Technical and Economic Feasibility of Multi-Family Social Housing and Nearly Zero-Energy Buildings in Southern Brazil

Abstract

Several studies have shown that social housing in Brazil usually fails to provide thermal comfort to its occupants. This study aimed to define energy efficiency measures for a representative social housing model to, together with local production of renewable energy, achieve the nearly zero-energy target. The thermal performance and energy efficiency of the representative model were evaluated using computer simulation, considering the southern Brazilian climatic context. An analysis of the economic feasibility of energy efficiency measures was also carried out. The results showed that the nearly zero-energy model with energy efficiency measures on the envelope (EPS and gypsum board in the external walls and rock wall in the roof) and a solar water heating system reduced 13.1% of the annual primary energy consumption in comparison with the representative model. Considering the renewable energy generation in the nearly zero-energy building, the electricity consumption was 38,777.6 kWh/year lower than that in the representative model (57.0% reduction). The economic analysis of the energy efficiency measures indicated a positive net present value and a payback of nearly six years. It was concluded that using energy efficiency measures and an on-site renewable made it possible to reach the nearly zero-energy target in a representative social housing model.

1. Introduction

The residential sector was responsible for 28.3% of electricity consumption in Brazil in 2021, second only to the industrial sector [1]. By 2100, the demand for cooling energy in residential buildings is expected to increase by 72% worldwide, influenced by global warming and increased access to air-conditioning appliances in developing countries [2]. Another factor that should increase the representativeness of electric energy consumption by the residential sector is the perspective of an increase in the number of houses built, given the housing deficit in the country, which in the urban area is 5.9 million dwellings [3]. To reduce the housing deficit in Brazil, in 2009, the Federal Government created the My House, My Life Programme (Programa Minha Casa, Minha Vida), which aims to create mechanisms to encourage the acquisition of new housing units for low-income families [4]. Through this programme, about 7 million homes have been built since its creation, which is still insufficient to meet the existing housing deficit in Brazil [5].

Triana et al. [6] analysed and defined representative social housing designs built through the My House, My Life Programme. The authors assessed the performance of representative models in bioclimatic zones 3 and 8 of Brazil and concluded that the current construction techniques show low thermal performance. The poor thermal performance of these buildings can lead to the need for greater use of air conditioning, increasing energy consumption in residential buildings. One of the reasons for the poor performance is that the designs are designed in a standardised model throughout the country, without concern for the regional specifics of climatic conditions or construction materials. The main barriers to the implementation of energy efficiency and environmental comfort strategies in social housing built under the My House, My Life Programme were the perception of high costs related to the implementation of energy efficiency strategies, the limited knowledge of bioclimatic architecture and environmental comfort on the part of designers and builders, the lack of in-depth and practical knowledge of renewable energy implementation, and the lack of incentives for the development of efficient designs [7].

A challenge for this area of knowledge is conducting studies with even broader scopes, such as evaluating representative models in other climatic contexts. In this sense, it is possible to simulate the representative models and vary the limits of the aspects that impact the energy efficiency of Brazilian social housing. Consolidating the knowledge produced can result in objective information regarding the design characteristics and the effect of thermal performance from the combination of several characteristics. By applying the same case studies in different climatic regions, it would be possible to quantify the impact of each parameter when subjected to specific conditions. Based on this, studies on the cost of these measures can help designers and builders choose the most cost-effective adaptations for their buildings [8].

In Brazil, the Inmetro Normative Instruction for the Energy Efficiency Classification of Residential Buildings—INI-R [9]—establishes two conditions for a building to be considered nearly zero-energy. The first is that it is energy-efficient, proven by obtaining an A rating for the overall energy efficiency of the building, i.e., the building must have a highly energy-efficient envelope. The second condition is to have 50% or more of its annual energy demand, measured in primary energy, supplied by renewable energy generated within the building or the land on which it is located.

Some strategies that may improve the energy performance of buildings are optimised building shape [10] and window-to-wall ratio [10,11], enhanced window glazing [12], appropriate thermal transmittance [13,14], solar shading [10] and photovoltaic energy generation [15,16]. Although nearly zero-energy buildings have existed in the European Union for more than ten years, it is not recommended that Brazil imports the same NZEB strategies used in the countries of that bloc but rather create its own strategies. This is justified because the climatic, cultural and economic conditions of Brazil differ strongly from the same conditions in those countries [17]. Implementing a nearly zero-energy building depends on three principles: passive energy savings, efficient equipment (lighting, air conditioning and water heating, among others) and local renewable energy generation [18]. In cold climates, such as northern Italy, achieving a zero-energy balance in a home requires high levels of thermal envelope insulation, forced ventilation systems with heat exchangers and robust photovoltaic systems [13]. In Brazil, achieving a zero-annual-energy balance is technically feasible only by adopting photovoltaic systems since residential energy consumption is low compared with that in the European Union and the United States. Thermal insulation is an example of an energy efficiency strategy used in cold-climate countries, such as North America and Europe, and may not be appropriate for most regions of Brazil, where the climate is predominantly warm. This strategy can cause the building to overheat and demand even more energy from the air conditioners.

