Parametric Urban-Scale Analysis of Space Cooling Energy Needs and Potential Photovoltaic Integration in Residential Districts in South-West Europe

Abstract

The energy needs for space cooling are becoming a significant share of the energy balance of different Member States of the European Union, in particular the Mediterranean countries. This trend has been observed and monitored by the European Union, which has started a number of initiatives to promote the reduction in the energy demand for space cooling and have it satisfied by renewable energy sources, such as photovoltaic electrical energy. Nevertheless, even if the potential of those solutions has been widely investigated at the single-building level, this scale of analysis seems not fully adequate to support the definition of the energy policies addressed towards the renovation of the current cities into smart ones, with a large share of their energy demand satisfied with renewable energy. In this framework, this research aims to investigate the topic of building energy performance for space cooling services by adopting an urban-scale approach. In detail, a parametric simulation plan was run with CitySim in order to assess the impact of different quantities, i.e., climate conditions, districts’ and buildings’ geometry features, and the thermal quality of the building envelope, on the overall cooling energy need for districts and the specific building energy performance. Furthermore, the advantages of the integration of photovoltaic systems to supply power to the cooling system were analyzed, identifying the district configurations with the highest potential. For instance, in Athens, the share of space cooling demand satisfied by PV in high-rise nZEB configurations ranges between 64% (Building Density = 0.25) and 87% (Building Density = 1), while in the low-rise nZEB configurations it ranges between 81% (Building Density = 0.25) and 75% (Building Density = 1).

1. Introduction

1.1. Energy Consumption in the Building Sector and European Framework for 2050 Decarbonization

The exponential growth of critical issues related to climate change and natural resources deployment is strongly influencing the policies of the European Union (EU) [1], pushing towards the development of more sustainable and green economies [2]. The European Green Deal is one of the most iconic examples among the plans adopted by the EU for damping its impact on the environment by reducing its energy consumptions, increasing the efficiency of energy production systems and energy-consuming sectors, and enabling the decarbonization of all its processes [3]. Among the most energy-consuming sectors, there is the building one. Specifically, about 40% of European energy consumption is related to buildings, which are also responsible for about 36% of CO₂ emissions in the EU [4]. Considering that the world population is predicted to grow to 9.7 bn by 2050 [5], inducing about a 70% growth in the number of households [6], the energy consumption related to the residential sector is expected to furtherly increase. It is thus clear how retrofitting strategies and energy savings in this sector will be crucial for reaching the long-term European net-zero energy targets [7,8].

Among the critical aspects related to buildings’ energy consumption, there is the drastic and continuous growth of the space cooling energy demand, which is pushed by factors such as climate change, improved mean quality of life, and the increase in occupants’ indoor comfort requirements and expectations [9]. Among the most relevant actions taken in the last decade, it is worth mentioning the 2012/27/EU Energy Efficiency Directive [10], defining a roadmap for reaching a “highly efficient and decarbonized building stock by 2050”. A further milestone in the road to decarbonization is the Energy Performance of Buildings Directive (EPBD recast), i.e., the new Directive 2018/844/EU [11], pushing Member States to develop long-term strategies for the retrofitting of building stock and the increase in performance of the buildings sector. Despite the numerous actions taken for stimulating the retrofitting of European building stock, only a current 1–1.2% renovation rate has been assessed [12]. This value cannot be considered satisfactory, thus long-term renovation strategies are pushing to bring the renovation rate to 3% per year and reduce building energy demand in the EU by 80% by 2050, with the remaining 20% supplied by renewable sources [13].

Concerning space cooling specifically, the state of the art of the European space cooling market shows that about 99% of the whole space cooling load is covered by traditional vapor-compression space cooling technologies [14,15,16]. An increase in the use of both renewables and high-efficiency technologies could drastically reduce the share of the space cooling energy consumption in the European total energy balance. Indeed, improving energy efficiency [17,18] and increasing the renewable energy share [18], together with fixing the “performance gap” [19,20], accelerating the building renovation rate [13,21], and assessing “embodied energy” [22], are the five challenges most frequently listed in the literature for achieving the decarbonization of the building sector. Although recent studies show how building-sector retrofitting is currently dominated by step-by-step renovations [23], implementing urban-based retrofitting would allow for an increase in the intervention rate from 1% to more than 3%.

1.2. Simulation of Space Cooling Needs

The numerical assessment of the building energy needs for space cooling is of growing interest in the literature, also because of the importance the cooling demand assessment will have in the future. Bezerra et al. [24], for example, considered three specific warming level scenarios (i.e., +1.5 °C, +2 °C, and +4 °C) and identified a growth of both Cooling Degree Days and space cooling energy consumption, respectively, by +70%, +99%, and +190%. This trend, confirmed by the IEA [25], is leading to a growing importance of space cooling assessment. In this context, the researchers’ aims are often to identify active or passive solutions and strategies to improve energy performance, especially in hot climates [26,27], exploit renewable energy sources, and in particular solar energy [28,29], introduce simplified models for preliminary assessments, or define methodologies to tackle this problem at district or citywide scale [30,31].

