1. Introduction
Energy vulnerability is considered as the susceptibility of a household to energy poverty, resulting from changes in the energy system. Energy poverty may be defined as a household’s lack of access to essential energy services, where such services provide basic levels and decent standards of living and health, including adequate heating, hot water, cooling, lighting, and energy to power appliances, in the relevant national context, existing national social policy and other relevant national policies, caused by a combination of factors, including at least non-affordability, insufficient disposable income, high energy expenditure and poor energy efficiency of homes [
1]. Considering energy poverty indicators such as the inability to keep homes adequately warm, the delay in the payment of utility bills, and the occupation of defective dwellings, Castano-Rose et al. [
2] have estimated that, in Europe, about 10% of the population live in energy poverty conditions, unable to warm their houses during the winter. The consequences of energy poverty are far-reaching, encompassing profound impacts on health, economic stability, and social cohesion [
3]. For many households, the inability to afford adequate energy services exacerbates existing inequalities. Failure to address energy poverty not only perpetuates social injustice but also undermines efforts toward sustainable urban development and climate change mitigation.
In Europe, there are several examples of public interventions to reduce energy poverty. For example, the REELIH project helps create and develop an investment market for building rehabilitation of low-income homeowners living in multi-apartment buildings in Central and Eastern Europe to secure financial and political support from governments placing greater emphasis on improving existing buildings by working with local communities [
4]. EnergyMeasures is working to tackle energy poverty in seven European countries through direct engagements with households, complemented and informed by policy innovations. ENPOR is another project involving several EU countries aiming to alleviate energy vulnerability in the private rental sector (PRS) by selecting existing policies based on monitoring and creating customized policies for energy poverty alleviation by 2030 [
5]. The EmpowerMed project aims to contribute to reducing energy poverty and improving the health of people affected by energy poverty in the coastal areas of Mediterranean countries, with a particular focus on women. Other initiatives are more focused at the country level. In Barcelona, the city council organizes and provides audits and interventions in the homes of energy-poor and low-income households. In Lancashire, in the UK, a local program targets energy efficiency for home heating, offering an affordable and direct means of accessing grants from energy companies and other sources to fund new heating measures, insulation and renewable technologies in domestic properties for low-income residents. In Cyprus, the government project will subsidize the implementation of small-scale energy renovations such as thermal insulation, lighting replacement, etc., in 300 disabled and energy-poor households [
6], while in Italy, the Energia su Misura project aims to support vulnerable families living in social housing belonging to local authorities, improve energy consumption and reduce energy costs by reading energy bills and installing smart devices connected to electrical appliances and the central electricity meter [
7]. Romania and Hungary are developing programs which, in addition to other target groups, include households without access to electricity [
6]. In Portugal, 36% of families were living in buildings in need of repair in 2021 [
8]. There are at least two organizations contributing to enhancing energy conditions in houses. One is “Just a Change”, a non-profit association that rebuilds homes for people in need, and LIGAR-Energia para Todos has a team of experts that identify, map, and execute local interventions in at-risk households [
6].
As a broader impact action, the deployment of decentralized rooftop photovoltaic (PV) systems became a viable and sustainable solution to help mitigate energy vulnerability [
9,
10]. Self-consumption PV systems can have a significant impact on the reduction of energy bills since the consumption of locally generated solar power avoids both active energy and grid access costs (at least partially, depending on the specific regulatory framework), with typical payback times below 10 years, for an operational lifetime beyond 20 years [
11]. For example, in Seoul, South Korea, the metropolitan government is funding a large-scale PV deployment project of small PV modules on verandas, targeting one million households, with subsidies amounting to 80% of the system cost, and 100% for low-income families, with expected payback below 2 years despite the less-than-optimal module tilt [
12].
Emerging solar energy communities are expected to play a particularly important role in the alleviation of energy vulnerability. A citizen-led initiative in the city of Getafe (Madrid, Spain) to design and implement a collective PV self-consumption infrastructure on the roofs of various public buildings was shown to be able to supply about 100 flats, reducing the energy bill in energy-vulnerable households by about 15 EUR/month [
13]. In an energy community with seven households in southern France, each with its own PV systems, energy sharing led to a bill reduction of 11.7% [
14]. Positive energy districts (PED), combining local solar energy generation and building energy retrofit at the neighborhood level, have also been explored as a relevant approach for the alleviation of energy poverty [
15].
The role of energy communities may contribute to reducing energy vulnerability, such as described by Hanke and co-authors [
16], which reports on a renewable energy initiative on the Danish Island of Aero that provides zero-interest loans to facilitate membership for vulnerable households. In the Belgian city of Eklo, the municipal government acquires shares in local renewable energy communities, allocating them to energy-disadvantaged households, enabling access to reduced energy costs. In Portugal, the energy cooperative Coopernico extends solar installations to NGOs assisting vulnerable households.
