Comparative Analysis of BIPV Solutions to Deﬁne Energy and Cost-E ﬀ ectiveness in a Case Study

: The built environment remains a strategic research and innovation domain in view of the goal of full decarbonization. The priority is the retroﬁtting of existing buildings as zero-emission to improve their energy e ﬃ ciency with renewable energy technologies pulling the market with cost-e ﬀ ective strategies. From the ﬁrst age of photovoltaics (PV) mainly integrated in solar roofs, we rapidly moved towards complete active building skins where all the architectural surfaces are photoactive (Building Integrated Photovoltaics - BIPV). This change of paradigm, where PV replaces a conventional building material, shifted the attention to relate construction choices with energy and cost e ﬀ ectiveness. However, systematic investigations which put into action a cross-disciplinary approach between construction, economic and energy related domains is still missing. This paper provides the detailed assessment of a real multifamily building, taking into account retroﬁt scenarios for making active the building skin, with the goal to identify the sensitive aspects of the energetic and economic e ﬀ ectiveness of BIPV design options. By assuming a real case study with monitored data, the analysis will consider a breakdown of the main individual parts composing the building envelope, by then combining alternative re-conﬁgurations in merged clusters with di ﬀ erent energy and construction goals. Results will highlight the correlation between building skin construction strategies and the energy and cost parameters by identifying the cornerstones for enhancing e ﬃ ciency. The outcomes, related to the total life cost, self-consumption / su ﬃ ciency, in combination with di ﬀ erent building design options (façade, roof, balconies, surface orientations, etc.), provide a practical insight for researchers and professionals to identify renovation strategies by synergistically exploiting the solar active parts towards lower global costs and higher energy e ﬃ ciency of the whole building system.

which the goal is to reduce as much as possible the electricity injected or withdrawn from the grid, giving priority to self-consumption and storage of the energy surplus produced.
Numerous studies investigate the topic of Demand Side Management (DSM) as an alternative to storage to optimize the self-consumption value of a system. This strategy, however, is mostly convenient in the case of residential buildings. In fact, when office and commercial buildings are considered, where the load profiles follow fixed characteristics during the year, the mismatch of needs makes it more difficult to achieve feasibility [9]. In this case, as proposed by Martín-Chivelet et al. [9], optimization is carried out by simulating the differently oriented façade combinations that provide better output in terms of SCR and SSR. Sánchez et al. [11] also studied the performance of poorly oriented BIPV systems, but the study is not designed to evaluate the benefits in terms of self-consumption. As for batteries, even DSM cases in the literature are numerous for residential buildings. The concept of DSM concerns the temporary shift of energy demand by the user, due to the use of more energy-intensive equipment, such as dishwashers, washing machines and tumble driers. A large part of energy waste can in fact be charged to a non-virtuous behavior of users, which necessarily implies a lack of overlap between the needs profile and the production profile. Optimizing this profile, translating the load curves in an appropriate way, can therefore be a good way to increase self-consumption. Numerous references in the literature aim to improve self-consumption in residential buildings by combining the strategies previously exposed: DSM and storage [1,2,10,14,24,25].
Castillo-Cagigal et al. [26], define the concept of ADSM (Active Demand Side Management) as the automation control of user load management, potentially useful in the residential sector for the possibility of being combined with additional functions relating to the comfort and safety of the building. The interesting results of his study are the following: the discovery that the influence of storage systems in the energy balance has no linear growth based on its capacity but reaches an asymptotic level for high storage capacities.
The final assumption is, therefore, that the use of the ADSM strategy is the same as the use of a small electrical energy storage system and the combination of the two strategies significantly improves the direct use of photovoltaic production.
The investigations previously analyzed are mainly based on simulated PV production models created on the basis of real meteorological conditions. The only reference of real monitoring of both consumption and electrical production of a BIPV system is not aimed at optimizing the values of self-consumption and self-sufficiency. In [14], the power generation of the BIPV system and the building loads were investigated by monitoring the building to verify the worst array causing the reduction in generation performance and to verify the achievement of the purpose of zero-energy buildings.
From an economic perspective, the investigations previously analyzed show that the BIPV system and storage installation are becoming attractive investments due to decreasing trends in costs and feed-in tariffs, also in the case of energy efficient retrofitting.
Due to the assumption that, actually, electricity tariffs, PV system cost and system efficiency do not allow us to define BIPV solar energy technology, a cost-effective option is reported in some studies, as in the research of Radhi H. [27]. According to this, the only way such a system will be convenient is if the electricity tariff increases or if the total system cost drops drastically. However, as Scognamiglio A. specifies in her studies, in the case of the BIPV system, the economic evaluation is more complex because not only the direct costs of BIPV installation should be considered, but also the benefits deriving from PV energy production and the avoided costs of the traditional building materials [28].
Other studies focus on the evaluation of the cost-effectiveness of BIPV for the building skin, in addition to the energy performance [29,30]. In [29], Hammond, G. P. et al. study the performance of a domestic BIPV system utilizing energy analysis, environmental life-cycle assessment (LCA) and economic analysis. The results illustrate that the systems are unlikely to pay back their investment over the 25 year lifetime, even if the energy analysis determines that the system paid back its embodied Energies 2020, 13, 3827 5 of 23 energy in just 4.5 years. They also demonstrate the importance of government support to the future uptake of BIPV.
Bonomo et al. illustrate a new methodology for evaluating the cost-effectiveness of BIPV, based on a life-cycle costing approach, assessing the whole building envelope solution, taking into account the multi-functionality of the system and, therefore, all the advantages of BIPV [29].
The economic sustainability of BIPV technology is a crucial aspect of its feasibility and market success. So far, the main efforts towards cost-effectiveness have been focused, similarly to conventional PV, on minimizing the final installation price per kWp. When using PV, the economic analysis and the afford ability assessment are generally focused on the energy payback time and on the return of investment [31]. The present study will base the analysis of a BIPV solution on the following innovative aspects beyond the state of art: • Multi-functionality. BIPV is considered as a building material, which replaces a conventional layer of the building skin, thus providing a cost saving due to the cost of the replaced building cladding; • Unitary concept of BIPV and building envelope. The energy and economic assessments are conducted considering the BIPV system together with the whole building, as a unique system.