Given the poor thermal performance of social housing in Brazil, this research is justified by the need to design and implement energy-efficient social housing appropriate to Brazilian economic and social standards. The studies published in Brazil considered particular climatic contexts or failed to address part of the theme, such as local renewable energy generation. In addition, many studies do not evaluate the economic viability of the proposed measures through costs or payback periods, a fundamental issue for decision-makers [19]. Also, it is important to mention that the space for renewable energy is limited in multi-family housing, making it challenging to achieve NZEBs without a high-quality building envelope [20], which is rare in Brazilian social housing. Therefore, the research gap that this paper intends to explore consists of the technical and economic feasibility assessment of a near zero-energy multi-family social housing in the climate of the southern region of Brazil. Thus, this work aims to apply energy efficiency measures to a representative social housing design and adopt a system of local production of renewable energy to achieve the NZEB goal.

2. Materials and Methods

Figure 1 shows the method used. A thermal performance evaluation of the representative social housing was performed using the computational method of the NBR 15575 Standard [21], while the energy efficiency evaluation was performed in accordance with the INI-R method [9].

In the next step, energy efficiency measures were selected for the representative model to achieve the condition of a building with Class A overall energy efficiency. The two models (with and without energy efficiency measures) had their performance compared in terms of electricity and primary energy consumption estimated using computer simulation. Next, local renewable energy generation source sizing was performed so that the Class A model would be considered a nearly zero-energy building.

Finally, an economic feasibility analysis of the representative Class A model was performed.

2.1. Building Typology

The most common social housing typology in the study region is multi-family housing. Thus, we used a representative social housing design of the multi-family type characterised by Triana et al. [6], which is a 16-flat multi-family building. Representative buildings are essential tools for large-scale studies about thermal and energy performance. The building has sixteen flats, and all of them were considered in the analysis. Each floor has four housing units (flats) with 43 m² of private area, as shown in Figure 2.

The construction characteristics and thermal properties of the representative model materials are given in Table S1 in the Supplementary Material. The influence of the surroundings and the reflectance of the ground and neighbouring buildings were considered similar to those of the existing housing developments in the region, as shown in Figure S1 in the Supplementary Material.

2.2. Climatic Data

For this study, climatic data from Caxias do Sul, Brazil, were considered. The city is located at latitude 29°10′4″ S and longitude 51°10′46″ W. The warmest month is January, with average monthly maximum temperatures of 26.85 °C, while the coldest month is July, with average monthly minimum temperatures of 8.73 °C. The city is inserted in Bioclimatic Zone 1 according to the zoning established by the NBR 15220-3 Standard [22], which groups the Brazilian territory into eight bioclimatic zones.

2.3. Simulation Parameters

According to the NBR 15575 Standard [21], to obtain the upper level of thermal performance (see Section 2.4), the building models must be simulated under two conditions, i.e., with and without natural ventilation. In the computer simulation, the windows open whenever a long-stay room (living room and bedroom) is occupied and the internal dry bulb temperature is equal to or greater than 19 °C. The internal dry bulb temperature must be higher than the external one for the windows to open.

The annual thermal loads were obtained from the simulations without natural ventilation [21]. The thermal load was evaluated in the simulation model in periods in which the long-stay rooms were occupied and with operative temperatures between 18 °C and 26 °C, which is applicable in the South region. Cooling and heating setpoints were 23 °C and 21 °C, respectively. The annual thermal heating load was used to determine the energy consumption of air conditioning, considering an energy efficiency coefficient of 5.3 W/W for cooling and 3.0 W/W for heating, according to the INI-R procedure [9].

A solar water heating system was considered an energy efficiency measure. This system is the most common water heating system in Brazilian social housing [8,23,24]. Furthermore, the electricity consumption for water heating in the representative model was significant (see Section 3.2). Thus, a solar thermal water heating system was necessary to achieve Class A overall energy efficiency for the building under study (a requirement for the building to be considered nearly zero-energy).

For the representative Class A model, in addition to using a solar water heating system, solutions pointed out in previously published studies focused on social housing in cold climates in Brazil were analysed for the envelope, such as those of Tubelo et al. [25], Silva [26] and Dalbem et al. [27]. The selection of solutions considered the characteristics of the representative model envelope and the results of its thermal performance evaluation. The selection of energy efficiency measures aimed at reducing the thermal transmittance of the envelope is a guideline recommended for cold climates in the NBR 15220-2 Standard [28].

