Integration of PV Systems into the Urban Environment: A Review of Their Effects and Energy Models

Abstract

Building integrated photovoltaics (BIPVs) consist of PV panels that are integrated into a building as part of its construction. This technology has advantages such as the production of electricity without necessitating additional land area. This paper provides a literature review on recent developments in urban building energy modelling, including tools and methods as well as how they can be used to predict the effect of PV systems on building outdoor and indoor environments. It is also intended to provide a critical analysis on how PV systems affect the urban environment, both from an energy and a comfort point of view. The microclimate, namely the urban heat island concept, is introduced and related to the existence of PV systems. It is concluded that urban building energy models (UBEMs) can be effective in studying the performance of PV systems in the urban environment. It allows one to simultaneously predict building energy performance and microclimate effects. However, there is a need to develop new methodologies to overcome the challenges associated with UBEMs, especially those concerning non-geometric data, which lead to a major source of errors, and to find an effective method to predict the effect of PV systems in the urban environment.

1. Introduction

Energy has been the focus of much research due to its importance to the economy, climate change, and political decisions. On one hand, countries want to ensure energy independence and security relative to other nations, competitiveness, and good economic standing [1]. On the other hand, it is necessary to reduce the emissions of carbon dioxide (CO₂) and greenhouse gases to mitigate climate change.

Some solutions allow for the achievement of energy sustainability, which can be defined as “the provision of energy such that it meets the needs of the present without compromising the ability of future generations to meet their needs” [2]. In addition to reducing energy demand, energy sources in the energy mix need to be diversified and energy production from renewable sources needs to be increased. One of the main renewable sources used for energy production is the sun. Solar energy can be converted into thermal or electrical energy [3]. To generate electricity, solar photovoltaics (PVs) can be used. In 2020, solar energy and geothermal heat only represented about 2% of buildings’ share of renewables [4]. This means that is important to increase distributed energy systems, namely, solar PVs. There are still challenges to increase the penetration of solar PVs. For example, cities combine high levels of electricity consumption with limited space, which restricts their access to eligible solar surfaces [5].

Incorporating PV technologies into buildings not only increases electricity production but can also affect energy loads and comfort. These aspects must be considered and optimized to increase sustainability and, therefore, the share of electricity produced from solar PVs.

Cities are responsible for more than 50% of the global population, 80% of global gross domestic product (GDP), two-thirds of global energy consumption, and more than 70% of annual global carbon emissions [6]. Considering that cities are the main drivers of the European Union’s (EU) economy, the EU is promoting policies to make urban areas more sustainable, competitive, and healthier, while tackling climate challenges [7].

As the International Energy Agency (IEA) stated [6], countries cannot meet their climate targets without optimizing building energy efficiency and energy demand. Buildings play an important role in world energy consumption and, consequently, in the environment. In 2022, building operations were responsible for 30% of global final energy consumption and 26% of global energy-related carbon dioxide (CO₂) emissions [8]. In the EU, building energy consumption is more significant. According to the European Commission, buildings account for 43% of final consumption and 36% of CO₂ emissions in the EU [9]. Overall, buildings are responsible for a large share of total energy consumption and CO₂ emissions.

It is also known that people spend most of their time inside buildings, whether at work or at home. Therefore, it is necessary to guarantee adequate indoor conditions to maintain thermal comfort and air quality. In ideal circumstances, this will lead to the use of energy to satisfy the needs of building occupants. However, energy bills are sometimes too high for the end-users, which may lead to energy poverty, which happens when households do not have access to essential energy services [10]. Innovation can reduce the price of energy and thus increase the acceptance of new efficient technologies, which are crucial to alleviating energy poverty and adhering to the energy transition.

1.1. Building Energy Modelling

Building energy simulation tools are important to predict and analyze building energy consumption, CO₂ emissions, and indoor conditions. This is particularly important in order to analyze the impact of new efficient technologies in buildings without having to install them. They focus essentially on heat transfer through envelopes [11].

There are many software tools that are used for building energy simulations. However, some tools are better suited than others, according to the desired objective. In fact, in addition to the experience of the user and hardware availability, the first criterion for selecting a building energy analysis program is the ability of the program to deal with the application [12]. Pan et al. [13] categorize 157 studies into five application scenarios: performance-driven design, operational optimization, building-to-grid interaction, digital twin, and urban modelling. The distribution of the studies across the various application scenarios is presented in Figure 1.

According to VanDerHorn and Mahadevan [14], a digital twin can be defined as “a virtual representation of a physical system (and its associated environment and processes) that is updated through the exchange of information between the physical and virtual systems”. Using this definition, it is possible to state that, in the case of urban building energy modelling, a digital twin will consist of a virtual representation of city buildings that is updated in terms of energy systems through the exchange of information between the physical building and its virtual representation.

To understand digital twins for building energy simulations either for single buildings or at an urban scale, it is important to know some concepts, like building information models (BIMs), building energy models (BEMs), urban building energy models (UBEMs), and the Internet of Things (IoT). The first one, a BIM, is a comprehensive digital representation of a building and typically contains information about its geometry and systems, spaces and zones, and the project structure/schedule [15]. However, a BIM does not contain all the energy information about a building. For that, it is necessary to create a BEM that is a physics-based model of building energy use [16]. Usually, the creation of a BEM starts with a BIM, which adds information about, for example, lights, occupancy, heating, ventilation, air conditioning (HVAC) systems, and renewable systems. A BEM can be simulated by software to analyze the energy performance of buildings. BIMs and BEMs offer high-fidelity projects at the component level [17]. However, the integration of real-time sensing data with the static information provided by BIM models is possible with, for example, the emergence of the IoT [18]. The IoT can be defined as the “interconnection of sensing and actuating devices providing the ability to share information across platforms through a unified framework, developing a common operating picture for enabling innovative applications” [17,18].

The interaction between real-time data, a virtual entity, and a physical entity is indeed an advantage compared to simulations with static and historical data. This allows one not only to evaluate the performance of a building, but also to improve decision making through continuously updated information, monitoring, and predictions of actual and future conditions of buildings, both in the design and operational phases [19].

An UBEM follows an identical concept to a BEM. It means that the physical models of heat and mass flow through the buildings applied to a BEM are the same for a UBEM but applied on a larger scale, to more buildings. Also, a BEM and a UBEM have different objectives. The first one is used to analyze the energy performance of individual buildings to optimize the energy efficiency at the building level during the design and construction phases. A UBEM is used to support decision making for properly designing and optimizing energy systems at the urban level [20]. Due to the larger scale of UBEMs compared to BEMs, simulations require significant automation procedures and computational prowess during data input, model generation, simulation, calibration, and execution [21]. Also, modelling an urban energy system has challenges due to the complexity of urban systems, a lack of data, and the extensive amount of time and modelling efforts needed to achieve accurate urban-scale modelling [22].