To assess the potential for PV rooftops to contribute to alleviating urban energy vulnerability, this work combines geospatial urban PV potential with measured hourly consumption and energy vulnerability indicators, identifying priority areas for intervention. The method is illustrated for the case study of Lisbon, Portugal. The results depend on the specific local conditions, including the available solar resources, rooftop characteristics, local energy demand, and PV regulatory framework, but the methodology and the general conclusions are expected to be valid for a wide range of urban areas across the globe.
The structure of this paper is as follows.
Section 2 presents the methodology developed, the case study, and the data used, including georeferenced estimates of the solar energy rooftop potential, hourly electricity demand, and energy vulnerability. In
Section 3, the results are presented, identifying the highest priority parishes for PV deployment to mitigate energy vulnerability.
Section 4 discusses how public policies can be developed to target PV deployment for energy poverty mitigation, and the conclusions are presented in
Section 5.
2. Methods
Adequate public policies are critical to the dissemination of PV deployment, and their effectiveness can largely benefit from data-informed methodologies that may quantify the benefits associated with different alternatives customized to specific contexts of climate, buildings, or energy utilization patterns. This paper offers a methodological framework based on geospatial data analysis, as described in
Figure 1. In this framework, energy demand, solar potential, and energy vulnerability are mapped for the area under study. Then, the energy demand and solar potential maps are used to identify the urban areas with higher self-consumption potential. Finally, the self-consumption potential map is combined with the energy vulnerability assessment to identify the priority urban areas.
Once the areas with potential for self-consumption are identified, the next step involves integrating the self-consumption potential map with an energy vulnerability assessment. By combining these two datasets, priority urban areas can be determined. The energy vulnerability assessment allows for considering additional factors, such as economic or social disparities that may influence the effectiveness and impact of public policy interventions. In the following sections, the case study is presented, followed by the methodologic details for assessing local energy demand, solar potential, and energy vulnerability.
2.1. Area of Study
The case study is developed for the city of Lisbon, in Portugal. The city is divided into 24 administrative regions, known as parishes or civil parishes. The total resident population is about 540 thousand inhabitants, with a significant population density inhomogeneity, ranging between 10 and 45 thousand people per parish with a median of about 20 thousand people, as shown in
Figure 2.
2.2. Energy Demand
Electricity demand data, corresponding to spatially-aggregated data from transformer substations, by parish, were provided by E-REDES, the Portuguese distribution system operator (DSO). Data aggregation is required by data privacy compliance issues since the DSO is only allowed to share demand data from substations that comprise a minimum number of consumption points as well as to allow comparison with the energy vulnerability analysis described in
Section 2.4 since the data is only available at the parish level. The consumption data used in this analysis corresponds to hourly data obtained in 2019, to avoid the influence of the pandemic with exceptional load features, providing a more accurate representation of typical energy consumption patterns.
2.3. Solar Potential
The hourly PV potential for the building rooftops was estimated by r.sun (considering clear sky) using hourly solar irradiance estimates for each building [
17], for each characteristic day [
18] of the year. The solar irradiance is determined by the r.sun algorithm [
19], integrated with GRASS software, based on a high-resolution LiDAR-based digital surface model considering rooftop tilt and mutual shading between buildings and their rooftop structures. The PV generation of the rooftop assumes an overall conversion efficiency of 21%, a performance ratio of 80%, and that 80% of the rooftop area can be used for the PV system.
Figure 3 illustrates the solar potential output.
The assessment of the self-consumption potential was mainly addressed by estimating self-consumption and self-sufficiency coefficients. As usual, the self-consumption coefficient is defined as the ratio between the portion of the PV generation consumed locally and the total generation and, therefore, it is a measurement of the locally consumed solar energy. As elsewhere, the self-consumption regulation in Portugal determines that excess generation is fed into the grid at a significantly lower price than the cost of consumption of electricity from the grid [
21]. This arrangement provides a strong incentive for local consumption and PV systems are often sized to minimize excess generation for maximum profitability. The self-consumption coefficient is thus a proxy measure of the profitability or returns on investment of the PV system since a high self-consumption rate describes a system where there is a minimal amount of excess energy curtailed or sold to the grid and all PV generation is highly valued.
The self-sufficiency coefficient, on the other hand, measures the fraction of total demand satisfied by local PV generation, being defined by the ratio between the amount of electricity locally generated and consumed and the total consumption. It is thus a measurement of the avoided consumption from the grid. Since PV-generated electricity is significantly less costly than electricity bought from the grid, the self-sufficiency coefficient is a proxy measure of the savings due to the locally generated solar energy.