•
Life-cycle costing. The balance of costs and benefits is extended to the entire life cycle of both PV and the building envelope.

•
A project-based approach. The analysis is performed for a specific case study in which characteristics and peculiarities are known and provided by real data by avoiding uncertainties on simulated cost and energy parameters. Accordingly, the paper will focus the analysis on the energy performance and the cost effectiveness of the BIPV and BAPV systems installed in the case study building.

Description of the Case Study: Multi-Family Building with a Fully-Active Building Skin
The following investigation will be based on a reference case study of an 8-storey multi-family house built in 1965 and renovated in 2013, which has been equipped with an integrated photovoltaic system on all façades and roofs, in addition to a single-axis solar tracker. The building, which represented a pilot example in Switzerland for building renovation according to a plus-energy target and with a completely active skin, is located in Chiasso ( Figure 1). The building energy demand index before retrofit was 291 kWh/m 2 y with a total energy yearly demand of 399.5 MWh.
After renovation, according to Passivhaus and Minergie protocols, the energy consumption was 56 kWh/m 2 y. Electrical loads in the current state were 77 366 kWh, which was at the base of the load profile considered in the following. Currently, the whole electricity production of the whole PV installations, consisting of about 16 381 kWh/y for the facades and 33 105 kWh/y for the roofing (roof and garage) and the tracker, was injected into the grid and was subjected to a Feed in Tariff (FiT) remuneration scheme. The FiT is an energy supply policy. Long-term (10-25 years) purchase agreements support the development of new renewable energy projects [32].
PV modules installed on the façade were installed as a cold ventilated system supported by a metallic structure with an air gap of 27 mm between the cladding and the thermal insulation layer behind. Façade modules were divided into glass-glass opaque amorphous silicon for about 35 kWp in all orientations; glass-glass opaque mono-crystalline Silicon for about 12 kWp in the south orientation; semi-transparent glass-glass modules for balconies for about 6 kWp in the east and west orientation. Details are displayed in Table 1.

Assumptions and Methodology
In this section, the necessary steps and key elements that permitted us to reach the results of the analysis are explained. The chapter is divided into three parts considering the energy aspect (5.1), method and aspect of analysis (5.2) and the assessment of the optimal scenario (5.3).