1.2. PV Systems in the Urban Environment

Building integrated photovoltaics (BIPVs) consist of PV panels that are integrated into buildings as part of their construction. This technology has advantages such as the production of electricity without necessitating additional land area. Also, BIPVs can be used as exterior protection, for example, windows, which means that one’s cooling needs decrease while there is a simultaneous production of electricity [23]. In general, according to Zhang et al. [24], building surfaces available for the integration of PV systems are facades (Figure 2), roofs (Figure 3), windows and overhead glazing (Figure 4), and sunshades (Figure 5). On the other hand, building attached photovoltaics (BAPVs) consist of PV systems that are attached to existing building envelopes [25], for example, through an installation with brackets onto a roof.

The performance of BIPV systems is affected by several parameters, such as building orientation, panel slope, solar cell type, ambient temperature, geographical location, and shading [29]. In fact, in terms of geographical location, due to a higher amount of solar radiation near the zenith, in tropical regions, BIPVs are mostly installed on roofs [30]. In contrast, in non-tropical regions, the amount of solar radiation near the zenith decreases with latitude, which reduces the electricity production in roof solar PV panels, while favoring electricity production in panels placed on vertical walls.

It is known that in the northern hemisphere in non-tropical regions, the best orientation for a solar panel is towards the south. However, in BIPVs, along with energy performance, the installation of solar PV panels is also dependent on aesthetics and architectural aspects, which means decisions must be made during the design phase [31]. Other factors, such as the daylighting performance and the energy consumption of a building, are fundamental to designing an efficient BIPV system.

The outdoor environment also influences PV system performance. In the same way, PV systems have an impact on the outdoor environment, for example, on the air temperature or building energy use. This idea about how the built environment and PV system performance are related is synthesized in a review by Sailor et al. [32]. It was stated that the air temperature, air pollution, soiling, and shading provoked by the urban environment impact the electricity output of PV systems. In fact, in terms of temperature, the higher the PV surface temperature, the lower its efficiency. Therefore, due to the low thermal mass of PV panels, these surfaces heat faster than their urban surroundings, which can reduce the PV output. In addition, urban heat island (UHI) effects can decrease the efficiency of PV panels. In terms of shading, it has been stated that partial shading from panels’ surroundings or trees is a challenge to improving PV output. On the other hand, concerning the effect of PV panels on air temperature or building energy consumption, Sailor et al. explained that the results of studies are not convergent. In terms of air temperature, although PV panels can warm the air during the day and cool during the night, it is necessary to use realistic values for urban surfaces, PV efficiency, and PV installation characteristics. Also, the air conditioning loads can differ according to the albedo of the surfaces, building insulation, and building characteristics and construction. It was also noted that integrating PV systems into shade structures can be beneficial in hot climates while producing electricity. However, regarding thermal comfort, high-solar-reflectance structures are preferable because they have a lower mean radiant temperature.

The thermal impact of PV panels on the surrounding environment is usually neglected, and often only the electricity production from this type of technology is analyzed. However, the low thermal inertia of these systems compared to urban constructions, the lower reflectance of radiation at higher wavelengths, and shading can influence the air temperature, which is directly related to microclimate effects and the formation of UHIs, consequently affecting the thermal loads of buildings. Furthermore, these types of converters can also influence the comfort of people. In the literature, there are recent studies and reviews that address the impact of microclimate effects, namely the formation of UHIs in urban environments. Mirabi and Davies [33] reviewed the impact of different types of urban infrastructures, such as rails, road networks, and system and utility corridors, on UHI effects. Zhou et al. [34] reviewed the interaction between PV systems in the urban environment through a CFD analysis. Zhu et al. [35] developed a review addressing the impact of urban green infrastructures on microclimate effects and building energy consumption. Susca et al. [36] conducted a review on the relationship between green wall installations, UHI effects, and building energy use.

1.3. Aims and Paper Structure

This paper intends to present a literature review on the effect of PV systems in the urban environment. The energy consumption of buildings and its relationship with PV systems is considered. In addition, the existence of PV systems, how they influence the formation of urban heat islands, and how they influence thermal comfort are important aspects to consider.

In the literature, there are few critical reviews on the effect of photovoltaic systems and urban surfaces in the urban environment, namely, the formation of UHIs, building performance, and outdoor and indoor comfort. Consequently, there is also a lack of methodologies to evaluate the impact of PV systems in the urban environment. Therefore, it is an objective of this paper to provide a systematic review on how it is possible to use urban building energy modelling as a tool to assess the effect of PV systems in the urban environment.

For this, Section 2 provides a literature review about recent developments in urban building energy modelling. The bottom-up approach is explained, as well as the steps for the creation of an urban building energy model. Data acquisition, especially the acquisition of non-geometric data, is suggested as one of the major difficulties in the creation of a precise urban building energy model. UBEM softwares and methodologies to assess microclimate effects are also included.

Section 3 reviews novel studies on how PV systems affect building indoor conditions, specifically energy consumption and thermal needs. The effect of PV systems in outdoor environments is also considered, particularly their effect on the microclimate and the formation of UHIs.

Section 4 presents the main conclusions of the literature review on urban building energy modelling and on the effect of PV systems in urban environments.

2. Urban Building Energy Modelling

UBEMs can be categorized into top-down and bottom-up models. The former rely on aggregated historical data to express the relationship between energy use and drivers like socio-economic variables and climatic conditions. This approach is based on statistical methods that do not require detailed technological descriptions and that are simpler. Nevertheless, this modelling is less suitable for examining changes in technology in current and future development studies.

On the other hand, bottom-up models are made from extensive data on a disaggregated level, which allows one to estimate individual energy consumption and, consequently, the total energy use of a city or district. Based on the level of detail in the end-use information and the applied methodology, bottom-up models can be divided into three main categories: statistical, engineering (physical), and hybrid models [37]. Statistical models use historical data and regression analyses to establish relationships between the energy use of buildings and their characteristics. The second one takes advantage of the physical and technological properties of the individual buildings to estimate the energy consumption of each one. These models have a high degree of flexibility and allow one to evaluate different energy efficiency scenarios. Finally, hybrid models refer to when these two methods are combined, for example, modelling the buildings according to their physical characteristics and estimating the occupancy schedule from statistical data. Nevertheless, Reinhart and Cerezo Davila [38] classified bottom-up engineering models, which include physical models of heat and mass transfer in and around buildings, as “urban building energy modelling”.