2.4. Energy Vulnerability
The assessment of energy vulnerability across various parishes was derived from the methodology developed by Gouveia and Palma [
22]. Their approach involved constructing an energy vulnerability index, which takes into account several building characteristics such as construction period, building form, number of floors, and roof type. This index enables the identification of the proportion of the population susceptible to energy poverty and assesses the likelihood of encountering energy-poor consumers within a given area.
The indexes for heating and cooling needs during winter and summer, respectively, were determined through an analysis of the energy gap between a building’s final energy demand and its actual consumption. These sub-indexes also incorporate various socio-economic factors such as unemployment rates, educational levels, and the overall condition of buildings.
The calculation of the energy needs takes into account key construction characteristics including wall type, thickness, insulation, and roof materials. By establishing optimal indoor temperatures for both seasons, the energy demand for space heating and cooling could be quantified across the different climatic conditions.
3. Results
In this section, maps and results regarding energy demand, solar potential and energy vulnerability are presented and discussed.
3.1. Energy Demand
The annual energy demand per capita and parish is shown in
Figure 4. We can observe that there is a relevant geospatial variability in the annual demand, with a few parishes in the city center with average per capita consumption significantly above the city average, which is attributed to electricity demand associated with non-domestic consumption.
The detailed hourly demand is shown in
Figure 5, for the city as a whole (all parishes) and two very different parishes: Santa Maria Maior and Santa Clara.
One may notice that the overall demand profile is strongly influenced by the demand in the high-consumption parishes, with strong demand during working hours throughout the year. Santa Maria Maior (bottom plot) has the highest number of non-domestic consumers, both in high and medium voltage. It is relevant to point out that Santa Maria Maior is in the historic city center, witnessing a decrease in resident population despite an increase in the number of dwellings in recent years. According to the National Statistics Institute, Lisbon has the highest number of guests staying in local accommodations, as well as in hotels, between May and September [
23]. Another interesting fact is that according to the list of establishments registered on the Lisbon municipality’s restaurants and commerce platform, in 2020, the parish of Santa Maria Maior had 154 establishments, and Santa Clara only had 15 [
24].
In the residential area (Santa Clara, middle plot), energy demand is rather low throughout the year, peaking mostly during the winter evenings, presumably for electric heating. These insights already suggest a lack of general match between PV electricity, available during the day and higher in the summer, and the electricity demand of residential customers, mostly at night and in the winter.
3.2. Solar Potential
The average rooftop area per parish can be determined from the digital surface model of the city. It shows a relevant variability (17 to 65 m2/person) with a median of about 30 m2/person. The rooftop area is the critical element for the estimation of the solar potential of the buildings in the parish, although the assessment also takes into consideration the detrimental impact of shading and less-than-optimal inclination and orientation.
The analysis of annual solar potential is shown in
Figure 6. One can observe that, in general, parishes with higher population density feature less rooftop area per inhabitant, hence lower PV generation per capita. Deviations are again associated with less-than-optimal rooftop orientation/inclination and, in most cases, shading from neighboring buildings, which are expected to be more densely packed and/or taller in higher-density areas. Nevertheless, a clear correlation between the number of inhabitants and the number of buildings, average area per building, total area, or average height of buildings was not observed.
3.3. Energy Vulnerability
The energy vulnerability per parish is shown in
Figure 7. We can observe that there is no general correlation between the summer and winter vulnerability indexes, although some parishes feature low vulnerability throughout the year (Lumiar, Carnide or Parque das Nações) while others are always vulnerable (Beato, Ajuda or São Vincente).
The energy vulnerability during the winter is anti-correlated with the local energy demand. As shown in
Figure 8, the most vulnerable parishes during the heating season have low per capita energy demand, suggesting that local inhabitants do not consume electricity for heating during the cold season. For the cooling season, we can observe that vulnerable areas have a wide range of potential solar options; the parishes Misericórdia and Santa Maria Maior feature high solar potential and energy vulnerability in summer.
It is also interesting to analyze the relationship between the energy vulnerability index as a function of the solar self-consumption and self-sufficiency ratios. The distribution of the self-consumption and the self-sufficiency ratios as a function of the energy vulnerability is shown in
Figure 9 for both the heating and the cooling seasons. It shows that the variation in the self-consumption index is quite small around the whole urban area (in the 30 to 40% range). The impact on the energy bill, on the other hand, can vary widely, from about 20 to 70%, depending on the parish and the season.