For the roof, rock wool, aluminium foil, precast ceramic slab, and precast expanded polystyrene (EPS) slab were evaluated as energy efficiency measures. For the external walls, autoclaved cellular concrete blocks, EPS and rock wool were evaluated, the latter two being combined with gypsum board (Figure 3). These elements were considered in addition to the components of the original representative model. The EPS and rock wool layers were considered for the inside of the building, and they were covered with drywall for finishing. Only in the case of autoclaved cellular concrete blocks was the additional layer considered for the external side, as this arrangement tends to give better results [29], and this material can be subjected to weathering. This scenario also considered covering the blocks with a 2.5 cm layer of gypsum. For the windows, the use of double glazing was evaluated. The thickness of the layers of the proposed materials was defined based on the commercial dimensions existing in the market. The absorptances of the external surfaces were kept in accordance with those in the original representative model. No fixed or mobile shading elements were considered for the windows because it is recommended to allow the direct incidence of sunlight during the cold period [28].

Table 1. Energy efficiency measures for roofing and its thermal properties.

Element	Layers	Thickness (cm)	Conductivity (W/m·K)	Density (kg/m3)	Specific Heat (J/kg·K)
Aluminium foil	-	0.01	230	2700	880
Rock wool (5 cm)	-	5.0	0.045	96	750
Precast ceramic slab (12 cm)	Ceramic	1.20	1.05	2000	920
	Air chamber	4.6	Thermal resistance = 0.081 m2K/W
	Ceramic	1.20	1.05	2000	920
	Mortar	1.00	1.15	2000	1000
Precast EPS slab (12 cm)	Concrete slab	4.00	1.75	2200	1000
	EPS and concrete	7.00	0.223	373	1000
	Mortar	1.00	1.15	2000	1000

Source: NBR 15220-3 Standard [22], and Ortiz and Bavaresco [30].

and

Table 2. Energy efficiency measures for external walls and their thermal properties.

Element	Thickness	Conductivity	Density	Specific Heat
	(cm)	(W/m·K)	(kg/m3)	(J/kg·K)
Autoclaved cellular concrete block (7.5 × 30 × 60 cm)	7.50	0.17	450	1000
EPS (2 cm)	2.00	0.04	15	1420
Rock wall (5 cm)	5.00	0.045	96	750
Gypsum board (2 cm)	2.00	0.35	900	840

Source: NBR 15220-3 Standard [22].

show the thermal properties of the proposed energy efficiency measures for the roof and exterior walls, respectively.

Initially, the energy efficiency measures of the building envelope were evaluated separately through computer simulation, considering the impact of their use on the thermal performance of the building. The measures that did not bring significant impacts, i.e., that resulted in an increase in the percentage of occupied hours within the operative temperature range of less than 1% were discarded. Subsequently, the impact of the combined use of the remaining efficiency strategies was evaluated.

2.4. Definition of the Thermal Performance Level

According to NBR 15575 [21], the computer simulation procedure represents the most comprehensive and representative way of analysing thermal performance, allowing evaluations to obtain minimum, intermediate and upper performance levels, where higher levels represent more efficient building envelopes. This procedure evaluates the annual thermal performance of the building envelope in relation to this envelope with reference characteristics. In this procedure, two models are elaborated: (1) the real model, with the geometric characteristics of the housing unit, and the thermal properties and compositions of the transparent elements, walls and roof; (2) the reference model, which represents the evaluated building, but with reference characteristics, as shown in Table S2 in the Supplementary Material.

When assessing the thermal performance to meet the minimum level, the real and reference models must be simulated, considering only natural ventilation in long-stay rooms. In order to obtain the intermediate and upper levels, the real and reference models must be simulated under two conditions, with and without natural ventilation [21].

From the simulation with natural ventilation, the following should be determined: (a) the percentage of occupied hours in long-stay rooms (living room and bedrooms) within an operative temperature range, considering that the operative temperature range for Bioclimatic Zone 1 is from 18 °C to 26 °C; (b) the maximum and minimum annual operative temperatures of each long-stay room, considering only their occupation periods.

From the simulation without natural ventilation, the following should be determined: (a) the annual sum of the hourly cooling load values; (b) the annual sum of the hourly heating load values.

The Supplementary Material presents the equations used to calculate these parameters in Section 2. The computational simulation procedure was performed using EnergyPlus software version 9.3. This software was used because it aligns with the standard requirements and is free. It is also the most widely used for simulating the thermal performance of buildings [31]. Other software available in the market can also be used since they meet the standard requirements.

At a minimum level, the simulation results for the real computer simulation model must show, over a year and during periods of occupancy in long-stay rooms, the percentage of occupied hours within the operative temperature range of the real housing unit that is greater than 90% of that obtained for the reference model. To reach the lower and upper levels, the building must show the minimum increase in the percentage of occupied hours within the operative temperature range and the minimum percentage of reduction in the total thermal load. These values are presented in Tables S3 and S4 in the Supplementary Material. At all levels (minimum, intermediate and upper), the criteria of maximum and minimum annual operative temperatures must also be met. The purpose of determining the thermal performance level is to classify the energy efficiency of the envelope.