The most frequently adopted approach is based on single-building simulations, which allow to characterize the cooling demand at an hourly or subhourly scale (i.e., with multiple calculation timesteps in an hour) [32] and support the development of simplified models. For instance, aiming to discuss the suitability of simplified approaches for the assessment of space cooling energy demand, Alhayek et al. [33] made a comparison between detailed simulations (using EnergyPlus) and simplified parametric regressions for a number of reference case-study buildings in Gaza, Palestine. Frayssinet et al. [31] performed an overview of city energy simulation models from the point of view of short energy dynamics, which are critical aspects for an efficient use of energy at the district and urban scale.

Although there are some examples of urban-scale simulations addressed towards the analysis of the cooling energy demands of districts and cities, there is not yet a comprehensive analysis able to support decision makers in the development of general strategies to implement at the EU or national scale. Indeed, many findings already present in the literature are often specific to the considered case studies, preventing the generalization required for the development of an energy policy.

In order to summarize, it appears fundamental to find approaches able to (1) identify the main drivers of the space cooling demand, (2) assess the potential of renewables to reduce the energy consumption, (3) investigate the topic of energy efficiency for space cooling at the district or urban scale. In this framework, the tools allowing building energy simulation at the city scale can become a fundamental instrument for preliminary studies aimed at developing new energy policies addressing building energy retrofitting issues.

1.3. Overview of Urban Building Energy Modeling

The advancement in building simulation in the last decades allowed the development of tools capable of performing accurate complex energy modelling and analyses for individual buildings [34]. The growing need in expanding these simulation tools to a broader urban scale led to the development of building energy models capable of considering energy/emission-related aspects on a wider urban scale [35,36,37]. According to Sola et al. [38], significant examples of Urban Building Energy Modeling (UBEM) tools and Urban-Scale Energy Modeling (USEM) tools are the following:

-

Urban building energy modeling tools (UBEM): BEM-TEB [39], CHREM [40], CityBES [41], SimStadt [42], TEASER [43], and UMI [44].

-

Urban-scale energy modeling tools (USEM): the model by Best et al. [45], City Energy Analyst CEA [46], CitySim [47], HUES [48], IDEAS [49], LakeSIM [50], MESCOS [51], SynCity [52], URBANopt [53], USEM platform with BCVTB [54], and USEM platform in UMEM project [55].

The aforementioned tools can also be categorized based on the number of required inputs, modelling approach, and spatial/temporal resolution [56]. Furthermore, according to Swan and Ugursal [57], they can be grouped as:

-

Top-down models: Models based on a black-box approach, allowing to calculate the aggregated energy demand of an entire region or country relying on historical data series [56,57].

-

Bottom-up models: Models based on statistical methods or engineering techniques allowing to simulate single-building/urban-district energy consumptions by end use [58].

1.4. Aims of This Research

As already stated in the previous paragraphs, it is fundamental for an effective retrofitting and for meeting EU energy efficiency targets to start working more on the urban scale rather than the single-building level. The state of the art of the research shows how, at the moment, the majority of studies are based on single-building simulations and not on the urban scale. Thus, developing the base of a parametric study for urban space cooling could serve as a starting point for future more complex developments of such models/studies. Furthermore, integrating renewable energies is recognized as an essential step for the sustainable management of the energy issues related to building space cooling.

This research aims at demonstrating how district- and urban-scale simulations can be effectively used for the definition of energy policies related to building space cooling in different scenarios of climate conditions and built environment, adopting simplified and parametric approaches with the final goal of ensuring integration with renewables (PV in this case). The research focused on understanding which factors were the most critical affecting factors for space cooling energy demand, in order to understand which retrofitting solutions should be considered, as well as how effective the integration with renewable energy production plants could be (e.g., PV systems). The results are discussed both in terms of annual energy needs for space cooling and peak loads, underlining which factors can contribute the most to their reduction.

2. Methods

A parametric approach was adopted in order to evaluate the impact of different factors on the energy needs for the space cooling of an urban district (Figure 1). Specifically, 8 different district geometries, 6 climates, and 3 levels of insulation of the buildings’ envelopes were considered. As a whole, 144 different configurations were simulated, obtaining both hourly profiles of energy needs as well as annual aggregated values. Furthermore, 2 different photovoltaic systems were modelled, leading to a total of 288 cases.

