittal et al. [
12] have highlighted the opportunities of achieving zero-energy consumption by developing a framework for net-zero energy neighborhoods. In an energy consumption scenario analysis involving typical central European neighborhood typologies, Nematchoua et al. (2021) have reported a 90% reduction of actual energy consumption at the neighborhood scale. However, as presented by Abbà et al. [
13], the simplest energy community is a multi-family house in which energy sharing among the apartments is permitted.
The European Union recently enacted the European Directives EU 2018/2001 and EU 2019/944 [
14,
15], with which the energy communities become formally recognized. Italy has started an experimental period, which sparked the realization of energy communities, some of which have been the object of research studies (e.g., [
16,
17,
18]). In Viti et al. [
19], energy communities, in comparison to individual consumers, were demonstrated to achieve not only higher
SC but also higher profitability. Likewise, Fina et al. [
20] have shown that the installation of PV-systems in energy communities brings economic benefits. Some studies have compared different SC maximization strategies, concluding that energy communities offer great potential [
18,
21]. An early study by Baetens et al. (2012) [
22], regarding the electrical limitations to self-consumption in single-family houses, reported an increase in self-consumption of 7% at the neighborhood level.
Although the number of studies on the benefits of energy communities to
SC is rapidly growing [
23], there is no common definition of energy communities in the scientific literature [
24] and most of the studies often regard specific cases, consisting mainly of single-family houses. In addition, as highlighted by Caramizaru and Uihlein [
25], more research is necessary to clarify and quantify the energy community potential benefits for supporting the EU’s climate and energy goals. Among the main obstacles, energy communities face technological issues during implementation [
23] and the demand mismatch is one of these [
26].
To address these aspects and to generalize the results of the benefits of energy communities on the
SC, this work performs a preliminary general simulation study, focusing on a generic energy community composed of multi-family buildings equipped with heat pump-driven heating systems and with photovoltaic panels. The buildings are representative of the building typologies of Italian heating-dominated climates. Together with the formation of the community, other
SC maximization strategies are studied to provide an order of magnitude of the achievable levels of
SC when each or a combination of these strategies is implemented. Specifically, a control algorithm to maximize self-consumption by storing thermal energy (developed by Pinamonti et al. [
27]) and appliance load shifting are considered.
2. Materials and Methods
The impact on the SC of forming energy communities is evaluated focusing on typical Italian multi-family buildings in mountain and heating-dominated climates. Two kinds of energy communities can be recognized when multi-family buildings are considered: either a multi-family house itself acts as an energy community if the sharing of energy occurs among its apartments, or a set of multi-family houses can form an energy community. This work first analyzes the benefits of turning a multi-family house into an energy community, then it broadens the perspective considering energy communities at the neighborhood level. The simulation studies were performed in TRNSYS 2018 with a one-minute time-step and the data analysis was performed in MATLAB.
2.1. Case-Study Buildings
The buildings’ geometry characteristics and thermal properties were retrieved from [
28], which characterizes the main features of Italian residential buildings. This report categorizes the building according to their location, size, and year of construction. However, the construction characteristics are like to much of the European context as demonstrated by the TABULA project [
29].
The selected buildings were respectively the medium condominium (MC) and the large condominium (LC) located in the Italian climatic zone “E” i.e., the zone with heating degree days from 2101 d to 3000 d. According to the report, the average number of floors and apartments in MC are respectively 3 and 12, whereas LC is generally made of 36 apartments subdivided into 6 floors.
The thermal characteristics were assigned according to the age of the building, considering four envelope quality classes spanning from 1976 to today: class V5 from 1976 to 1990, class V6 from 1990 to 2005, class V7 from 2005 to 2015, and VR after 2015. Class VR also includes all the buildings that were renovated in compliance with the European Energy Performance of Building directive (2010) [
30], transposed into the Italian decree 26 June 2015 [
31].
The thermal zoning of the buildings subdivided the volume based on the floors and the orientation, splitting the building along the west to the east axis, which resulted in 6 and 12 zones respectively for MC and LC.
Additionally, the single apartments were simulated to evaluate the energy performance of multi-family buildings where each apartment is provided with its heating system. This case represents the actual state of most multi-family houses in Italy. The apartment model was obtained considering the multi-family house volume equally split and simulated by assigning the adiabatic conditions to the adjacent walls of the apartments.
Table 1 presents the geometrical characteristics of the two building typologies considered. Only the glazed surface orientation is reported since the solar gains through walls are negligible.