2.5. Definition of the Energy Performance Level

The Inmetro Normative Instruction for the Energy Efficiency Class of Residential Buildings (INI-R) determines the procedures for determining the energy efficiency rating of residential buildings in Brazil. This instruction was under public consultation at the time of writing this paper. We used the base text published by PBE Edifica through Public Consultation No. 18, 12 July 2021 [32].

The INI-R [9] allows a building to receive a classification for overall efficiency or individual systems: the envelope and water heating system. The classification is based on the percentage of reduction in primary energy consumption, comparing the consumption of the building in the real model with that of the same building with reference characteristics (the reference model), equivalent to the C classification [9]. The equations used to calculate the percentage reduction and the primary energy consumption are in Section S4 of the Supplementary Material.

The envelope evaluation considers the heating and cooling loads of long-stay rooms, as well as the percentage of occupied hours within the operative temperature range and the maximum and minimum annual operative temperatures. Individual systems are rated from Class A (most efficient) to Class E (least efficient) (Table S5 in the Supplementary Material).

The classification of the envelope follows the computer simulation procedure described in Section 2.4, establishing the equivalence relationship shown in

Table 3. Equivalence relationship between thermal performance levels and energy efficiency classes of the envelope.

Energy Efficiency Class of the Envelope	Thermal Performance Level
Class A	Upper
Class B	Intermediate
Class C	Minimum
Class D	Minimum
Class E	-

Source: Based on Inmetro [11].

. The Class A envelope aligns with the upper thermal performance level for equivalence with INI-R [9] energy efficiency classes. Thus, since INI-R [9] establishes that a building must have an A energy efficiency rating to be considered nearly zero-energy, the building must also present a higher level of thermal performance.

The adjustments considered for equivalence to the efficiency classes are presented in Section S6 of the Supplementary Material.

The energy efficiency rating of the water heating system is based on the percentage reduction in primary energy consumption required to meet the hot water demand of the building. The lower limit of the percentage reduction for each rating range varies according to the type of system employed (with or without storage), as shown in Section S7 of the Supplementary Material. However, it is important to highlight that achieving Class A on individual systems is not required for an overall Class A rating.

2.6. Sizing of the Renewable Energy Source

A photovoltaic system was considered for the local renewable energy supply. Photovoltaic power generation system sizing was carried out using the Solar Simulator, an algorithm made available by the Ideal Institute created in partnership with the German Cooperation for Sustainable Development [33].

It is necessary to inform about the annual electrical consumption for sizing through the simulator. Since the building to be supplied is hypothetical, the data regarding electric consumption was estimated using the INI-R method [9]. The monthly cost of the electric bill was estimated based on the current taxes of the distribution company in the city of Caxias do Sul.

The simulator determines the power required to meet the building energy demand from the data provided. It considers the supply of the annual electricity demand provided by the user, deducting the availability cost, which varies according to the type of connection (single-phase, two-phase or three-phase). Photovoltaic generation represents data from the first year of operation of the photovoltaic system. For the photovoltaic modules, polycrystalline silicon technology is employed, with installation oriented to the north and an optimal inclination corresponding to the latitude of the chosen location.

The simulation follows the Energy Compensation System [34], which considers an annual balance without accumulating credits for the following year. Therefore, if energy is generated in addition to monthly demand, the exceeding amount is converted into credits, which are used to offset the electricity bill in subsequent months.

By sizing the photovoltaic system, the peak power required to supply 50% of the annual electricity consumption of the case study building was determined—a condition for the building to be nearly zero-energy, according to INI-R [9]. These data were used for the economic feasibility analysis of the energy efficiency measures.

2.7. Economic Feasibility Analysis of Energy Efficiency Measures

The direct costs of the energy efficiency measures for the envelope were calculated considering the unit compositions, cost of inputs and services of the National System for Research of Costs and Indices of Civil Construction (SINAPI). The value of benefits and indirect expenses was added to the direct costs. The photovoltaic system cost considered the reference values per Watt peak (Wp) installed used in the market [35].

The economic savings due to the adoption of energy efficiency measures were estimated by comparing the electricity consumption of the representative model with the consumption of the building with Class A efficiency. The electricity consumption data were obtained using the INI-R method [9]. The tariff charged by the electricity distribution company in Caxias do Sul is BRL 0.90 /kWh (BRL stands for Brazilian Real; on 20 February 2024, USD 1.00 equalled BRL 4.96.) for low-income consumers under the yellow flag (intermediate scenario between the green and red flags, which indicate the energy value depending on generation conditions).

The economic feasibility analysis of the energy efficiency measures was performed using three indices: net present value (NPV), internal rate of return (IRR) and discounted payback. These indices were used to consider both the cost of energy efficiency measures and the savings resulting from decreased energy consumption over the life span, which is a common approach in the literature [25,36].

The NPV (Equation (1)) was obtained by summing all the flow values of a cash flow, brought to the present date and discounted by the minimum attractive rate of return (MARR). It was considered that MARR was the average value of the Selic rate (the basic interest rate of the economy controlled by the Central Bank of Brazil) in the last 36 months before we carried out this research (4.23% per year).