2.1. CitySim

CitySim [47] was used to run urban simulations on some geometry configurations drawn with Rhinoceros 3D. The developed geometries are based on the extrusion of squared buildings with flat roofs (for more information see Section 2.2.2). Citysim provides a decision support for urban energy planners aiming to evaluate the space heating and cooling loads of single groups of buildings or urban districts, minimizing the use of nonrenewable energy sources and, consequently, the related greenhouse gases emissions. The CitySim software was specifically selected for the simulations, given its capability of simulating the energy demand of groups of buildings, respecting the stochastic nature of occupants’ presence and behavior and accounting for a range of commonly used heating, ventilation, and air conditioning systems (HVAC systems).

To prepare a CitySim model, a “dxf” format of the geometry model describing the urban district is required. Different buildings’ envelope characteristics and surface properties can be included, together with some HVAC system features, in a dedicated “xml” file. The provision of specific files related to climatic conditions and horizon shape complete the data are required to perform the simulations.

2.2. Overview of Parameters and Variables in the Simulation Plan

2.2.1. Climate Conditions

Concerning the energy consumption for space cooling in the EU27 + UK, Italy, Spain, Greece, and France are the most impacting countries [59]. For this reason, it was decided to select some cities belonging to these nations, respectively, Milan, Rome, Madrid, Athens, and Paris. Furthermore, for the sake of comparison, the continental climate of Berlin, Germany was included in the list of cases as well. Hourly weather files were extracted from the dataset of EnergyPlus [60] and further elaborated to make them compatible with the Meteonorm “cli” format used by CitySim.

Although CitySim also allows to simulate the effect of the horizon profile, in particular regarding shadows cast by mountains, a flat horizon was set for all simulations in order to simplify the comparison. The insertion of nonflat horizons could be considered for future implementation of this model.

2.2.2. Urban and Buildings Geometry

For each climate, different geometries of the urban district were modelled, based on the layout shown in Figure 2.

This two-dimensional grid represents a simplified urban district divided by two main roads in four subsections. Each subsection is divided into six squares with a 20 m side (i.e., an area of 400 m²). They are furtherly subdivided into four squares of 10 m sides and 100 m² area. Four building density (BD) levels (BD = 0.25, BD = 0.50, BD = 0.75, and BD = 1) were modelled, corresponding, respectively, to one, two, three or the whole four 100 m² cells occupied by buildings. Finally, the number of floors was varied in order to have configurations with either low-rise (i.e., 2-floor buildings—6 m height) or high-rise buildings (i.e., 10-floor buildings—30 m height). The obtained eight geometrical configurations are shown in Figure 3.

As regards the windows, for each building a surface equal to 1/8 of the floor area was imposed, in agreement with the minimum value prescribed by the current Italian law [61].

2.2.3. Building Envelope

As far as the opaque components are concerned, clay-block and concrete elements were selected, since they are commonly used in Europe in the residential building stock. Three levels of insulation were set, representative, respectively, of (1) nearly zero-energy buildings nZEB, (2) partially insulated buildings, and (3) uninsulated buildings.

While different compositions were modelled for vertical and horizontal elements, for sake of simplicity the same layers were applied to both roofs and floors.

Nearly Zero-Energy Buildings (nZEB)

In the nZEB case, the building envelope was modelled in agreement with the requirements currently applied in Italy for this kind of buildings [62].

Table 1. Wall composition for the nearly zero-energy buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Clay block	0.20	0.25	893	840
Polystyrene	0.15	0.04	40	1470

The U-value is 0.21 W m⁻² K⁻¹.

and

Table 2. Composition of the horizontal components for the nearly zero-energy buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Concrete 1	0.06	1.4	1800	1000
Concrete 2	0.18	0.6	1800	1000
Polystyrene	0.17	0.04	40	1470

The U-value is 0.22 W m⁻² K⁻¹.

describe the compositions of vertical and horizontal elements, respectively.

As regards the transparent surfaces, triple-glazed windows filled with argon and with low-e coating (5.7-12-5.7-12-5.7), characterized by a U-value of 0.79 W m⁻² K⁻¹ and a g-value of 0.57, were modelled.

Partially Insulated Buildings

This second group represents buildings constructed usually before the 1990s and characterized by limited insulation levels [63]. As it can be seen in

Table 3. Wall composition for the partially insulated buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Clay block	0.09	0.9	2000	1000
Air gap	R = 0.18 m2 K W−1
EPS	0.08	0.059	2000	1000

The U-value is 0.52 W m⁻² K⁻¹.

and

Table 4. Composition of the horizontal components for the partially insulated buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Concrete 1	0.06	1.4	1800	1000
Glass fiber	0.06	0.046	30	1030
Concrete 2	0.18	0.6	1800	1000

The U-value is 0.55 W m⁻² K⁻¹.