Table 2 presents the thermal properties of the envelope classes.
2.2. Heating System
Each building and apartment is equipped with the same heating system configuration, which is shown in
Figure 1 and consists of an inverter-driven heat pump (HP), a buffer storage (BS) for space heating (SH), and thermal energy storage (TES) for domestic hot water (DHW) preparation.
The DHW is prepared instantaneously via a submerged heat exchanger. Both storage tanks are provided with a double-stage auxiliary electric heater. The terminal units are modeled by
Type 1231 of the TESS library, which models the heat transfer according to the equation
where
U is the heat transfer coefficient of the heating device,
A is its area, and
the logarithmic mean temperature of the heating fluid and air temperatures. The exponent
n is varied to model either radiators (
) or radiant panels (
), depending on the envelope quality class of the building (i.e., radiators for V5 and V6, radiant panels for V7 and VR). The terminal units are fed by the same amount of flow rate, which is bypassed when the temperature set point is reached within the thermal zone. A mixing valve downstream the buffer storage controls the flow temperature to the terminal units, which is particularly useful when not all zones have reached the set point or the buffer storage is loaded at a higher temperature, as explained in the following section.
The buildings are also provided with a PV system installed on the roof. Three levels of roof covering are considered: 50%, 75%, and 100%. The system is made of panels with peak power at nominal operating conditions of 140 /, and it assumed to be operated always at the maximum power point, neglecting the losses of the inverter.
2.3. Control Strategy
The control of the system is split into two parts: one controlling the primary loop i.e., the HP, and the other controlling the secondary loop i.e., the terminal units. The first part controls whether the HP needs to be turned on or off according to the temperature level of the storage tanks. As usual, whenever the temperature probe within the TES measures a temperature below the set point of 50 , the HP is switched to DHW mode, which consists of running the HP at maximum capacity. The HP in SH mode is controlled based on the BS set point determined by the outdoor temperature reset. Additionally, the working capacity is controlled proportionally to the temperature difference of the BS probe and the set point. The secondary circulation pump is turned on whenever a zone thermostat measures an air temperature lower than the set point. The thermostat implements an on-off control strategy with a centered dead-band of 2 .
To maximize the
SC of PV generation, the controller adopts a rule-based control (RBC) strategy for inverter-driven air-to-water heat pumps developed by Pinamonti et al. (2020) [
27]. This control algorithm, shown in the figure, consists of controlling the frequency of the HP such that the load matches the generation. To allow the HP to work, the excess thermal energy is stored in the TES by raising its temperature set point to 60
, and in the thermal mass of the building by increasing the indoor temperature set point of 2
. When the TES is fully charged, the maximum value of the outdoor temperature reset curve is raised by 10 K to allow for energy-storing within the BS as well.
Figure 2 shows the control algorithm flow chart. A comprehensive explanation of the RBC strategy is available in Pinamonti et al. [
27].
2.4. Appliance Load Profile
The complete characterization of the electric energy consumption requires the definition of a suitable appliance load profile. For this purpose, the database of monitored annual appliance consumption for residential buildings with a one-minute resolution provided by Tjaden et al. (2015) [
32] was used to generate the load profile. The power consumption profiles in the database were averaged and normalized by the annual energy consumption. Then, in the simulation studies performed for this work, the normalized profile was multiplied by the appliance and lighting energy consumption set to a value of 4
/a per apartment.
Figure 3 shows the obtained hourly averaged annual power profile for an apartment. It is interesting to note the decrease of appliances consumption during the summer months, which adds to electric generation-production mismatch of PV panels.
The choice of an average, yet realistic, appliance consumption profile is justified by the purpose of the study. Indeed, the heterogeneous appliance profiles that could be obtained by using stochastic models could lead to better matching of load and PV powers. Thus, a homogeneous profile, representative of average residential consumption, was preferred to provide a lower bound of the achievable levels of SC in energy communities.
To evaluate the SC in presence of demand-side management, the appliance power profile was managed assuming to concentrate the power consumption in the central hours of the day, which represents the behavior of smart household appliance connected, for instance, to an energy management system. A baseline consumption of 600
is subtracted from the profile, in order to identify the peak load. This load is then equally distributed between 10 AM and 3 PM throughout the year. The results of the consumption profile rearrangement for a day in winter are shown in
Figure 4.