The future cash flow and electricity tax values were readjusted based on the National Wide Consumer Price Index (IPCA) average over the last twenty years (6.17% per year). The analysis period was fifty years, equivalent to the useful life of the building [37].

$$\mathrm{N}\mathrm{P}\mathrm{V}=\left[\sum _{\mathrm{k}=1}^{\mathrm{n}}\frac{\mathrm{C}\mathrm{F}}{{\left(1+\mathrm{M}\mathrm{A}\mathrm{R}\mathrm{R}\right)}^{\mathrm{k}}}\right]-\mathrm{I}\mathrm{I}$$

(1)

where NPV is the net present value (BRL); CF is the cash flow; k is the period of each cash flow (1 ≤ k ≤ n); n is the analysis period (years); MARR is the minimum attractive rate of return (%); and II is the initial investment.

3. Results and Discussion

3.1. Thermal Performance of the Representative Model

The flats were evaluated individually by comparing the criteria of the percentage of occupied hours within the operative temperature range, maximum and minimum annual operative temperatures, and the total thermal load reduction of the real building in relation to that of the reference building. The most satisfactory performance in the three criteria was observed in the ground-floor flats, while the least satisfactory was in the penthouse flats, which is related to the building envelope area in each flat exposed to the external environment. All flats met the criteria for the minimum thermal performance level. However, none of them simultaneously demonstrated the ability to increase the percentage of occupied hours within the operative temperature range and reduce the total thermal load to obtain the intermediate or upper level of performance. One of the factors that caused the low thermal and energy performance of the building is the composition of the envelope, the materials of which are selected based exclusively on their low cost, to the detriment of thermal properties [6,36].

In the Supplementary Material, Table S10 presents the values of the increased percentage of occupied hours within the operative temperature range and the reduction in total thermal load when comparing the reference and real buildings. According to these values, all flats evaluated presented a minimum thermal performance level. Thus, based on the thermal performance, none of the flats of the representative model could be considered nearly zero-energy without modifications in the building envelope or the energy source.

3.2. Energy Performance of the Representative Model

Although the representative model does not have an air-conditioning system, the energy efficiency coefficient of a Class A appliance listed in the consumption and energy efficiency tables of Inmetro (the Brazilian agency responsible for assessing the conformity of products) was used to apply the evaluation method [38]. Table S11 in the Supplementary Material shows the parameters used to evaluate the water heating system.

All flats received a Class C rating of overall energy efficiency, as they presented a percentage reduction in primary energy consumption lower than 6% compared with that of the reference model (Table S12 in the Supplementary Material). The classification was carried out according to the primary energy consumption reduction ranges defined in Table S5 in the Supplementary Material for non-accumulation water heating systems.

Considering the energy consumption of all flats, the percentage of reduction in estimated primary energy consumption was 1.6%, comparing the representative model with the reference model. The energy consumption of the common areas was disregarded in this analysis. The representative model showed a total electricity consumption of 68,056.3 kWh/year, i.e., the flat consumes an average of 354.5 kWh/month, 126% more than the national average in 2019 of 157 kWh/month [39]. Compared with the average consumption in the state, 214.9 kWh/month [40], the consumption of the representative model was 65% higher.

Considering the total consumption shown in Table S12, it can be seen that the water heating system is responsible for most of the energy consumed in the representative model (48%), followed by the electrical equipment (42%). Energy consumption with air conditioning represents only 10% of the total, with the largest share going to heating (8%). Compared with average end uses observed in the Brazilian residential sector in 2019 [39], water heating represents the most significant share in the representative model. This end use corresponds to 16% of total electrical consumption on the national average and 48% in the object of study. This significant difference can be attributed to two main reasons. One is the use of solar heating systems and the expansion of the natural gas network, which has caused a reduction in the use of electricity for water heating in the national context [39], while in the representative model, an electric shower was considered to heat water. The other is the low temperature of cold water in the city (17.2 °C), which contributes to the increase in the consumption of the water heating system since this is one of the variables used for its estimation. Additionally, it should be noted that only 37.5% of Brazilian households have electric showers [40], which contributes to reducing the national average consumption of electric energy for water heating.

A factor that tends to increase the estimate of electric energy consumption through the INI-R method [9] is the calculation of cooling and heating loads, which considers the air-conditioning system the whole time the building is occupied. In this sense, natural ventilation is disregarded.

3.3. Energy Efficiency Measures

3.3.1. Energy Efficiency Measures of the Envelope

The measures were evaluated individually to verify their contribution to increasing the thermal performance of the representative model through the percentage of occupied hours within the operative temperature range, which is the main parameter for classification in the energy efficiency classes. This range represents the habitability limits related to the thermal performance of the building. These limits do not necessarily represent the thermal comfort of all occupants, which is also related to other psychological and physiological parameters [41].