, besides the massive clay-block and concrete layers, an additional insulation layer of EPS or glass fiber was included in the wall composition.

The transparent elements are made of double-glazed windows filled with air (4-12-4), with a U-value equal to 2.8 W m⁻² K⁻¹ and a g-value equal to 0.78.

Uninsulated Buildings

This last group of buildings lacks any insulation layer, as reported in

Table 5. Wall composition for the uninsulated buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Clay blocks	0.20	0.25	893	840

The U-value is 1.03 W m⁻² K⁻¹.

and

Table 6. Composition of the horizontal components for the uninsulated buildings (from the internal to the external side).

Material	Thickness s (m)	Thermal Conductivity λ (W m−1 K−1)	Density ρ (kg m−3)	Specific Heat Capacity c (J kg−1 K−1)
Concrete 1	0.06	1.4	1800	1000
Concrete 2	0.18	0.6	1800	1000

The U-value is 1.03 W m⁻² K⁻¹.

, and has the same type of windows described for the partially insulated buildings (i.e., double-glazed windows filled with air).

2.2.4. Fixed Parameters

This section describes the factors kept constant for all simulations, i.e., occupancy, ventilation, and shading systems. These factors surely have a contribution in the evaluation of the cooling load, but for the sake of simplicity they were not involved in the parametric analysis developed.

The same occupancy density was applied to all configurations. Specifically, a value of 0.04 persons m⁻² was set, in agreement with the technical standard UNI 10339 [64]. Concerning the sensible heat gain due to the occupants’ metabolic activity, 90 W person⁻¹ was imposed, as suggested by the ASHRAE Handbook of Fundamentals [65]. Furthermore, the radiative share was set at 0.6 and the latent heat gain neglected, since a sensible heat balance was analyzed with CitySim. Finally, the standard occupancy profile for residential buildings implemented in CitySim was adopted.

A constant average ventilation rate, equal to 0.3 ACH according to the UNI 10339 [64], was imposed.

In all configurations, shading systems were activated only when the solar irradiance on the considered surfaces is higher than 300 W m⁻², defined as the cut-off irradiance suggested in the UNI/TS 11300-1 [66]. A default value of 0.2 of the CitySim reducing factor λ was imposed.

2.2.5. Photovoltaic Systems

Vapor compression technologies are, as already stated, the most widely used technologies for cooling, thus causing problematic peak loads during the hottest days of the year. These peak loads on the electric grid can, however, be damped by the use of photovoltaic panels. Matching the cooling electric consumption profile and the photovoltaic electric production profile could be of great interest, allowing to evaluate which percentage of energy can be provided by self-consumption (direct coupling with PV panels) and which should be provided by the grid.

In the simulated models, the whole available surface of the roofs was covered (flat roofs in the geometric models) with south-oriented surfaces representing the PV panels (1.5 × 10 m). They were inclined by a tilt angle optimized for the specific latitude of each considered climate. A minimum distance between the PV panel rows was set to avoid mutual shadowing, which would dramatically decrease the PV production of the whole row. The following Equation (1) describes how the optimal tilt angle was calculated [67], while Equation (2) shows the evaluation of the minimum distance between the PV rows. Table 7 reports the obtained PV panels’ distances and tilt angles for each selected latitude. Maintaining the correct distance among PV modules is fundamental for avoiding mutual shading, which could lead to the incorrect functioning of the shaded PV cells strings, causing a decrease in efficiency of the entire plant.

For φ between 25° and 50°: β = 3.7 + 0.69 ∙ φ

(1)

$$d=L+\frac{sin\left(\beta \right)-\cos \left(\beta \right)\ast \tan \left(\delta \right)}{\left(\tan \left(\delta \right)+\tan \left(\gamma \right)\right)\ast l}$$

(2)

where:

φ = latitude (°).

d = distance between each row of photovoltaic panels installed (m).

β = optimal tilt angle of the photovoltaic panels installed (°).

γ = inclination of the roof on which the photovoltaic panels are installed (°).

δ = declination angle of the sun during the winter solstice (°).

l = length of the photovoltaic panels installed (1.5 m in our case study).

Table 7. Optimal tilt angle and minimum distance of positioning for the photovoltaic panels for each evaluated climate.

Location	Latitude (°)	Optimal Tilt Angle β (°)	Distance (m)
Athens	37.97	29.9	1.75
Berlin	52.52	39.9	2.05
Madrid	40.41	31.6	1.80
Milan	45.46	35.1	1.91
Paris	48.85	37.4	1.98
Rome	41.89	32.6	1.84

Table 7. Optimal tilt angle and minimum distance of positioning for the photovoltaic panels for each evaluated climate.