2.5. Climatic Conditions
The lowest annual
SC values are expected to be obtained in heating-dominated climates because of the mismatch between peak energy demand of buildings (i.e., in winter) and PV supply (i.e., in summer). Northern Italy encompasses different climatic conditions, ranging from alpine to coastline weathers but, overall, the climate can be considered as heating-dominated. The cities selected as reference climatic conditions were Trento (TN), Belluno (BL), and Padova (PD), which span different levels of heating degree days. According to Koppen classification they are in class
,
, and
, respectively. Moreover, the study included the climate of Strasbourg (SXB) as a reference average European climate classified as
according to Koppen (
Figure 5).
Table 3 reports the annual average air temperature, the design temperature, and the heating degree days of these locations.
2.6. The Energy Community
This study focuses on two types of energy communities:
Technically, the first kind of community in Italy is known as collective self-consumers. However, for the purpose of the study, there is not a substantial difference between the two options and in the following both will be regarded to as energy communities. The first is also the simplest to implement and could offer greater potential for rationalizing the energy flows within multi-family buildings. Indeed, in Italy, the state-of-the-art in multi-family houses is the decentralized heating system, i.e., each apartment is provided with its heating system, although the best option would be to centralize the heating system.
The neighborhood in this study is built considering all the possible combinations of the analyzed buildings—MC and LC—and envelope qualities, such that the installed PV power does not surpass a threshold value, but neglecting combinations that are too homogeneous (e.g., made of the same building types). The threshold is set according to the Italian law transposing the European directive that, during the experimental phase, is set to 200
. The neighborhood was built considering covering 75% of the available roof surface of the buildings and then the 100% and 50% covering of the same neighborhood were assessed.
Table 4 reports the composition of the neighborhood together with the number of resulting envelope combinations and the PV power installed for the different roof-covering fractions.
These energy communities were further analyzed considering implementing the previously presented SC maximization strategies. Therefore, the obtained scenarios were:
Scenario 1: no implementation of control or load shifting to maximize SC.
Scenario 2: RBC strategy presented in the section is implemented.
Scenario 3: load-shifting presented in the section is implemented.
Scenario 4: both RBC and load-shifting are implemented.
These scenarios were all compared to the baseline case, represented by the same type of community (same buildings and envelope qualities) not sharing energy among them and without implementing any maximization strategy.
2.7. Key Performance Indicators
The analysis mainly focuses on net-energy consumption, which is the total energy imported by the buildings during a year, and self-consumption metrics. In the literature there are several metrics quantifying self-consumption. However, as reported in Salom et al. (2011) [
33], these metrics can be grouped into two major groups: load-matching and grid interaction indicators. The first group describes the overlap of generation and load profiles, whereas grid interaction targets the unmatched part of these profiles. This study is based on load-matching indicators, specifically, it relies on the metrics adopted also by Luthander et al. (2015) [
6]. These are the self-consumption (
SC), which is the fraction of generated energy that is self-consumed, and self-sufficiency (
SS), which is the fraction of total energy consumption covered by the onsite generation. These are computed as
where,
is the total power consumption,
is the power generation, and
is the self-consumed energy, i.e.,
In the following, the results report annual
SC,
SS, and net-energy consumption, i.e.,
. By knowing these three quantities, the other two can be determined by
As shown by Luthander et al. [
34], particularly meaningful is the ratio of
SC to
SS, which is
Thus, if both energy generation and total consumption remain unchanged, every measure improving self-consumption will have the same ratio between
SC and
SS. On an
SC-
SS chart—known as energy matching chart [
34]—this means that a building with a given
SC and
SS, after the implementation of, say, load-shifting, all the achievable levels of
SC and
SS lies on a line with slope
, unless the load or the generation are changed. The bisector of the quadrant—
—represents net-zero energy buildings. Therefore, buildings falling on the left of this line would be net-producer, or plus-energy buildings, whereas those on the right would be net-consumer.
Ideally, both indicators should be maximized, such that the maximum possible amount of generated energy is locally consumed (i.e., maximum SC), and this covers the largest possible fraction of total consumption (i.e., maximum SS). Depending on the goal, SC should be higher than SS when the aim is to reduce the burden on the grid, whereas larger self-sufficiency signifies less dependence on the grid.
4. Discussion
The analysis performed in this paper allows assessing the benefits to self-consumption of forming energy communities together with other maximization strategies.