Figure 4 shows the increase in the percentage of occupied hours within the operative temperature range in the model with the energy efficiency measures compared with the representative model. These results represent the minimum, average and maximum increase in the percentage of occupied hours within the operative temperature range among the sixteen flats evaluated. Regarding the energy efficiency measures used on the roof, the average increase was calculated based only on the results for the four penthouse flats.

The following energy efficiency measures were discarded because they presented insignificant (less than 1%) increases in the percentage of occupied hours within the operative temperature range: double-glazed windows, aluminium foil and precast ceramic slab on the roof. After discarding these measures, the combination of the remaining measures was evaluated. It is noteworthy that no energy efficiency measure was able to provide the representative model with the upper thermal performance level individually. The description of the evaluated combinations is presented in

Table 4. Combinations of energy efficiency measures evaluated.

Nomenclature	Energy Efficiency Measures
	External Walls	Roof
Combination 1	Autoclaved cellular concrete block (7.5 cm) and gypsum (2.5 cm)	Rock wool (5 cm)
Combination 2		Precast EPS slab (12 cm)
Combination 3	EPS (2 cm) and gypsum board (2 cm)	Rock wool (5 cm)
Combination 4		Precast EPS slab (12 cm)
Combination 5	Rock wool (5 cm) and gypsum board (2 cm)	Rock wool (5 cm)
Combination 6		Precast EPS slab (12 cm)

.

In order to verify whether or not the combinations of envelope energy efficiency measures would allow all flats of the representative model to achieve the upper thermal performance level (and consequently the Class A rating of envelope efficiency), the results of the increase in the percentage of occupied hours within the operative temperature range were compared with the respective minimum increase (Table S13 in the Supplementary Material). Only combinations 1, 3, 5 and 6 provided upper thermal performance for all flats of the representative model.

Figure 5 shows the percentage of occupied hours within the operative temperature range for the flat evaluated, comparing the performance of the representative model with and without the combinations of energy efficiency measures. These combinations also met the criteria for maximum operative temperature, minimum operative temperature and total thermal load (Figure 6) for the upper level of thermal performance.

Combinations 5 and 6 showed the highest percentage of occupied hours within the operative temperature range (20.8% and 17.5%, respectively), and combination 1 had the lowest increase (15.2%). Rock wool in the roof provides lower thermal transmittance than does precast EPS slab. Using gypsum board combined with rock wool or EPS in the walls provides lower thermal transmittance than does using an autoclaved cellular concrete block. These results corroborate the findings of Sorgato et al. [42] and Eli et al. [43], who observed favourable thermal performance in naturally ventilated envelopes with lower thermal transmittance.

On average, the alternative that provided the lowest maximum temperatures was combination 1 (28.4 °C). On the other hand, combination 5 was the one that provided the highest minimum temperatures (15.5 °C) and the lowest total thermal loads (469.9 kWh/year) on average. Combination 3 was the one that resulted in the smallest increase in wall thickness (+4 cm), while combination 1 resulted in the largest increase (+10 cm). The evaluation of this factor is significant because the flats have a relatively small area (43 m²), and this increase in thickness results in a significant loss of space.

According to INI-R [9], the envelope has Class A energy efficiency without air-conditioning systems if the occupied percentage of hours within the operative temperature range is higher than 95%. Only the flats on the ground floor achieved this performance in combinations 5 and 6 (Figure 5). Therefore, with none of the combinations, achieving a Class A rating of energy efficiency of the envelope in all flats was possible without considering air-conditioning systems.

Considering the social character of the representative building used for this study, we chose energy efficiency measures for the envelope that presented the lowest cost for further analysis. Combination 3 (EPS and gypsum board in the external walls and rock wall in the roof) had the lowest cost. The use of this combination results in an additional cost of BRL 77,997.36 (Table S14 in the Supplementary Material). Furthermore, this alternative resulted in the smallest increase in the thickness of the external walls, from 10 cm to 14 cm.

3.3.2. Energy Efficiency Measures of the Water Heating System

Incorporating a solar thermal water heating system was necessary due to the high electricity consumption for water heating in the representative model. Without an energy efficiency measure to reduce the electricity consumption for this purpose, achieving Class A overall energy efficiency for the building under study would not be possible.

For this, we selected an energy-efficient Class A system in the consumption and energy efficiency tables of Inmetro, the Brazilian agency responsible for providing the details regarding the products available [38]. The calculation of the energy for water heating from the solar thermal energy system and the energy consumption associated with the thermal losses of the system was performed in accordance with Annex B.III of INI-R [9].

Then, the electricity consumption for water heating was recalculated considering the energy contribution from the solar heating system. The proposed water heating system reduced electricity consumption by 14% compared with the representative model system (Table S15 in the Supplementary Material).

The estimated solar water heating system cost was BRL 28,852.88 (Table S16 in the Supplementary Material).

3.4. Energy Efficiency of the Nearly Zero-Energy Building

In the context of this paper, the nearly zero-energy building is the representative model with the energy efficiency measures described in Section 2.3. In the case of the envelope energy efficiency measures, the lowest cost combination was considered (Section 3.3.1). In the case of the water heating system, the solar heating described in Section 3.3.2 was considered complementary to the electric heating of a single consumption point (electric shower).