Location	Latitude (°)	Optimal Tilt Angle β (°)	Distance (m)
Athens	37.97	29.9	1.75
Berlin	52.52	39.9	2.05
Madrid	40.41	31.6	1.80
Milan	45.46	35.1	1.91
Paris	48.85	37.4	1.98
Rome	41.89	32.6	1.84

Two different types of photovoltaic panels, respectively, a monocrystalline and a polycrystalline silicon panel already included in the CitySim database [47], were used. It was decided to use these two different technologies since they are the most widespread PV technologies currently implemented. Generally, monocrystalline silicon PV modules are more expansive than the polycrystalline ones but guarantee a higher efficiency. Their main characteristics are reported in

Table 8. Selected photovoltaic panels.

	MONOCRYSTALLINE SILICON PANEL	POLYCRYSTALLINE SILICON PANEL
	1Soltech—1STH-235-WH	APOS Energy—APOS130
Efficiency ηmpref (-)	14.5%	12.7%
Power Pmax (W)	235	130
Voltage at Pmax (V)	29.3	17.38
Voc (V)	37	22.07
Normal operating cell temperature Tcnoct (°C)	49.9	49.1
Reference temperature Tref (°C)	25	25
AC (m2)	1.625	1.023

.

For each case, the hourly power production was compared with the cooling energy uses in order to characterize the share of self-consumption of the generated electricity and the one to be integrated by the grid. Considering the specific focus of this research on space cooling, the share of power demand for space cooling satisfied by PV electricity was estimated as well. In these calculations, the power consumption due to space cooling was estimated assuming an EER = 3 in agreement with [62], and no batteries were considered.

3. Results

3.1. Simulated Energy Needs for Space Cooling

Hourly energy needs for space cooling, PV power production, and the self-consumption results were analyzed for the set of 144 configurations. The annual cooling needs for space cooling in the whole district (Figure 4) range from 53 MWh (i.e., configurations with low-rise partially insulated buildings and 0.25 building density in Berlin) to 2064 MWh (i.e., configurations with high-rise uninsulated buildings and unitary building density in Athens). The annual specific energy needs for space cooling were reported in Figure 5, ranging from values around 1 kWh m⁻² (e.g., configurations with high-rise buildings in Paris or Berlin) to values almost equal to 90 kWh m⁻² (e.g., configurations with low-rise uninsulated buildings in Athens).

The climate is the most impactful variable, as expected. Paris and Berlin show similar trends, with energy needs and specific energy needs for space cooling often negligible and, respectively, in the range 460–570 MWh and slightly larger than 20 kWh m⁻² in the worst case. Madrid, Rome, and Milan are characterized by larger energy needs, ranging from about 15 kWh m⁻² to about 65 kWh m⁻² (i.e., from 153 MWh to about 950 MWh) in the configurations with low-rise buildings and from about 3 kWh m⁻² to almost 23 kWh m⁻² (i.e., from almost 283 MWh to about 1420 MWh) in those with high-rise buildings. Athens, finally, shows the largest variability of energy needs and specific energy needs for space cooling—from 272 MWh to 2064 MWh and from 10 to almost 90 kWh m⁻², respectively.

Although high-rise districts usually have larger cooling energy needs, the simulated neighborhoods with low-rise buildings are characterized by significantly larger specific cooling energy needs compared with those with high-rise buildings. This can be motivated by (1) the mutual shading present in the cases with high-rise buildings and (2) the relatively lower external surface of high-rise buildings compared with the low-rise ones with respect to the conditioned volume. Indeed, the low-rise buildings have shape factors S/V equal to 0.73 m⁻¹, 0.63 m⁻¹, 0.60 m^−1, and 0.53 m⁻¹, respectively, for building densities (BD) of 0.25, 0.50, 0.75, and 1.00. On the contrary, high-rise buildings are much more compact, with S/V equal to 0.47 m⁻¹, 0.37 m⁻¹, 0.33 m^−1, and 0.27 m⁻¹, respectively for building densities (BD) of 0.25, 0.50, 0.75, and 1.00.

The building density has an impact on the cooling demand as well and the configurations, with BD = 1.00 characterized by slightly lower specific energy needs for space cooling. Moreover, in this case, the phenomenon can be explained with the shadows cast by a building on the adjacent ones, reducing the solar radiation gains, and by the buildings’ shape factors.