Load Profile
The building load profile is made up of space heating (SH), domestic hot water (DHW) and appliances. The DHW and SH is produced with a heat pump of 33 kW. In addition, a solar thermal plant of 46.5 m 2 installed on the roof produces yearly, according to design, 26′272.5 kWh of thermal energy (net income evaluated of 565 kWh/m 2 ). Because of the lack of hourly-based monitored consumption data, the load profile was calculated through the aggregated data given by the Distribution System Operator (DSO) and the energetic certification (Minergie ® certificate). The hourlybased load profile was determined as explained in the following.   cSi  33  3  576  1 671  Balcony W  90  110  cSi  35  3  116  441   Total  652  53  311  16 381   Garage  10  20  cSi  70  15  660  9 699  Roof  10  20  cSi  48  10  1 216  12 155  Tracker  --cSi  58  12  938  11 251   Total  176  37  902  33 105

Assumptions and Methodology
In this section, the necessary steps and key elements that permitted us to reach the results of the analysis are explained. The chapter is divided into three parts considering the energy aspect (5.1), method and aspect of analysis (5.2) and the assessment of the optimal scenario (5.3).

Load Profile
The building load profile is made up of space heating (SH), domestic hot water (DHW) and appliances. The DHW and SH is produced with a heat pump of 33 kW. In addition, a solar thermal plant of 46.5 m 2 installed on the roof produces yearly, according to design, 26 272.5 kWh of thermal energy (net income evaluated of 565 kWh/m 2 ). Because of the lack of hourly-based monitored consumption data, the load profile was calculated through the aggregated data given by the Distribution System Operator (DSO) and the energetic certification (Minergie ® certificate). The hourly-based load profile was determined as explained in the following.
The thermal load for SH (Qsh) and for the DHW (Qhw) changed in electric load considering the hourly external temperature that influences the COP of the heat pump installed. The contribution Energies 2020, 13, 3827 7 of 23 of solar thermal was taken into account, as a DHW and SH thermal load share reduction (Qsh and Qhw). As reported on the energetic certification, the reduction corresponded to 74% for the DHW and 41% for the SH. The remaining 59% of the SH was produced with a heat pump during the 24 h of the day. The remaining 26% of DHW was produced with a heat pump every day for six hours, from 0 until 6 am. The thermal load profile Qsh and Qhw is shown in the Figure 2. Furthermore, a monthly correction factor of the heating produced with the heat pump according with the influence of the solar thermal seasonal production was individuated. The temperature of the hot water was assumed to be 58 • C and the indoor temperature was assumed to be 20 • C, from SIA 380/1. The supply heating temperature was assumed to be 30 • C.
Energies 2020, 13, 3827 7 of 23 The thermal load for SH (Qsh) and for the DHW (Qhw) changed in electric load considering the hourly external temperature that influences the COP of the heat pump installed. The contribution of solar thermal was taken into account, as a DHW and SH thermal load share reduction (Qsh and Qhw). As reported on the energetic certification, the reduction corresponded to 74% for the DHW and 41% for the SH. The remaining 59% of the SH was produced with a heat pump during the 24 h of the day. The remaining 26% of DHW was produced with a heat pump every day for six hours, from 0 until 6 am. The thermal load profile Qsh and Qhw is shown in the Figure 2. Furthermore, a monthly correction factor of the heating produced with the heat pump according with the influence of the solar thermal seasonal production was individuated. The temperature of the hot water was assumed to be 58°C and the indoor temperature was assumed to be 20°C, from SIA 380/1. The supply heating temperature was assumed to be 30°C.
The remaining electric consumptions (mechanical ventilation, lift, lighting and appliances) were m   The remaining electric consumptions (mechanical ventilation, lift, lighting and appliances) were merged, and an hourly load profile of a typical day was generated. The hourly load profile generated was obtained from the weighted hourly consumption of the 365 days of the year considering five typical profiles of five different families simulated with the software Load Profile Generator [33] Figure 3. The thermal load for SH (Qsh) and for the DHW (Qhw) changed in electric load considering the hourly external temperature that influences the COP of the heat pump installed. The contribution of solar thermal was taken into account, as a DHW and SH thermal load share reduction (Qsh and Qhw). As reported on the energetic certification, the reduction corresponded to 74% for the DHW and 41% for the SH. The remaining 59% of the SH was produced with a heat pump during the 24 h of the day. The remaining 26% of DHW was produced with a heat pump every day for six hours, from 0 until 6 am. The thermal load profile Qsh and Qhw is shown in the Figure 2. Furthermore, a monthly correction factor of the heating produced with the heat pump according with the influence of the solar thermal seasonal production was individuated. The temperature of the hot water was assumed to be 58°C and the indoor temperature was assumed to be 20°C, from SIA 380/1. The supply heating temperature was assumed to be 30°C.
The remaining electric consumptions (mechanical ventilation, lift, lighting and appliances) were merged, and an hourly load profile of a typical day was generated. The hourly load profile generated wa