Given the equivalence between the thermal performance levels of the NBR 15575 Standard [21] and the energy efficiency classes of the envelope of INI-R [9], the nearly zero-energy building of this work has an energy efficiency rating of Class A for the individual envelope system.

The reduction in energy consumption with the proposed water heating system (14% compared with that of the reference system) is equivalent to a Class B energy efficiency rating of the individual system, according to the ranges presented in Table S9 in the Supplementary Material. Classifying individual systems as energy efficiency Class A is not an eligibility condition for obtaining an overall energy efficiency Class A rating for the building.

The flats of the representative model were evaluated separately to calculate the overall energy efficiency class of the representative model with energy efficiency measures. Table S17 in the Supplementary Material shows the energy consumption and overall energy efficiency class of each flat of the representative model. All flats received a Class A overall energy efficiency rating, showing a percentage reduction in primary energy consumption higher than 12%. The classification was performed as a function of the primary energy consumption reduction intervals defined in Table S5 in the Supplementary Material for water heating systems without storage. Although the solar water heater has an accumulation tank, the classification intervals are selected as a function of the water heating system of the reference model, which, in this case, is an electric shower (without accumulation).

Considering the consumption of the flats, the estimated primary energy consumption was 110,706 kWh/year in the reference model and 94,574 kWh/year in the representative model with energy efficiency measures, representing a reduction of 14.6% (see Table S17 in the Supplementary Material).

When comparing the representative model with and without energy efficiency measures (Figure 7), the electricity consumption reduction was 8947.6 kWh/year (13.1%). When the end-uses were analysed separately, the most significant reduction was in consumption with air conditioning for heating (78.2%), followed by air conditioning for cooling (32.7%). These two consumptions, however, have low representativity in the overall consumption. In absolute numbers, the most considerable reduction in consumption was with the water heating system, at 4558.0 kWh/year.

Different results are observed in the literature, depending on the type of building, the climate and the efficiency measures adopted. Khunatorn et al. [44] developed a case study of energy improvement in an existing building in Thailand to make it nearly zero-energy. The energy reduction measures used were replacing all the bulbs in the building with LED bulbs and replacing low-efficiency air conditioners with high-efficiency ones. The results showed a reduction of 45.03 kWh/year (38.8%) in energy consumption. Matana Júnior et al. [19] also proposed retrofit measures in the lighting and air-conditioning system to reduce energy consumption in a university building in Brazil. The measures resulted in 18.0% reduction in the energy consumption of the building.

According to INI-R [9], a local renewable energy source must supply at least 50% of its annual energy demand for a building with Class A overall energy efficiency to be considered nearly zero-energy. Considering that the representative model with energy efficiency measures would have a total primary energy consumption of 94,574 kWh/year, the local renewable energy source would be required to have an annual generation of at least 29,554.4 kWh/year.

According to the data obtained through the Solar Simulator, a 23.1 kWp photovoltaic system would be necessary to provide the minimum generation indicated [33]. Table S18 in the Supplementary Material presents the data for the proposed local renewable energy generation system and its cost. For calculating the cost of the sized photovoltaic system, the price charged per Watt peak obtained from the local market of BRL 4.2 per Wp installed was used [35].

3.5. Economic Feasibility

The economic feasibility evaluation considered the cost of the energy efficiency measures and the electricity savings over the lifespan of the building when comparing the original representative model with the Class A overall energy efficiency design.

The cost of the energy efficiency measures for the representative model to be Class A was presented in Section 3.3.1 and Section 3.3.2. The local renewable energy source must also be considered for the representative model to be nearly zero-energy, the cost of which was presented in Section 3.4. Thus, the total cost of the energy efficiency measures for the nearly zero-energy building was BRL 199,250.24.

The electricity savings due to the use of the energy efficiency measures were calculated by comparing the consumption of the representative model with the consumption of the representative near-zero energy design, already discounting the local renewable energy generation (Table S19 in the Supplementary Material). The difference in consumption was 38,777.6 kWh/year, representing a 57.0% reduction.

In a case study carried out in an office building in the Brazilian city of Brasília to make it an NZEB, energy consumption decreased by 46%, comparing the original model with the model optimised with efficiency strategies (lighting system, air conditioning, natural ventilation and solar photovoltaic generation) [45]. Jung et al. [46] developed two possible scenarios for making an NZEB in South Korea, the first with a hydrogen-based energy supply and the second based on electricity. The NZEB included passive, active, and renewable systems. The scenarios were compared with a typical building using a gas boiler and air conditioning. The results showed that the first and second scenarios had primary energy savings of 53.0% and 62.4%, respectively.

In the case study of this work, there were annual financial savings of BRL 34,948.01, given the cost of electricity of BRL 0.90 per kWh. In this analysis, the performance loss of the photovoltaic system was not considered.