In absolute terms, although the configurations with BD = 0.50 have a double volume compared with those with BD = 0.25, the total cooling energy need rises far less than +100%. On the contrary, even if BD = 0.75 has just +50% more volume than BD = 0.50, the percentage increase in the cooling needs is much larger and, in some cases, exceeding +100%. Finally, increasing the conditioned volume of another +33% with respect to BD = 0.75 (i.e., considering the cases with BD = 1.00), the cooling needs remain approximately the same (in the configurations with low-rise buildings) or even decrease (in the configurations with high-rise buildings). This can be explained by analyzing again the variation of the shape factor S/V (Figure 6), which is equal to 0.10 m⁻¹ for the cases with BD = 0.25 and those with BD = 0.50, to just 0.03 m⁻¹ for the cases with BD = 0.50 and those with BD = 0.75, and to 0.07 m⁻¹ for the cases with BD = 0.75 and those with BD = 1.00.

Moreover, the quality of the building envelope affects the energy spent for space cooling, although to a lower extent compared with the other investigated variables. Especially in the warmest climates and for the configurations with small buildings, the more the envelope is insulated, the lower the energy need for space cooling. However, for colder climates (such as Paris and Berlin) and neighborhoods with high-rise buildings and intermediate building densities, that is not true anymore, and partially insulated configurations perform better than the nZEB ones.

This can be explained by analyzing the hourly and daily data of cooling energy needs (Figure 7). Indeed, it is possible to observe that in the case of well-insulated buildings (i.e., nZEB configurations) the space cooling load is generally lower in absolute terms. Nevertheless, the cooling period for the building is longer. Lower peak loads combined with a longer usage of cooling systems result in a worsening of the cooling energy performance of the nZEB buildings compared with the partially insulated ones. Comparing instead uninsulated and partially insulated buildings, a similar duration of the cooling period can be detected, with lower cooling peak loads for the latter configurations.

A correlation analysis was performed in order to confirm what was observed through descriptive statistics. By applying the Spearman test with a statistical significance of 0.05, the specific annual cooling energy needs were correlated to the buildings’ height (expressed in number of floors), density, insulation level (expressed through the U-value of the vertical walls), and the climate (characterized by means of the Cooling Degree-Days CDD calculated with a base temperature of 18 °C). Considering all six climates, correlations with building density and insulation level were found nonstatistically significant with respect to the chosen 0.05 level. On the contrary, the correlation between the specific annual cooling energy needs and the building height and the one between the specific annual cooling energy needs and the climate were found significant, respectively, with a Spearman correlation coefficient equal to −0.55 and 0.68.

3.2. Simulated Contribution by the PV Systems

Simulated hourly energy needs for space cooling were used as input in order to assess the power demand, assuming a standard cooling system with an EER equal to 3. With the hourly profiles of power demand, the self-consumption of electric energy generated by PV systems was calculated.

Considering that similar trends were observed for the different analyzed locations, and the configurations simulated in Athens showed the largest energy needs and the largest variability, this section presents and discusses in detail the potential integration with PV systems, first for this climate.

For each configuration in Athens,

Table 9. Annual results regarding simulated electrical energy demand for space cooling, electrical energy supplied by the grid, and self-consumption to satisfy the cooling energy demand for the low-rise configurations in Athens, considering both (1) monocrystalline silicon PV modules (gray rows) and (2) polycrystalline PV modules (rows in italics).

Building Density	Quality of the Building Envelope	Annual Electrical Energy Demand for Space Cooling (MWhel yr−1)	Annual Electrical Energy Integration from the Grid (MWhel yr−1)	Annual Self-Consumption of Generated PV Electricity (MWhel yr−1)	Share of Cooling Demand Satisfied by PV Electricity (%)
0.25	nZEB	90.9	17.3	73.6	81
			18.1	72.7	80
	Partially insulated	100.3	30.8	69.5	69
			31.5	68.8	69
	Uninsulated	139.0	58.6	80.4	58
			59.3	79.8	57
0.50	nZEB	109.3	27.2	82.1	75
			27.9	81.5	75
	Partially insulated	148.3	51.0	97.2	66
			51.9	96.3	65
	Uninsulated	239.2	104.0	135.2	57
			105.2	134.0	56
0.75	nZEB	205.5	47.9	157.6	77
			49.1	156.4	76
	Partially insulated	266.4	94.2	172.2	65
			95.7	170.7	64
	Uninsulated	418.9	188.3	230.6	55
			190.4	228.5	55
1.00	nZEB	200.3	49.5	150.8	75
			50.7	149.6	75
	Partially insulated	272.7	92.1	180.5	66
			93.7	178.9	66
	Uninsulated	438.7	187.8	250.9	57
			190.2	248.5	57

and

Table 10. Annual results regarding simulated electrical energy demand for space cooling, electrical energy supplied by the grid, and self-consumption to satisfy the cooling energy demand for the high-rise configurations in Athens, considering both (1) monocrystalline silicon PV modules (gray rows) and (2) polycrystalline PV modules (rows in italics).