According to Figure 5, a UBEM requires four steps before its application [21]: data acquisition, model generation, simulation, and calibration. The first consists of acquiring geometric and non-geometric data, weather data, energy measurements, and building templates. The second is the creation of a geometric model using the geometric data collected. After that, the simulation involves the generation of load profiles, indoor temperatures, and energy use. It is created based on the geometric model developed in the step before, non-geometric data, such as occupancy and equipment schedules, weather data, and building templates. The simulation is calibrated using historical energy measurements to achieve the energy model closest to reality. The energy model created can be used for different applications.

2.1. Data Acquisition

An overview of the acquisition of geometric and non-geometric data is given next, based on the works by Wang et al. [39] and Johari et al. [37]. Starting with geometric data, this type of information is necessary to design a geometric model, such as building footprints and heights, window–wall ratios, the number of stories, and terrain data. This information can be achieved using existing databases, for example, using a shapefile that can be used for a geographic information system (GIS) such as QGIS. Other direct modelling approaches are achieved using Light Detection and Ranging (LiDAR) and oblique photogrammetry. When studying a larger number of buildings, a common practice is to use archetypes to estimate the non-geometric data, because it is difficult to model and acquire data for each building. In this way, the archetypes can be classified into deterministic or probabilistic classifications. Buildings are classified as deterministic according to their typology, age, shape, and floor area. Also, if HVAC system data are available, this could be another indicator to classify them. In comparison to buildings classified as probabilistic, the archetypes are formed based on historical energy demand data, which could increase the accuracy of deterministic classifications. They both show that the major current challenges related to data acquisition rely on non-geometric data, especially data for archetype development, due to the higher associated degree of uncertainty. Access to data to develop archetypes is still a challenge due to data privacy and high costs. In addition, there is a lack of methodologies to determine occupants’ behaviors and energy use profiles.

Abolhassani et al. [40] developed a novel workflow for detailed urban energy modelling, which used an occupancy archetype approach, intending to reduce the uncertainty associated with occupancy behavior. They categorized archetypes based on building use type and the year of construction, obtaining occupancy, lighting, electrical equipment, and ventilation schedules. The construction archetypes follow an identical procedure. It was reported that occupancy behavior has the greatest impact on heating and cooling needs, and its uncertainty can be reduced through more accurate feedback and smart energy management systems.

Sasso et al. [41] follow a common approach and categorize archetypes based on the construction period, type of heating system, location, and urban setting. This work presented a study on the thermal performance of office buildings in Switzerland and could potentially influence future works on thermal optimal retrofits through energy-efficient measures for envelope and heating systems. However, this study does not consider ventilation systems or the accuracy of the results in terms of archetype characterization.

Borges et al. [42] studied a hybrid approach, combining deterministic and cluster methods to characterize archetypes for urban models. As seen in Figure 6, first, the buildings are grouped according to their use. After that, with machine learning, they are analyzed as a cluster considering energy consumption. Finally, fragmentation is performed according to the construction period. This technique showed a granularity of results when compared to separate deterministic or clustering methods: it represented the heterogeneities of buildings more accurately than deterministic methods, while better distinguishing key aspects of buildings compared to cluster approaches. However, the application of this approach is complex and requires more time than a purely deterministic or clustering method. Also, this study does not show the impact of the hybrid approach on UBEM simulation results.

A reference building could be characterized from deterministic methods through the average values obtained after an audit of different buildings of the same archetype. On the other hand, probabilistic methods could be used, in which the values are not obtained from the average but instead from uniform or normal distributions, which could improve the simulation accuracy but increase the simulation time. In addition, buildings could be characterized using not purely deterministic methods, such as the standards of the American Society of Heating, Refrigerating, and Air-Conditioning Engineers (ASHRAE), which presents reference values for each category.

The unreliability of inputs can lead to errors in energy consumption. Occupants’ behavior, namely the occupancy schedule, due to its variability and prediction difficulty, is one of the major sources of uncertainty in building energy consumption and has been studied in recent years [43]. Fu et al. [44] developed a study in China that generated an approach to identify residential buildings’ occupancy patterns and to describe profiles of typical families. Four occupancy patterns for workdays and five for weekends were found, and the average occupation time was 20.3 h per day during weekdays and 20.1 h per day during weekends. By far, the bedroom is the room with the greatest occupation time. This paper was developed considering factors like the characteristics of households, namely occupants’ age, work status, and educational experience. However, although the study was carried out in different geographical locations across China, it did not consider other countries. Another major drawback of this study is the fact that it only includes data from the summer season, from June to September 2017.

Ferrando et al. [45], based on a study on 21 buildings in Milan, created data-driven schedules for electric use and occupancy from smart meters and assessed the impact of these schedules on the energy results of a UBEM at different time and spatial scales. They found that fixed and predefined schedules tend to underestimate the energy results when compared to schedules with measured data. Still, for simulations at the urban scale, fixed schedules seem to be enough to describe energy patterns. On the other hand, for small groups of buildings, randomizing schedules can bring variability to the model energy results. Although this study was performed over a year and with different time and spatial scales, it was only developed for the residential sector and a specific region of Italy.

Fu and Miller [46] show an approach to predict energy consumption in buildings using data collected from Google Trends. It was found that the searched volume of information on certain topics is correlated with building energy use. This fact helps to improve the prediction of the energy model. This study was conducted only for educational buildings and office buildings, and only 293 power meters out of 2380 had a high correlation with Google Trends.

Park et al. [47] developed an approach named CROOD, which allows one to estimate the building occupancy from mobile device connections without ground-truth calibrations. It was found that CROOD can estimate with reasonable accuracy the number of occupants of the building from the number of mobile devices. On the other hand, this approach is limited by building size, type, and system. In addition, this approach works better with a limited variable number of people in the building. Also, this approach was only used for one educational building, and it is necessary to test it in larger and different buildings to confirm the results.

The number of people in the building can have a significant influence, for example, on the HVAC design, and inaccurate estimates can be one of the causes of errors in the design. Moreover, they may influence individual and urban building modelling and lead to errors in the estimation of energy consumption. This fact has led, in recent years, to new approaches to predict the number of people in buildings with accuracy. Although innovative approaches have been proposed, a major gap is the impossibility of extrapolating the general approach to estimate the number of people in other types of buildings. In addition, it is not possible to extrapolate these data for different countries and regions or different seasons.