The simplest energy community is represented by the multi-family building. The results presented in
Section 3.1 show a marginal improvement in net-energy consumption, particularly for buildings with good envelope qualities. Nevertheless, the energy generation is always rationalized, leading to improvements in both
SC and
SS. Furthermore, according to this study, larger buildings with centralized heating system benefits more than medium-sized ones. As a result, switching to a centralized heating system and allowing for the sharing of energy generation might be particularly effective to reduce the consumption of old large condominiums.
An energy community at the neighborhood scale is proven to reduce the energy consumption with respect to the baseline case, albeit by a small amount, especially in locations with lower solar energy availability. However, the benefits are evident when the SC and SS are considered. Neighborhoods of buildings with low envelope quality could experience the largest improvement in SC and SS. The implementation of maximization strategies together with the formation of an energy community leads to further improvements in all the metrics considered. Despite the somehow small improvement in net-energy consumption, the exploitation of the onsite energy generation can be improved up to 15% when all the strategies are applied together, leading to an increase in the self-sufficiency of up to 10%. Communities located in climates with lower solar availability—such as Strasbourg—present lower improvements, specifically when only the energy community is formed and no other self-consumption strategies are implemented.
The formation of an energy community alone leads to a marginal reduction of energy needs and an improvement in SC and SS metrics that varies between 5% and 3% respectively, both in multi-family houses and in neighborhoods. Nevertheless, the study focused on heating-dominated climates, which is the worst scenario for analyzing the self-consumption of PV energy. Moreover, the standard and invariable appliance load profile makes the energy demand of the buildings or the apartments less heterogeneous, which is also a factor that penalizes self-consumption. Additionally, considering that there is room for system sizing optimization (e.g., PV power based on real needs, storage volumes, etc.), the results presented here can be regarded as the lowest bound of improvements achievable by energy communities. The implementation of DSM and RBC for maximizing self-consumption in energy communities leads to larger improvements. Therefore, energy communities can be considered as an interesting solution for reducing the burden of distributed energy generation on the electricity grid, but it is not the ultimate solution to the problem, and it should be considered together with other strategies.
In this study, electric storage was not considered. From the technical point of view, there is no doubt about its benefits, whereas its contribution should be analyzed from an economical perspective. In general, an approach integrating different technologies and strategies seems to be the right choice, as also presented by Battaglia et al. [
5].
Although in a not-so-far future electric vehicles (EVs) will represent a significant load for buildings, considering them in the computation of the achievable levels of self-consumption in energy communities of residential buildings might not be appropriate. Indeed, considering the standard usage of cars, it is unlikely that several vehicles would be connected to residential buildings during the daytime. On the other hand, the load of EVs becomes relevant in energy communities comprising residential office and commercial buildings.
5. Conclusions
This paper contributes to the discussion on the benefits of energy communities to the self-consumption of on-site generated energy from PV panels. The focus is on energy communities located in heating-dominated climates, which represent the most difficult situation for the maximization of PV energy self-consumption. To enable the largest share of self-consumption the heating system must be electric-driven, hence heat pump-based heating systems. The study focuses on air-to-water heat pumps.
The obtained results quantify the contribution to the self-consumption of energy communities, demand-side management, and RBC control. In multi-family buildings with decentralized heating systems, switching to a centralized heating system and enabling the share of on-site generated energy improves the
SC and
SS by almost 5% and 3% respectively. The formation of such a community was particularly effective in poorly insulated large condominiums, enabling also an energy savings of 10%. These results could be further improved by applying RBC, DSM, or both strategies. In general, an energy community at the multi-family building level ensures better levels of
SC and
SS that—considering also the financial convenience demonstrated by Abbà et al. (2021) [
13]—makes this kind of community attractive and easy to implement.
Communities at the neighborhood level ensure an increase in self-consumption up to 5% at the Italian latitudes. Locations with lower solar availability—such as Strasbourg—are affected by lower energy generation. Combining the formation of the energy communities with DSM and RBC the SC can be improved by up to 15% and the SS by up to 10%. The energy needs can also be reduced up to 12%.
These results are intended to quantify the lower bound of the achievable self-consumption in energy communities. In an increasingly complex energy system and in the rush towards climate neutrality, there is not a single solution and several alternatives should be combined. According to the results obtained, energy communities are to be considered among the solutions for rationalizing the energy consumption of buildings. By aggregating individual energy demands, communities can offer local flexibility services such as relieving network congestion and avoiding peak withdrawal from the electricity grid. The results of this study can assist local governments in understanding the achievable performance of new energy communities, developing supportive policies, and promoting them to end users.
In future works, commercial and office buildings should be included in the analysis and the presence of electric energy storage should also be considered.