PV Production Profile
The PV production profile was obtained by monitoring the energy production of Palazzo Positivo, with sensors downstream of each string inverter (AC current). The sensors detected the hourly electric production of each inverter, the temperature of the environment and modules and the incident radiation. Figures 4 and 5 show the profile of PV production and electric consumption generated of the typical summer and winter day. The PV production profile was obtained by monitoring the energy production of Palazzo Positivo, with sensors downstream of each string inverter (AC current). The sensors detected the hourly electric production of each inverter, the temperature of the environment and modules and the incident radiation. Figures 4 and 5 show the profile of PV production and electric consumption generated of the typical summer and winter day.

PV Production Profile
The PV production profile was obtained by monitoring the energy production of Palazzo Positivo, with sensors downstream of each string inverter (AC current). The sensors detected the hourly electric production of each inverter, the temperature of the environment and modules and the incident radiation. Figures 4 and 5 show the profile of PV production and electric consumption generated of the typical summer and winter day.

Method and Steps of Analysis
As shown in Figure 6, this study was conducted in three steps: (i) step 1: breakdown and recomposition of the active building skin; (ii) step 2: economic analysis; (iii) step 3: comparison after the application of a storage system. The starting point of this step is a breakdown of the building envelope of Palazzo Positivo. Each main functional element of the PV building skin is isolated and identified as a group, called a typological cluster (TC), according to building component (façade, railing and roof), orientation (north, south, east, west) and PV technology (a-Si, c-Si). This permitted us to generate a mapping of the building envelope by BIPV functional parts useful for the analysis (Figure 7).  The starting point of this step is a breakdown of the building envelope of Palazzo Positivo. Each main functional element of the PV building skin is isolated and identified as a group, called a typological cluster (TC), according to building component (façade, railing and roof), orientation (north, south, east, west) and PV technology (a-Si, c-Si). This permitted us to generate a mapping of the building envelope by BIPV functional parts useful for the analysis (Figure 7).

Step 1: Breakdown and Recomposition of the Active Building Skin Façade Layout
The starting point of this step is a breakdown of the building envelope of Palazzo Positivo. Each main functional element of the PV building skin is isolated and identified as a group, called a typological cluster (TC), according to building component (façade, railing and roof), orientation (north, south, east, west) and PV technology (a-Si, c-Si). This permitted us to generate a mapping of the building envelope by BIPV functional parts useful for the analysis (Figure 7). TCs were combined with each other in order to recompose groups of PV elements, called merged clusters (MC). MCs represented new configuration of PV systems thanks to a combination of different building skin parts, as shown in Table 2, so that it was possible to analyze a range of different solutions. MC1 was a union of each TC; it represented the actual state (all PV plants). The retribution of the MC1 did not represent the current state of the building (see MC10). MC2 was composed of only façade BIPV systems without roofs, railing applications and without the north oriented facade. MC3 stood for the BIPV solution with all four facades, namely MC2 including the northern facade. MC4 represented a combination of BIPV and roof BAPV systems; it was composed of BIPV modules on the four façades in addition to PV panels on the roof and the garage. MC5 stood for a conventional rooftop PV installation. MC6 was composed of only the PV modules with a final yield higher than 400 kWh/kWp (e.g., thus excluding some portions of facades non optimally exposed). MC7 considered BIPV on facades (apart from a-Si part installed on the south façade and the north facade) TCs were combined with each other in order to recompose groups of PV elements, called merged clusters (MC). MCs represented new configuration of PV systems thanks to a combination of different building skin parts, as shown in Table 2, so that it was possible to analyze a range of different solutions. MC1 was a union of each TC; it represented the actual state (all PV plants). The retribution of the MC1 did not represent the current state of the building (see MC10). MC2 was composed of only façade BIPV systems without roofs, railing applications and without the north oriented facade. MC3 stood for the BIPV solution with all four facades, namely MC2 including the northern facade. MC4 represented a combination of BIPV and roof BAPV systems; it was composed of BIPV modules on the four façades in addition to PV panels on the roof and the garage. MC5 stood for a conventional rooftop PV installation. MC6 was composed of only the PV modules with a final yield higher than 400 kWh/kWp (e.g., thus excluding some portions of facades non optimally exposed). MC7 considered BIPV on facades (apart from a-Si part installed on the south façade and the north facade) and roofs apart from the railings and the tracker. MC8 was the complete active skin excluding the garage, the tracker, the north facade and all the accessories such as balconies. All the MCs are reported in Table 2.
Furthermore, the other three MCs that considered special solutions were added in the analysis: MC9, MC10 and MC11. MC9 considered a solution without PV, namely the alternative solution of the building without any active surface where the electric energy was fully withdrawn from the grid. MC10 and MC11 corresponded to the MC1 and MC8 with the Feed in Tariff retribution rather than the self-consumption model. MC10 represented the retribution and the solar configuration that represented the current state of the building.