For the calculation of the economic indicators, a cash flow was established where the cost of the energy efficiency measures is the initial investment in year zero of the analysis. The annual economic savings, resulting from the difference in electricity consumption between the representative and the representative near-zero energy designs, are the future revenues. In this analysis, neither the maintenance costs of the energy efficiency measures nor the loss of performance of the photovoltaic system was considered.

The net present value was BRL 2.5 million at the end of the analysis period, the internal rate of return was 18.7% per year and the discounted payback was approximately 5.7 years. Figure 8 shows the discounted cash flows of this analysis. The investment is considered economically viable, given that the net present value is positive and the internal rate of return is higher than the minimum attractive rate of return.

Table 5. Sensitivity analysis of the inflation rate (IPCA).

Inflation Rate (% Per Year)	NPV (BRL)	IRR (% Per Year)	Payback (Year)
2.0	835,744.20	14.7	6.3
4.0	1,388,445.74	16.6	6.0
6.17	2,526,506.69	18.7	5.7
8.0	4,347,139.29	20.4	5.5
10.0	8,144,250.42	22.3	5.3

Considering the minimum attractive rate of return to be 4.23%.

shows a sensitivity analysis of the economics using different assumptions for the inflation rate adopted (6.17% per year) to analyse the attractiveness of the investment with possible future prices.

In a lower inflation rate (2.0% per year) scenario, the net present value would decrease by 66.9% due to the lower electricity cost savings. Despite this, the payback period would increase only by about seven months. On the other hand, in a higher inflation rate (10.0% per year) scenario, the net present value would be 222.4% higher, and the payback period would decrease by approximately five months.

Another uncertainty is related to the minimum attractive rate of return, which was considered to be 4.23% per year in this study. The net present value would increase by 104.9% and 47.1% if the minimum attractive rate of return was considered to be 2.0% and 3.0% per year, respectively, and would decrease by 20.4% and 40.0% if an 8.0% and 10.0% per year minimum attractive rate of return was adopted, respectively. With an increase in the minimum attractive rate of return from 2.0% to 6.0% per year, the payback period would increase from 5.3 to 6.0 years.

Although there is a great variation in the net present value due to the uncertainty in future prices, in all scenarios, the investment would be considered economically viable (with positive NPV values and an internal rate of return higher than the minimum attractive rate of return).

4. Conclusions

For all flats of the representative Brazilian social housing to achieve Class A overall energy efficiency, energy efficiency measures were proposed for the envelope and the water heating system. For the envelope to achieve Class A efficiency as an individual system, the less expensive alternative was to use a layer of EPS and gypsum board on the inside of the external walls and rock wool on the roof. A solar water heater was proposed for the water heating system. With the addition of these measures, the representative project received Class A overall energy efficiency, with a reduction in primary energy consumption of 13.1% compared with that in the representative model.

Once classified as Class A for overall energy efficiency, a photovoltaic power generation system sufficient to supply 50% of the annual primary energy demand was sized. In this way, the representative project could be considered a nearly zero-energy building, according to INI-R [9]. Considering renewable energy generation, the electricity consumption was 38,777.6 kWh/year lower in the nearly zero-energy building than that in the representative model (a 57.0% reduction).

From the economic point of view, it was verified that the nearly zero-energy representative project is economically feasible since the net present value is positive and the internal rate of return is higher than the minimum attractive rate of return considered. It was also found that the discounted payback was approximately 5.7 years.

The results showed that Brazilian social housing, which usually shows poor thermal performance, could achieve the nearly zero-energy target through energy efficiency measures and installing renewable energy on-site. For southern Brazil and similar climatic regions, EPS and gypsum board in the external walls, rock wall in the roof and a solar water heater are viable measures to improve envelope efficiency and reduce energy consumption. These measures can be considered in future social building design.

This study contributes to sustainability by addressing environmental, economic and social aspects. The proposal of energy efficiency measures to reduce building energy consumption and a renewable energy system may help mitigate environmental impacts. The results showed that achieving the nearly zero-energy target in social housing buildings is also feasible and may improve the well-being of the occupants.

Although economically feasible in this work, the nearly zero-energy building faces difficulties due to market issues and the lack of public policies to make it a reality. The builders of social housing, usually private companies, would have to pay for the extra cost of energy efficiency measures without enjoying the resulting benefits, which occur throughout its useful life. This over cost could make it unfeasible for low-income people to purchase their homes. In this context, public policies that encourage and subsidise near-zero energy housing would be necessary.

Despite the research findings, some limitations can be identified. This study evaluated a social housing representative model in the southern Brazil climate. Therefore, the results cannot be generalized to other building typologies in different climates. Furthermore, the results of the thermal performance and electricity consumption were obtained from computer simulations. Therefore, their accuracy cannot be verified, as no experimental measurements were carried out. Future studies could evaluate other building typologies and analyse other energy efficiency measures, such as external shading elements and variations in the external painting colour, in addition to different local renewable energy generation sources.