2.2. Calibration and Validation

The calibration and validation of simulations is a key part of the modelling process [48]. Urban energy modelling is no exception, and this step is required to improve the precision of results and bring virtual energy models closer to reality. For urban energy modelling calibrations, Bayesian inference is “the process of fitting a probability model to a set of data and summarizing the results by a probability distribution on parameters of the model and unobserved quantities such as predictions of new observations” [49]. It is a topic that has gained attention and has been studied by different researchers in recent years. Tardioli et al. [50] exploit a non-deterministic Bayesian framework to identify sets of representative parameters for building clusters and to calibrate their model to overcome problems of data scarcity and to scale the solution for large applications. They found that the method has a better performance for large groups of buildings when compared to single buildings, because errors tend to eliminate each other. However, more studies must confirm these results, especially at longer time scales. Dilsiz et al. [51] conducted research studying the influence of using different spatial and time scales in urban building energy simulation results, using a validated Bayesian calibration approach. They stated that for an annual time period, the results’ accuracy can increase by performing a calibration with building-level data. However, this brings up concerns about privacy and complexity. On the other hand, performing calibrations of annual data instead of monthly data can increase the results’ accuracy at an aggregated level. In addition, this study highlights the need for studies detailing the performance of building-integrated solar energy systems with more accuracy.

2.3. Weather Data and Microclimate Modelling Tools

Weather data are important information for UBEM simulations, because weather conditions, such as humidity and exterior temperature, have a direct influence on building energy demand. Therefore, the precision of weather data selection is important to obtain highly accurate results. In addition, it is also reported that considering the urban microclimate affected by urban heat islands (UHIs) significantly influenced the energy consumption of the buildings [52].

Liu et al. [53] carried out research in Hong Kong, a high-density city, comparing buildings’ thermal performances with and without considering microclimate effects. MesoNH-TEB was used for urban microclimate data in EnergyPlus. They found that the thermal performance of buildings is highly affected by their surroundings, especially during hot seasons and in high-density urban areas. Because of this, not considering the microclimate can lead to an underestimation of building overheating by about 140%. Still, there are a lot of uncertainties present in the measures.

Xu et al. [54] investigated the influence of weather datasets on building thermal performance in a high-density city through a comparison between microclimate datasets and throughout a typical meteorological year. They found that the bias error is halved when using microclimate datasets. The microclimate data were developed by coupling measured weather data with an increment-temperature local model due to anthropogenic heat.

Katal et al. [55] studied a dynamic approach to integrate urban microclimate and building/thermal energy models in Montreal. They used CityFFD, which predicts microclimate features due to aerodynamics, coupled with CityBEM. It was stated that the spatial variation in the local air temperature during the hottest time period could reach 15 °C, which has a great impact on building thermal performance. In addition, the energy consumption of buildings using microclimate data results in a 5 to 23% variation when compared to uniform weather data.

A novel review conducted by Sezer et al. [56] considered different urban microclimate software tools, such as Envi-MET, OpenFoam, and CityFFD. Depending on their characteristics, they can be coupled with different energy simulation tools. For example, ENVI-met can be used to increase the accuracy of building energy load estimations with TRNSYS, ESPr, or EnergyPlus, and OpenFoam can be integrated with CitySIM or EnergyPlus to determine the convective effect in urban energy simulations. They stated that coupling computational fluid dynamics (CFD) with building energy models is a common coupling strategy but requires high computational costs and an increased simulation time. In addition, there are still few studies in the literature on coupling strategies.

Table 1. Software tools used to model the urban microclimate.

Software	Characteristics	Reference
ENVI-met	Allows for energy retrofit analysis, urban energy planning, building operations improvement, and energy benchmarking; Simulation time and computational requirements are too high; Adequate to study the impact of common energy efficiency measures; Does not consider the microclimate effect.	[57,58,59,60]
OpenFOAM	Suitable to assess the building demand and the performance of energy systems and renewable energy resource potential; Only allows one to estimate the solar potential at rooftops.	[61,62,63]
CityFFD	Contains thermal radiation, behavior, plant, and equipment models; Detailed radiation models based on simple electrical circuit analogy, allowing one to determine surface longwave and shortwave radiance.	[64,65]

presents the major characteristics and the references of the software tools used to model the urban microclimate.

ENVI-met was developed in 1998 by Bruse and Fleer [57]. This software can simulate microscale interactions between the atmosphere, vegetation, and surfaces. To model the atmosphere, it calculates the mean air flow using Navier–Stokes equations, temperature and humidity with advection–diffusion equations, and turbulence and radiative fluxes. In addition, the software models the soil, vegetation, ground surface, and walls.

Thomas et al. [58] used ENVI-met to study the effectiveness of green walls in the modification of the microclimate of an urban academic campus. It was possible to conclude that the green walls reduced the ambient average temperature by 1.3 to 1.6 °C in winter and 0.4 to 0.5 °C in summer. The software was used to model the interactions between the plants and air to estimate the ambient air temperature, water vapor pressure, relative humidity, and mean radiant temperature.

Forouzandeh [59] predicted the surface temperature of buildings using ENVI-met. The effect of different variables, such as the inside building temperature, was explored. It was found that ENVI-met cannot predict the surface temperature with accuracy for all days and hours of the year. The winter results showed a greater accuracy than the summer ones.

Aleksandrowicz et al. [60] assessed the accuracy of the new version of ENVI-met in determining the mean radiant temperature. It was shown that the accuracy of the simulation to determine reflective radiation fluxes increased with the new version. However, the mean radiant temperature obtained is still low and less accurate. Another problem described was the high simulation time, which limits the application to large-scale areas.

Wong et al. [61] coupled OpenFoam, which is an open-source 3D CFD software, with weather research forecasting (WRF) and EnergyPlus to create an urban microclimate model considering multi-scale effects, combined with building energy models. In this study, the outputs of WRF, which reproduced the mesoscale climate conditions, served as boundary and initial conditions for OpenFoam, which was used to refine the microclimate model and was combined with EnergyPlus to assess the building energy demand.

Mirza et al. [62] used OpenFoam to analyze the effect of building temperature on air temperature and wind velocity and profile. It studied the changes in infrastructure, including the addition of vegetation coverage to assess its cooling effect. OpenFoam was used to model the microclimate effect and was proven effective in studying the effect of variables such as water bodies, humidity, and albedo.

Kadaverugu et al. [63] improved the accuracy of the urban microclimate model, downscaling the wind flow from the mesoscale to the building scale using OpenFoam. Therefore, it was possible to assess microscale effects, such as vegetation, and to study several parameters, like tracing hazard residues or pollutants for hot-spot identification and thermal comfort due, for example, to changes in the wind.