Step 2: Life-Cycle Cost Analysis
The analysis consisted of the cost effectiveness in installing PV systems according to the MC previously defined. The main analysis parameters were total cost of energy, yearly average revenues, self-consumption rate, self-sufficiency rate and cost/extra cost of PV.
The Total Cost of Energy (TCE) in a certain period of time was the main parameter of this analysis and represented the entire actualized amount of costs incurred during the lifetime of the solar system [Swiss Francs-CHF]. TCE included the initial cost (investment) to install the solar and/or the battery system, the energy costs to pay during the operation due to building energy consumption, the cost of maintenance and the revenues obtained due to PV self-consumption. Within the TCE, both the building skin system (cladding and structures) and the BOS costs were included, with the cost to replace the inverter after 10 years. TCE did not consider the increase in the energy cost over years and possible subsidies on the investment. The Equation (1) expresses the TCE calculated as the present value of all the annuity costs and revenues.
where: This was assumed as the extra cost to make a similar non-active façade active (in order to specify that, in most cases, the BIPV envelope is an active cladding installed in substitution of a similar non active system with the same substructures. Thus, the cost considered in this paper as the initial investment was defined as the extra cost compared to a traditional non-active system.). Included within the extra cost was the BOS (707 CHF/kWp) and the cost of the new inverter substituted after 10 years (312 CHF/kWp). This resulted from real examples analyzed from the Swiss market. For the BAPV element the extra-cost was assumed as the cost of the plant since it did not replace any building material [CHF].
The revenue of the PV system was considered as a benefit. The remaining energy taken from the grid and the extra cost of the BIPV system were considered as negative factors. According to the electric society, the energy consumed outright cost 0.19 CHF/kWh, energy injected into the grid cost 0.07 CHF/kWh. In the MC9, the solution without PV, the energy withdrawn from the grid cost, as an assumption, 0.19 CHF/kWh. Considering the MC10 and MC11 with a Feed in Tariff retribution, Energies 2020, 13, 3827 11 of 23 the energy injected into the grid cost 0.28 CHF/kWh (Age SA retribution during the year 2013). The prices referred to the year 2013, when the PV system was installed.
The cost of the a-Si cladding module installed on the façade of Palazzo Positivo was 110 CHF/m 2 . As also mentioned by P. Bonomo et al. [34], if we considered the cost of a conventional laminated glass cladding 80 CHF/m 2 we could define 30 CHF/m 2 as the extra cost for our active BIPV glazed cladding. Other costs and extra costs are showed in Table 3, for all the claddings used in the case study. The costs were actualized in 2020 for the analysis (t = 1). Analyses were extended to each MC considering a period of calculation of about 20 years. Along with the analysis explained above, Scenario 2 considered the introduction of a storage system, in order to analyze the ratio cost/benefit of a different alternative of technology implementation in the building. The general goal of optimization is to introduce batteries to increase self-consumption and reduce the TCE through storage.
In Scenario 1, the current load profile was compared with Scenario 2's modified load profile with electrical storage implementation.