In general, in winter, the self-sufficiency ratio is lower for the most energy-vulnerable parishes. In the summer, however, we can identify more vulnerable areas, such as Beato and Ajuda, whose PV impact on the energy bills is lower, while there are other parishes, such as Santa Maria Maior or São Domingos de Benfica, where the impact of PV systems could be very significant, with a reduction greater than 50% on the energy bill.
4. Discussion
In the previous section, it was shown that some of the parishes, in particular Santa Maria Maior and São Domingos de Benfica, feature high vulnerability indexes and high self-sufficiency coefficients and thus have the potential for the high impact of the PV systems on the reduction of the energy bill. They feature a relevant self-consumption coefficient, of the order of 35–40%, which warrants, with current PV system prices, a payback period of about 3–4 years. The most vulnerable households in these parishes would thus be expected to strongly benefit from investment in rooftop PV systems.
4.1. Public Policy Implications
Assuming a strong correlation between energy vulnerability and limited individual capacity to invest in PV systems, it is arguably the role of public policies to help provide the means or contribute towards the installation of PV systems. This would simultaneously address the challenges of energy vulnerability and transition to a cleaner urban energy system. The natural question does arise: how could public policy support the deployment of PV systems in the most vulnerable parishes, such as Santa Maria Maior or São Domingos de Benfica?
One option is the implementation of direct support mechanisms, such as tax benefits or subsidies, for individuals willing to install PV systems. However, empirical evidence suggests their limited effectiveness and potential for exacerbating inequity [
25]. Higher-income communities tend to benefit from such incentives, as they possess the resources and incentives for initial investments. Additionally, renting is prevalent in energy-vulnerable areas and neither landlords nor tenants may find adequate motivation to invest in PV systems due to tenure constraints and the relatively prolonged return on investment.
Another important limitation of direct support mechanisms is the mismatch between PV supply and electricity demand, the latter being higher in winter evenings when there is no PV generation. This mismatch can be mitigated by storage solutions or by combining different demand profiles as in a solar community, with residential and non-residential consumption. Examples of these diverse solar communities could include a neighborhood with residential consumers and office spaces, shops, social services, or a school, with higher consumption during the day [
26].
Promoting solar energy communities requires tailored incentive mechanisms, regulatory simplification, and municipal facilitation. These could be accompanied by establishing decentralized local offices to support community development and disseminate information, possibly coordinating with local schools, to promote engagement and foster energy literacy, essential for the promotion of sustainable behaviors.
Encouraging third-party participation in energy communities can mitigate the capital constraints faced by individuals. Facilitating the development of PV systems on a collective self-consumption basis, wherein roofs are rented and electricity is sold to local residents at reduced prices, is an interesting approach. The city council itself can play a pivotal role by installing PV systems on social housing and/or municipal buildings, such as schools, to supply the neighboring population at subsidized rates.
It should be noted that, in the particular case of Lisbon as elsewhere, the public promotion of solar energy communities and PV rooftops in general needs to take into account the potential conflict between solar energy and protected heritage buildings, demanding suitable strategies including site-specific assessments, technological adaptations, and stakeholder consultations to ensure compatibility between solar installations and protected heritage structures [
27]. In this specific case, and according to the 2021 census [
28], the average age of buildings in the parish of S. Domingos de Benfica is relatively recent (46 years old), but the age of buildings in the parish of Santa Maria Maior was 86 years, making it the second oldest parish in town and, therefore, particularly prone to witness this potential conflict between solar generation and protected heritage buildings.
4.2. Method Limitations
The methodological approach of this work is conditioned by the low spatial resolution of the data, due to privacy concerns, hence not allowing for definitive conclusions regarding the impact on the energy bill of PV deployment in specific energy-vulnerable homes. Indeed, there are both significant PV potential and energy vulnerability variability within each parish that require further detailed analysis to assess the real impact of the PV systems on the energy bill of the residents. It should also be noted that the self-consumption rate (hence the return on investment) may be improved if only the most favorable fraction of the roof area is used, a factor that was not considered in the analysis. Of course, in that case, the PV systems would have a lower impact on reducing the bill. Nevertheless, the results help identify priority areas where further detailed assessment should be undertaken.
Finally, one ought to underline that merely providing electricity at a reduced cost does not automatically alleviate energy poverty since lower electricity prices may not always translate into increased demand, for the essential standard of living, particularly heating needs during winter months. Instead, households may redirect the savings towards addressing other pressing needs such as rent, food, or transportation expenses. Thus, while the reduction of electricity costs is a significant factor, broader socio-economic factors also influence the effectiveness of measures aimed at the mitigation of energy poverty.