Mortezazadeh et al. [64] presented the structure of a novel simulation tool called CityFFD, developed to respond to large-scale problems, including describing airflow around buildings, natural ventilation, and thermal stratification. It models the airflow using non-dimensional Navier–Stokes equations and introduces novel turbulence models based on the Smagorinsky large eddy simulation method. As weaknesses, a lack of wall functions, radiation, vegetation, and pollutant models was reported. Wang et al. [65] developed a study that first validated the CityFFD software and WRF-CityFFD combined method. It was concluded that with the combined method, it was possible to achieve an accurate wind distribution, especially in coastal areas and with a limited number of meteorological stations.

2.4. Urban Building Energy Modelling Tools

Urban building energy modelling tools should be chosen specifically according to the purpose of the work or the simulation approach, due to their great heterogeneity [66]. Other factors such as the experience of the user or the availability of the program, especially if it is an open-access software, influence the choice of software. Some of the tools that can be used for urban simulations are the following: CityBES [67], City Energy Analyst (CEA) [68], CitySim [69], UMI [70], and URBANopt [71]. The main characteristics of the existing software tools are indicated in

Table 2. Software tools to perform urban building energy simulations.

Software	Characteristics	Reference
CityBES	Allows for energy retrofit analysis, urban energy planning, building operations improvement, and energy benchmarking; Simulation time and computational requirements are too high; Adequate to study the impact of common energy efficiency measures; Does not consider the microclimate effect.	[67,72,73]
CEA	Suitable to assess the building demand, the performance of energy systems, and renewable energy resource potential; Only allows one to estimate solar potential on rooftops.	[69,74,75]
CitySIM	Contains thermal radiation, behavior, plant, and equipment models; Detailed radiation models based on simple electrical circuit analogy, allowing one to determine surface longwave and shortwave radiance.	[69,76,77,78]
UMI	Rhinoceros-based urban modelling design tool; Indicated to assess walkability and to perform daylight and building energy simulations; Allows one to evaluate retrofit measures, calculate operational and embodied carbon emissions, and estimate solar power potential.	[70,79,80,81]
URBANopt	Allows one to perform detailed simulations at individual building level; Allows one to evaluate building-to-building shading and solar access.	[71,82,83,84]

.

CityBES is a software developed in 2016 by Hong et al. [67]. It is a web-based software that allows for energy retrofit analyses, urban energy planning, building operations improvement, and energy benchmarking from a small group of buildings to all the buildings in a city. CityGML, which enables the 3D modelling of urban buildings, is an open-source software. Energy models are made using Energy Plus and OpenStudio. The architecture of the software is presented in Figure 7. On one hand, CityBES enables analyses not only for a group of buildings but also for each building separately. However, this can increase the time of simulation and the computational requirements. Another problem with this tool is that Energy Plus cannot consider inter-building effects, such as radiant heat exchange between building exterior surfaces, and microclimate effects.

Teso et al. [72] developed a study in Venice, in which CityBES was used to conduct an energy analysis of four common conservation retrofit measures. These were the replacement of existing windows for more efficient ones, applying insulation in the interior of the roof and the exterior of the external wall, and, finally, upgrading the heating system through a replacement of the boilers for more recent and advanced ones. This work does not consider microclimate effects either, due to software constraints, or the heat transfer through mutual radiation between exterior surfaces of surrounding buildings. As reported before, one limitation of CityBES is the fact that it does not consider microclimate effects. Therefore, Hong et al. [73] carried out research with CityBES in which a microclimate map for San Francisco was created and applied as weather data, allowing them to study the microclimate effect. Although it was possible to analyze the microclimate effect in San Francisco, for other cities, it is not possible yet to conduct the same study with CityBES.

CEA was first presented by Fonseca et al. [68] in 2016. It is an open-source framework that enables energy, carbon, and financial analyses of building and infrastructure retrofits. It can also find optimal energy generation schemes, promoting energy system optimization at neighborhood and district scales. With CEA, it is possible, in addition to other features, to assess, on a spatiotemporal scale, the availability of resources, to estimate the building energy demand, and to simulate conversion, storage, and distribution technologies. It seems to be an interesting tool to evaluate energy systems in urban building energy modelling. However, in terms of solar energy, it only allows one to estimate the rooftop solar potential and the technical potential of rooftop solar collectors, photovoltaic (PV) systems, and PV–thermal collectors. This fact makes this tool less suitable for estimating the potential of building-integrated PV technologies.

Oraiopoulos et al. [74] investigated the future energy demand in representative Swiss communities, using CEA, considering future scenarios relative to climate change, building envelopes, and energy system retrofits. This tool was suitable, as the objective of the work was an analysis of building envelope retrofitting and an energy system analysis.

Mosteiro-Romero and Schlueter [75] studied the effect of occupants and the microclimate on demand and, consequently, on the performance and cost of energy systems using CEA. The energy systems studied were PV systems, heat pumps, chillers, and thermal networks. In this paper, CEA seems adequate for the aim of the study, due to having the possibility of estimating the solar PV potential, modelling the thermal networks, and integrating a microclimate and occupancy approach.

Robinson et al. [69] introduced a software for urban building energy modelling in 2009 called CitySIM, which was conceived for urban settlements with sustainable planning. It allows one to simulate the energy demand of buildings and to determine the energy supply from renewable and conventional energy systems. The software contains thermal models, radiation models to consider the radiance effects between buildings, behavioral models to estimate, among other aspects, the occupancy behavior, and plant and equipment models, which include either HVAC or energy conversion systems.

Chen et al. [76] used CitySIM to calculate the solar-induced wall temperature in a group of buildings and used this parameter as the thermal boundary condition in a CFD model. Although the software was capable of determining the hourly irradiation in the walls and the respective temperature, the effect of the shielding of surrounding buildings on the wall temperature was not considered, and the stratification temperature was not calculated.

Adilkhanova et al. [77] conducted research in the high-density city of Seoul to study the climate and the integration of high-albedo materials to prevent urban overheating, as well as to study the energy implications for urban building energy modelling using CitySIM. The fact that this software has longwave and shortwave radiation models to take into account the radiance effects between the surroundings, neighboring buildings, and the ground, was a huge advantage to the authors in achieving these objectives.

Khan et al. [78] studied the combined impact of urban overheating and heatwaves on building thermal performance using CitySIM. The software allowed them to perform detailed dynamic energy simulations at the neighborhood scale. Some of the outputs obtained with the software consist of the building energy and electricity demand, surface shortwave irradiance, and surface longwave balance. Once more, the advantage of this tool was the fact that it was possible to use detailed radiance models, based on a simple electrical circuit analogy with a resistor capacity network.