Assessment of Optimal Scenarios
In the following section, the two scenarios that characterized the analysis of the study are explained. The difference between Scenario 1 and Scenario 2 is the introduction of batteries that permit the storage of electric energy during the day. This variation affects the value of the parameter TCE, explained on Section 5.2.2.

Scenario 1: Current Load Profile
In this scenario, the current load profile of the building was used. A total of 26% of the DHW and 59% of the SH were generated from the heat pump. DHW was produced during the night, from 0 until 6 am. Thus, Scenario 1 represented the current solution implemented in the building.

Scenario 2: Electrical Storage Implementation
In this scenario, a strategy of maximizing self-sufficiency and self-consumption was pursued by considering the presence of batteries. A complementary strategy to optimize the energy demand of the building did not consider the shift of the electric loads during the hours of PV production. Due to the shortage of hourly values of the whole electric loads, it was decided to show only the current configuration with the storage implementation solution.
The proposed approach for the evaluation was aimed at • optimal sizing of the battery capacity, on the basis of the output values of the self-sufficiency and self-consumption and of the values electricity input and output from the grid; • cost-benefit analysis of different battery capacities and choice of the optimal size; • comparison of the optimized scenario with Scenario 1 in terms of SSR, SCR and cost-effectiveness analysis.
Energies 2020, 13, 3827 12 of 23 The first step involved the optimal sizing of the battery capacity (kWh) according to the total annual electrical loads of the building (Equation (2)). The battery capacity (Cbat) was then established as a percentage of the electrical load, varying from a value of 10% up to a maximum of 100%. where: • R i and E load stand for capacity rate and daily energy consumption (kWh), respectively. Further input values that we considered are the following: • charging/discharging efficiency of the battery equal to 0.95, • Depth of Discharge (DOD) equal to 0.80.
From an economic point of view, the cost-benefit analysis was conducted considering the following assumption: • use of Lithium-Ion Batteries; • fixed cost of about 339 USD/kWh (ca. 387 CHF-August 2018), plus about 33% of the cost for labor, without considering the cost of the inverters (forecast for 2020) [35]; • life cycle of 13.6 years [34] was rounded off to 14 years for convenience; • annual operation and maintenance of the battery system 1.5% of the investment cost [19].
However, these data are constantly evolving. The future trend is in fact a drastic reduction in the installation costs of the batteries (equal to about half the cost for 2030) and an increase in the life cycle of the batteries (equal to about 1/3 more than the current life cycle for 2030). [35] In order to consider the effects of the battery, the TCE, after 20 years, is slightly adapted from the previous one, mentioned in Section 5.2.2 In this case, the cost of running energy and the revenues included in TCE were split into during the battery lifetime (assumed 14 years, with storage) and after the battery lifetime (for the remaining 6 years, without storage). The Equation (3) where the new parameters represent: where: • dr: degradation rate of the PV, assumed 0.8%/year;

Analysis and Results
In this section the results concerning the technical-economic assessment, according to the assumptions and reference scenarios described in the previous chapters, are presented and discussed. The goal was to bring to light how the different MCs affected the energy and cost-effectiveness of the whole building in different ways. To make the results more clear, the discussion and results were grouped in the thematic analysis as follows: • energy and cost effectiveness, • comparison between BAPV and BIPV scenarios, • business model alternative to self-consumption: FiT, • electrical storage implementation: Scenario 2.
As abovementioned, the outcomes were the results of the method actualized for a specific case study. In Table 4, an overview of the analysis after 20 years, considering Scenario 1, is shown. It is worthy to note the position of MC9 (building without PV) and the current status of MC1.