Building Density	Quality of the Building Envelope	Annual Electrical Energy Demand for Space Cooling (MWhel yr−1)	Annual Electrical Energy Integration from the Grid (MWhel yr−1)	Annual Self-Consumption of Generated PV Electricity (MWhel yr−1)	Share of Cooling Demand Satisfied by PV Electricity (%)
0.25	nZEB	252.7	85.4	167.2	66
			90.0	162.7	64
	Partially insulated	229.5	87.9	141.6	62
			90.6	138.9	61
	Uninsulated	280.4	139.0	141.4	50
			143.4	137.0	49
0.50	nZEB	397.1	123.5	273.7	69
			128.4	268.7	68
	Partially insulated	401.2	141.5	259.7	65
			145.3	255.9	64
	Uninsulated	456.4	207.5	248.9	55
			212.9	243.5	53
0.75	nZEB	685.1	192.4	492.7	72
			201.4	483.7	71
	Partially insulated	618.0	206.6	411.3	67
			211.7	406.2	66
	Uninsulated	659.4	290.3	369.1	56
			296.6	362.8	55
1.00	nZEB	324.5	40.6	283.9	87
			42.3	282.2	87
	Partially insulated	624.3	209.6	414.7	66
			213.6	410.7	66
	Uninsulated	686.5	286.7	399.8	58
			292.1	394.4	57

report, respectively, for the cases with low-rise and high-rise buildings, the annual power consumption for space cooling, the amount of electrical demand which can be satisfied with the power generated by the PV systems, the amount of power absorbed from the grid and consumed for space cooling, and the share of power demand for space cooling satisfied with PV electricity. The results are reported for both monocrystalline and polycrystalline silicon PV system simulations.

As regards the low-rise configurations (Table 9), the self-consumption for the space cooling service grows with building density and decreases with the quality of the building envelope. Indeed, the districts with high BD values and uninsulated envelopes are the most energy demanding in the case of low-rise buildings. The same is true considering the power supplied by the grid to satisfy the energy needs for space cooling. As regards the share of cooling demand satisfied by PV electricity, nZEB configurations show the largest percentages, ranging from 81% (BD = 0.25) to 75% (BD = 1). On the contrary, uninsulated buildings are characterized by a lower variability of the share of cooling demand satisfied by PV electricity, respectively, from 58% (BD = 0.25) to 55% (BD = 1).

Although the configurations with high-rise buildings are characterized by a higher energy performance regarding space cooling, the overall energy needs are larger (Table 10). Consequently, the annual self-consumption of generated PV power and that integrated from the grid are also larger in absolute terms.

Some similarities with the low-rise configurations can be registered. For instance, the share of cooling demand satisfied by PV electricity is generally higher for nZEB buildings and lower for the uninsulated ones. However, the trend is opposite: while for low-rise buildings a negative correlation was found with building density (i.e., the higher the building density, the lower the share of cooling demand satisfied by PV electricity), this is not true for high-rise buildings, and the share grows in agreement with BD. Furthermore, this trend applies not only to high-performance buildings but also to the partially insulated and uninsulated ones. nZEB buildings also show the largest variability in terms of the share of cooling demand satisfied by PV electricity, ranging from 64% (BD = 0.25) to 87% (BD = 1.00). On the contrary, uninsulated buildings range from 49% (BD = 0.25) to 57% (BD = 1.00).

As regards the comparison between the two types of PV systems, it can be concluded that, although the monocrystalline silicon module outperforms the polycrystalline one, the differences in terms of the share of cooling demand satisfied by PV electricity are limited—usually being less than 2% for both low-rise and high-rise building configurations.

Considering the whole set of climates assessed in the parametric simulation plan, some differences can be noticed between northern and Mediterranean climates.

At higher latitudes, such as in Berlin and Paris, both space cooling load and solar radiation availability—and thus PV power generation—are clearly lower. Nevertheless, the shares of cooling demand satisfied by PV electricity are very high. In the northernmost simulated location, i.e., Berlin, all configurations with low-rise buildings show a share larger than 90% in the case of nZEB envelopes. This figure is always larger than 55% if uninsulated buildings are considered. More specifically, the largest shares of cooling demand satisfied by PV electricity, both equal to 95%, are found for a building density of 0.25 in the case of low-rise buildings and a unitary building density in the case of high-rise buildings. In the climate of Paris, this performance index reaches even larger values, respectively, equal to 96% and 99% for the same best cases identified for Berlin.

In the Mediterranean climates, such as Athens, discussed above in detail, the shares of cooling demand satisfied by PV electricity are lower. This can be explained by the larger peak loads for space cooling, which cannot always be satisfied with PV production. Compared with the case of Athens, other climates, such as Madrid, show a better performance, on average, in terms of exploitation of renewable energy for space cooling purposes. For instance, low-rise buildings with nZEB envelopes in Madrid range between 81% and 86%, while in Athens they vary between 75% and 81%. In the case of high-rise nZEB building configurations, the shares range from 72% to 82% in Madrid and from 66% to 87% in Athens.