Reinhart et al. [70] presented in 2013 a Rhinoceros-based urban modelling design tool to perform evaluations of operational energy consumption, daylighting, and walkability, called UMI. It uses Rhinoceros as the CAD platform to build the 3D geometric model, Energy Plus to perform building energy simulations, Daysim for daylight simulations, and custom Python scripts that allow for the assessment of walkability.

Buckley et al. [79] used UMI in a study in Dublin to evaluate energy retrofit policies in terms of energy costs, CO₂ emissions, and energy use intensity. They tested a novel approach in which an EU database called Tabula was used, which contains more information than the one required by UMI. It was found that this approach allowed them to provide accurate assessments and district-level energy policy evaluations.

Wang et al. [80] conducted a study in Beijing in which the carbon emissions of a building through numerical simulations were evaluated with UMI. In this paper, the microclimate was considered, using a tool of UMI, called the Urban Weather Generator. To calculate the carbon emissions of the building stock, UMI calculates the operational energy consumption, and, also, the embodied energy consumption. This allows one to calculate, in a similar way, the operational and the embodied carbon emissions.

Buckley et al. [81] evaluated the energy profile of a neighborhood with commercial and residential buildings using UMI. The sharing of the solar power potential and district heating/cooling systems was considered. The software also allows one to estimate the actual and future energy demand of buildings, based on climate change projections.

URBANopt [71] is an open-source tool that was developed by the NREL (the National Renewable Energy Laboratory, USA) in 2016. OpenStudio allows one to perform detailed simulations at individual building levels using EnergyPlus, estimating the building energy demands. The software allows one to evaluate building-to-building shading and solar access. In addition, it is possible to assign several energy systems, such as community-scale PV, central heating and cooling plants, ground source heat pumps, and energy storage systems.

Wang et al. [82] used URBANopt to evaluate the impact of energy-efficient measures and distributed energy resources on a community’s energy usage, carbon emissions, and peak demand in Denver, Colorado. To achieve these goals, different modules were used to assess different energy efficiency measures that combine different scenarios of distributed energy resources and building models. In addition, grid-interactive models were used to optimize PV and battery system size and dispatch.

Flores et al. [83] developed an accurate energy model for a community where only minimal building and energy use data were available. URBANopt was used to achieve a detailed, accurate, and physics-based model that represents the energy usage of the community.

Ge et al. [84] studied the effect of vertical meteorological patterns in China through the creation of a building energy model developed in URBANopt. It was possible to estimate the heating and cooling loads of the buildings and assess the effect of the urban blocks.

3. Effect of Photovoltaic Systems on the Indoor and Outdoor Environment

3.1. Effect of PV Systems on the Indoor Environment

In recent years, different studies have been carried out in which the orientation of BIPVs and its effect on building indoor environments were analyzed. Uddin et al. [85] developed a research study on the potential of the application of different types of semi-transparent integrated PV window systems in Bangladesh for energy conservation through a numerical simulation in Energy Plus. It was found, as expected, that the south-facing integrated window systems generate more electricity and consume less electricity for lighting. Still, although east-facing windows produce less electricity, the level of energy consumption is lower due to air conditioning. It is important to note that this country has a typical tropical climate and is located in the east; therefore, the results cannot be extrapolated for Portugal and other EU countries.

Nicoletti et al. [86] developed a mathematical approach to assess the electricity production of solar PV blinds. The parametric study was developed in the city of Rome in Italy. It was found that the maximum electricity production was achieved for southeast or southwest orientations. In addition, automating the closing of PV blinds can increase electricity production. However, in this study, it was not possible to develop a method capable of achieving a compromise between electricity production, energy consumption, and daylighting performance.

In China, Feng et al. [87] carried out a study that assessed the potential of BIPVs for residential buildings. They concluded that rooftops are the best location to implement BIPVs, followed by south-facing facades (especially in high-latitude regions), and east-facing and west-facing facades, respectively. North facades can be adequate for areas with lower solar irradiation and with a lot of clouds. In this study, the use of electrical storage for BIPVs was not considered.

Along with BIPV orientation, shadows also have an effect on daylight performance, building consumption, and electricity production. Surrounding buildings may influence the amount of solar radiation that reaches the panels and need to be considered. In this context, BIPVs at an urban scale have been studied in recent years. Ye et al. [88] proposed a multi-objective framework for maximizing solar PV potential while minimizing PV area and enabling proper sunshade to improve electricity generation and the amount of indoor sunshade using PV-integrated sunshade devices. The study was developed in Hong Kong and considered real weather conditions and shading caused by obstructions, such as other buildings. An optimized solution was obtained in terms of PV shading, considering the balance between energy generation, cooling needs, and daylight performance. However, it was stated that the precise determination of window positions and building surface reflectance were a challenge in modelling the power potential of PV shades.

Liu et al. [89] conducted a multi-objective optimization study to improve the energy and environmental performance of residential blocks, including their energy consumption, PV energy potential, and daylight performance. It was found that 58 Pareto solutions increased the PV energy potential by 52.7% and sunlight hours by 50% and reduced the energy consumption by 1.5%. The study was conducted in China in an ideal residential block. Nevertheless, it was highlighted that a lot of time was needed to perform the simulations and only some specific variables were chosen to obtain the optimization, such as the building type, number of floors, and location of open spaces. None of these variables include parameters that affect the performance of BIPVs, such as shades or materials.

Xiang and Matusiak [90] explored the effect of balcony shading on facade-integrated PVs and on the daylight performance of high-rise buildings. The architectural aspects of the system were also assessed. It was found that it is important to avoid shading on lower-floor living rooms and southern surfaces because of the high solar potential. Although this study produced some interesting findings in terms of the effect of shading, in this case, due to balconies, it did not assess economic aspects or accurately evaluate the real efficiency of facade-integrated PVs.

Fan et al. [91] used the DaySim tool to investigate the indoor lighting environment to determine the optimal area ratio of different materials of skylight-integrated PVs. It was found that cities with sufficient sunshine need a higher area of PV materials. In addition, possible shading can be necessary to avoid excessive lighting and glare. A weakness of this study was that it did not consider factors that affect daylight, such as the type of skylight and the geometrical configuration of interior rooms.

To assess the economic potential of BIPVs, Mendis et al. [92] conducted a study in Sri Lanka, which evaluated horizontally inclined PV-integrated shading strategies. It was found that horizontally inclined installations with a distance-to-length ratio of four and inclination of 30º are significantly more efficient than typical vertical installations. However, the effect of PV-integrated shading devices on daylight performance has not been studied.

Besides energy consumption, BIPVs have the potential to improve thermal and visual comfort in buildings. The solar transmittance of the technology, allowing the sun to enter the rooms, affects the visual comfort of buildings. Also, the heat gains or shadow effects of BIPVs can affect both levels of thermal comfort and energy consumption.

Lee et al. [93] conducted research to evaluate the potential of applying PV modules to light shelf reflectors to improve levels of indoor comfort and energy efficiency. They found that in the summer and mid-season, the level of lighting consumption was lower due to an increase in the amount of natural light. However, the efficiency of the modules was higher in mid-season due to relatively lower temperatures, and with increasing module area, both cooling and heating energy consumption increased.

Wang et al. [94] studied the effect of semi-transparent PV windows on indoor thermal comfort. It was found that the change in illuminance in different ranges has a significant impact on thermal comfort. Furthermore, with semi-transparent PV windows, the indoor air temperature has a higher level of variation when compared with simple clean glass and also presents a higher value. This was also felt by the occupants, who reported a warmer and uneven environment.

Yadav et al. [95] compared the thermal performance of opaque PV Trombe walls with semi-transparent ones. It was found that with semi-transparent cells, it was possible to achieve a higher efficiency due to the lower PV cell temperature, although the indoor air temperature was higher. On the other hand, the opaque PV Trombe walls allowed for higher levels of indoor thermal comfort.

Taşer et al. [29] conducted a novel review that stated that BIPVs are more adequate for moderate, temperate, and sub-tropical regions, due to an increase in the indoor air temperature, which is not beneficial for indoor thermal comfort. In terms of the number of studies in the literature, it was indicated that there are more studies focusing on the effect of BIPVs on energy performance than on indoor comfort. As a consequence, the effect of BIPVs on both energy performance and comfort, either visual or thermal, has not been studied.

3.2. Effect of PV Systems on the Outdoor Environment

In the last few years, studies on the relationship between PV systems and building outdoor environments have been conducted, and there have been updates to the state of the art. Additionally, the integration of PV systems into building systems and structures, such as roads and parking lots, has been studied. Ding et al. [96] conducted research that analyzed the energetic, economic, and environmental effects of the integration of solar PVs into bus parking lots. It was found that solar PVs can meet about 50% of the electricity demand of electric buses, with the longest payback time being about 6.2 years, and, therefore, can reduce the emission of pollutants. However, the study does not evaluate the effect of these structures on the urban climate.

Furthermore, Serrone et al. [97] assessed the potential of the substitution of asphalt road pavements for cool pavements with integrated PV panels to mitigate the UHI. Although, in terms of the air temperature and predicted mean vote, the solutions do not present significant differences, the mean radiant temperature was 4 °C higher for the integrated PV roads, which may influence thermal comfort. Still, the PV roads could meet 20 times the demand for lighting, and they may be used for other applications, such as vehicle charging. However, the simulations were performed only in July.

To evaluate the impact of the integration of PVs on cooling loads in buildings, Garshasbi et al. [98] carried out research in Sydney during two summer months. It was found that PVs can increase the ambient air temperature between 0.6 and 2.3 °C, which translates into an increase in cooling loads, depending on the wind intensity and direction, and the characteristics of urban surfaces, such as the emissivity. They emphasized the importance of using PV panels with a high level of reflected light in higher wavelengths to mitigate the increasing air temperature caused by these systems.

In a review conducted by Elhabodi et al. [99], the increase in air temperature due to BIPV systems was assessed, its causes were identified, and solutions to mitigate it were identified. It was stated that the increase in cell temperature led to a decrease in efficiency, and the release of heat to the exterior increased the ambient air temperature. Also, PV panels are made with low albedo materials, which cause an increase in cell temperature and, consequently, enhance the UHI effect and the cooling loads. However, PV panels can act as shading devices for buildings, which may reduce their cooling needs.

As mitigation strategies, the thermal control of PV panels through ventilation systems can help to increase efficiency and mitigate the UHI effect. In addition, the adoption of green spaces with a high albedo in cities, PV panels attached to cool roofs, or semi-transparent PV panels integrated into buildings can help to decrease the air temperature. However, it is suggested that in the future, studies should focus on different applications and types of PV panels in urban areas to explore the link between PVs and UHIs.

4. Summary and Conclusions

Urban building energy modelling has great potential, as it may be used for several different applications. For example, it can be used to predict energy consumption and to estimate the energy and economic potential of retrofit measures at the district level, without the need for detailed simulations of each building. Additionally, it can be used as a basis for the creation of a digital twin, allowing for the instantaneous monitoring of energy consumption.

The methodology to create a bottom-up urban building energy model (UBEM) is identified in the literature. First of all, it is necessary to process and acquire the building data and use them in model creation and simulations. Afterwards, it is necessary to create geometric and energy models. Calibration and validation are final steps to guarantee the precision and accuracy of the urban building energy model. The methodology employed for selecting articles to review followed the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” (PRISMA) framework. The framework is illustrated in the Supplementary Materials.

However, in the literature, some problems and gaps are related to UBEM simulations. The complexity of modelling is still an issue. On one hand, creating a detailed energy model can require a lot of time, for example, for the creation of a geometric model. The creation of buildings and the assignment of shade and trees substantially increase the difficulty of modelling. In fact, it was found that most studies do not consider the effect of the terrain and the elevation. However, it is important to take them into account to estimate with precision the energy potential of PV systems. On the other hand, the more complex the system, the longer the simulation time will be. Additionally, the actual tools of UBEM do not simulate the heat dissipation of PV panels to the urban environment. This is necessary to estimate the effect of the integration of PV panels on the UHI.

It was concluded that data acquisition is still a weakness in urban building energy modelling, especially the acquisition of non-geometric data, which is one of the major sources of errors. The availability of data is a challenge because many times they are not accessible due to privacy concerns or high costs. Methodologies to develop precise archetypes for creating UBEMs are also lacking. Conducting energy audits on a sample of buildings can be an effective method to improve non-geometric data acquisition, and, consequently, building archetypes. Still, it is necessary to define the criteria of the buildings chosen in the sample to increase the accuracy.

In recent years, research related to Bayesian calibration has been carried out. These studies have revealed that this type of calibration approach can increase the accuracy of models. However, it requires a high level of complexity compared to common calibration approaches based on statistical indicators.