As a whole, it can be observed that in the central and northern European locations almost the entirety of the cooling load could probably be satisfied with properly designed PV systems. On the contrary, the climate conditions of the Mediterranean locations make this target more difficult to achieve, despite the larger production of renewable energy. As a consequence, the couple with storage solutions could be considered for an overall satisfaction of the energy demand for space cooling with PV electricity.

4. Conclusions

Among the most relevant issues related to space cooling are the growing annual energy needs and peak loads, especially in correspondence to extremely hot events. In order to analyze the most influencing factors governing these phenomena at the district and citywide scale and to discuss the potential benefits from the integration with renewable energy sources, a parametric analysis was carried out in this research. Specifically, a set of districts were simulated, varying the climatic conditions, the geometry of the urban district (e.g., the height of the buildings, their layout, and density), and the quality of the building envelope. This study focuses on more South-Western European countries, such as Madrid, Milan, Rome, Athens, and Paris. However, based on the proposed approach and by simply changing the climatic conditions, all European areas could be easily investigated. After simulating with CitySim the hourly cooling needs and peak loads for all configurations, the power load for space cooling was calculated and compared with the PV power generated by two different PV technologies, with the goal to evaluate the possible advantages for the satisfaction of the cooling demand.

It was observed that:

• As expected, the climate conditions are the main source of variability for the simulated space cooling energy needs of the considered districts. For instance, while the largest specific annual cooling energy needs can be around 20 kWh m⁻² in the in locations of Paris and Berlin, they can achieve almost 90 kWh m⁻² in Athens.
• Due to their larger conditioned volume, districts with high-rise buildings are characterized by an overall energy need for space cooling larger than districts with low-rise buildings. However, despite a five-time larger volume, the overall energy need is at least twice as large as the one simulated for the low-rise configurations just in Athens, Madrid, and Rome for nZEB buildings and a building density of 0.50 or 0.75. In most of the configurations, districts with high-rise buildings had an overall energy need for space cooling within 100% of the value estimated for those with low-rise buildings.
• In terms of specific annual cooling energy needs, districts with high-rise buildings had higher energy performances for space cooling (e.g., usually from −25% to almost −90% of the specific annual energy needs of the districts with low-rise buildings). This is due to both the buildings’ shape factor, which is lower for the high-rise buildings, and the mutual shadows cast in the case of taller buildings, which was accounted for in the CitySim simulations. For similar reasons, districts with higher building densities can be characterized by lower values of specific annual cooling energy needs.
• The quality of the building envelope has a minor impact on the cooling energy performance if compared with the geometry aspects. Nevertheless, in warm climates and for districts with low-rise buildings, the more the envelope is insulated, the lower the energy need for space cooling. On the contrary, in colder climates and districts with high-rise buildings and intermediate building densities, the partially insulated buildings can perform better than the nZEB ones.
• Assuming to install PV panels on all available roofs, the annual generated electricity is always higher than the energy needs for space cooling for all configurations. Nevertheless, due to the high peak load for space cooling, the share of the energy needs satisfied by PV renewable energy is lower than 100%. Larger shares of cooling demand satisfied by PV electricity can be found in colder climates, while in the Mediterranean locations those figures are often lower. Furthermore, nZEB buildings are characterized by larger shares of the energy needs satisfied by PV electricity, and different trends can be observed in low-rise and high-rise districts. Respectively, in the former ones, the share decreases with building density, while in the latter ones, it increases. Again, this can be attributed to the role played by different shape factors and mutual shading on the buildings’ cooling energy needs.

As a whole, this simplified parametric analysis has allowed to identify the different relationships among climate conditions, the geometry features of both district and buildings, and the thermal quality of the building envelope on the peak loads and energy needs for space cooling. Although specific microclimatic effects, such as the urban heat islands phenomena and street-canyon ventilation, were not accounted for, the analysis highlighted the importance of the design of the urban district to reduce the space cooling load demand. Furthermore, it has allowed to observe that the potential of PV panels to satisfy the space cooling energy needs can be very different and should be optimized considering the specific features of the district.

Additional developments of this research will focus on more realistic geometry features of the district, considering for instance randomized heights and building densities in order to allow for a robust generalization of the findings to real urban building fabrics. Furthermore, simulated values for the energy efficiency of the cooling machines could be implemented instead of the use of the simpler approach related to EER (energy efficiency ratio), using, for example, a BIM method for the performance evaluations. In addition to that, the next steps will focus on the optimization of the design of the integration with PV systems, accounting also for their economic performance and the presence of electrical storage systems.