In relation to weather data, studies emphasize the importance of considering the microclimate effect. Despite the underestimation of overheating due to the microclimate effect, some studies do not consider the urban heat island (UHI). This can significantly influence the energy loads of a building and the comfort of the occupants, either inside or outside. A common approach to consider the microclimate effect and estimate the comfort and energy consumption of a building is to combine CFD with energy modelling software. In this way, with CFD, it is possible to consider the convective effects and to predict wind effects in detail. The choice of the software needs to be made according to the desired application and considering the capabilities of the programs to create a coupling strategy. However, the process is complex, and it is necessary to explore different coupling methods, because there are few studies on the matter. Therefore, convective models integrated into UBEMs can increase the accuracy of predictions of convective heat dissipation in the urban environment, the formation of UHIs, and, consequently, the building energy performance. This could be achieved by coupling UBEMs with CFD in order to obtain an optimized temperature profile across urban surfaces, instead of average wind and temperature values from a weather file or meteorological archive. However, this increases the complexity of the modelling significantly.

PV systems can be either integrated or attached to buildings, which affects their energy and daylight performance, as well as the comfort of occupants. In recent years, the orientation and inclination of PV systems have been studied. In terms of orientation, south facades or rooftops (in the Northern Hemisphere) are the best locations to install PV panels to maximize electricity production. However, the literature is inconclusive and presents contradictory results concerning the most suitable location and orientation to find an optimal compromise between energy consumption, daylight performance, and electricity production.

Electricity generation, energy consumption, and daylight performance using PV shading devices have been addressed. These systems can ensure that excessive light does not enter rooms, which, consequently, influences energy consumption and production. However, like the case of orientation, there are still few studies addressing shading solutions to find the optimal balance between electricity production, daylight performance, and energy consumption. The accuracy of the models in the determination of the real PV system efficiency or building surface reflectance has been stated as a weakness of these studies.

The impact of PV systems relative to indoor thermal and visual comfort has been studied. Although thermal comfort is related to visual comfort, it depends especially on the air temperature, either in terms of magnitude or variation, and studies have addressed the general energy impact of PV systems on buildings. The optimization of the design of PV shades can lead to a decrease in heating loads at the same time that electricity is generated. Additionally, visual comfort is improved and shading reduces excessive lighting; thus, indoor strategies at the building level are carried out to evenly disperse the light in a comfortable way to occupants throughout the rooms. So, although PV shades can be important to generate electricity, they can reduce the amount of daylight reaching rooms, which will have an impact on occupants’ thermal and visual comfort. However, achieving the optimal level of electricity generation and comfort is challenging: first, electricity generation depends on the orientation and inclination of PV systems; and second, one’s comfort level is subjective and depends on the temperature of the room and the incidence of solar radiation.

PV systems also affect the outdoor environment. The studies pointed out that the integration of PV systems, especially in facades, can increase the air temperature and accentuate the UHI effect during hot periods with higher solar radiation. This happens due to the lower thermal mass of the PV panels, which makes the cell temperature increase faster than the surrounding urban surfaces and dissipates the heat to the urban air. On the other hand, the UHI effect can lead to an increase in cell temperature and cause the efficiency of PV panels to decrease. However, the studies are not consistent regarding the magnitude of the effect of PVs on the UHI effect, and there is a need for more studies with real values of PV characteristics and urban surfaces. In addition, the balance between the increase in UHIs and the reduction in building cooling loads due to shading has not yet been identified.

Together with the air temperature, the average radiant temperature is a parameter to consider to evaluate outdoor thermal comfort. The studies indicated that PV shades have a higher mean radiant temperature than urban shades, due to lower light reflectance at higher wavelengths. Therefore, the presence of PV shades can lead to a higher sensation of discomfort, when compared with conventional urban shades, due to the higher radiant temperatures. Still, studies have revealed the need to study different technologies of PV panels and types of urban surfaces in urban environments to evaluate the effect of PV panels on UHIs. Notwithstanding, it is necessary to explore different mitigation strategies for UHIs caused by PV technologies and use realistic values for PV panels and urban surfaces, specifically regarding emissivity and reflectance at each wavelength.

The present article presented an up-to-date review about urban building energy modelling and about the effect of PV systems in urban environments (building outdoor and indoor environments). It reviewed urban building energy modelling tools and their applications. It also reviewed the urban heat island effect and its causes, which include building surface types and PV systems. Our conclusions are summarized as follows:

• A bottom-up UBEM approach is identified in the literature; however, the complexity level of the modelling needs to be carefully assessed. On one hand, a detailed UBEM can lead to a precise simulation; on the other hand, its high complexity requires a lot of time for modelling and simulation;
• Poor data acquisition, especially that of non-geometric data, can lead to large errors in the results. The calibration and validation of the model, despite its complexity, can mitigate the differences obtained;
• Urban heat islands can influence building energy loads. To predict their impact, coupling between UBEMs and radiation and convective models must be considered;
• With the integration of PV panels in building envelopes, in addition to electricity generation, their daylight and energy performance are affected. The reviewed studies are not conclusive about the relationship between these variables;
• The relationship between PV systems and indoor thermal and visual comfort must be studied at the building level;
• Due to their lower thermal mass, PV cells’ temperatures increase faster than the surrounding urban surface temperatures due to solar radiation. Consequently, PV efficiency decreases and can increase the outdoor temperature and building cooling loads. However, the studies do not show consistent results;
• PV systems have higher mean radiant temperatures than urban shades due to the lower level of light reflectance at higher wavelengths; therefore, PV shades can lead to a lower outdoor comfort level than regular shades.

The reviewed literature suggests a relationship between PV systems and urban heat islands. Consequently, building energy performance and comfort can be affected by PVs. However, more studies are needed, because there are not consistent results in the literature.

As the main conclusion of this review, UBEMs can be interesting tools to predict the effect of PV systems in urban environments and building energy performance. This is possible because UBEMs can predict building energy performance and, simultaneously, model the microclimate and urban heat islands. It is necessary to choose adequate software and methodologies for one’s intended objective. For example, it may be necessary to couple UBEMs with convective models to increase the precision of results. By incorporating PV thermal models, it is also possible to study the relationship between PVs and urban heat islands.

So, the authors recommend further investigations of this issue. New methodologies to couple UBEMs with PV models are needed. These are important to overcome the challenges associated with UBEMs, namely, the acquisition of non-geometric data, to allow for effective studies of the effect of PV systems in urban environments. In particular, the authors recommend that thermal models are incorporated in UBEMs, allowing one to predict with accuracy the heat released from PV panels to the urban environment. This will allow one to study the impact of PV systems on the urban air temperature, occupants’ comfort, and the energy building performance. With the incorporation of estimated future weather data, a solid coupling between UBEMs and PVs will be important to predict their long-term effects concerning UHIs and building energy performance.