Energy and Cost Effectiveness
The quantification of the energy effectiveness of each MC is represented within this paragraph by the SSR, namely the amount of energy produced with a renewable source and self-consumed compared with the global amount of energy needed during the year for the building. This value is directly related to the PV energy production value.
However, as emerging from Figure 8, a high value of SSR itself did not entail a low value of TCE during the solar system lifetime (in this study 20 years; the lifetime value was an approximated value based on the average performance guarantee on 80% of nominal power under standard test conditions of some examples of PV modules of different brands and different technologies. The lifetime of a PV system could be longer). The MC1, representing the current state of the solar system installed on the building was the solution with the higher SSR. However, in terms of TCE, it did not represent the most affordable solution.
Energies 2020, 13, 3827 14 of 23 However, as emerging from Figure 8, a high value of SSR itself did not entail a low value of TCE during the solar system lifetime (in this study 20 years; the lifetime value was an approximated value based on the average performance guarantee on 80% of nominal power under standard test conditions of some examples of PV modules of different brands and different technologies. The lifetime of a PV system could be longer). The MC1, representing the current state of the solar system  For the MC9, MC10 and MC11, the SSR was zero because of the total injection of the energy produced into the grid.

Comparison between BAPV and BIPV Scenarios
Three different MCs were compared in order to evaluate the effectiveness of building integrated/added PV systems. The analysis on BAPV (MC5) and BIPV (MC3) systems showed typical aspects. The BAPV systems had a high value of PV energy production, corresponding with high yearly average revenues and SSR, which, despite the extra cost of the system (modules, cables, structures, etc.), was high. Namely, the extra cost used in the calculation corresponded to the entire PV plant cost, since in a BAPV system no building system was replaced. On the other hand, the BIPV system, MC3, had a restrained energy production due to the non-optimal exposition, but a lower initial extra-cost as indicated in Table 5.
The most affordable solution in terms of TCE was represented by the MC8, the solution combining BAPV on the roof and BIPV on facades where the TCE was significantly lower. Within this analysis, the MC5 (only BAPV) was more affordable than other BIPV or mixed BIPV/BAPV solutions.

Business Model Alternative to Self-Consumption: FiT
The MC9, the solution without any PV on the building, was not convenient, according with the TCE, in comparison with more than half of the MC due to the high price of the energy purchased   For the MC9, MC10 and MC11, the SSR was zero because of the total injection of the energy produced into the grid.

Comparison between BAPV and BIPV Scenarios
Three different MCs were compared in order to evaluate the effectiveness of building integrated/added PV systems. The analysis on BAPV (MC5) and BIPV (MC3) systems showed typical aspects. The BAPV systems had a high value of PV energy production, corresponding with high yearly average revenues and SSR, which, despite the extra cost of the system (modules, cables, structures, etc.), was high. Namely, the extra cost used in the calculation corresponded to the entire PV plant cost, since in a BAPV system no building system was replaced. On the other hand, the BIPV system, MC3, had a restrained energy production due to the non-optimal exposition, but a lower initial extra-cost as indicated in Table 5. The most affordable solution in terms of TCE was represented by the MC8, the solution combining BAPV on the roof and BIPV on facades where the TCE was significantly lower. Within this analysis, the MC5 (only BAPV) was more affordable than other BIPV or mixed BIPV/BAPV solutions.

Business Model Alternative to Self-Consumption: FiT
The MC9, the solution without any PV on the building, was not convenient, according with the TCE, in comparison with more than half of the MC due to the high price of the energy purchased from the grid. However, in comparison with the current configuration, both in self-consumption (MC1) and FiT (MC10), the TCE of the MC9 was about 20% and 15% lower, respectively.
The MC with Feed in Tariff (MC10 and MC11), showed different results. Only MC11 showed a low value of TCE, which represents a good convenience of this business model ( Table 6). The energetic benefit in terms of SSR was not calculated because the energy produced with PV was all injected into the grid. In spite of the low value of TCE, it is necessary to underline that the FiT for the energy injected into the grid was considered 0.28 CHF/kW, as paid by the electric society in 2013 for the building, according to Swiss FiT skills. Nowadays, it is important to remark that the subsidies are no longer available in most EU countries. Self-consumption also shows feasible scenarios in a non-subsidized situation. In this paragraph, the scenario concerning the storage implementation is discussed. Within the first part, we will provide an insight on the correlation between the batteries' capacity and the effect on building self-consumption and self-sufficiency, along with the main assumptions adopted for this scenario, with the goal to also define an optimal size of storage.

Battery Implementation. Preliminary Assumptions and Results for MC1
The values of both SCR and SSR and of electric injected and withdrawn energy from the grid were calculated and the output values of the various capacity scenarios were compared, with the goal to assess the effectiveness of battery implementation, as shown in Figures 9 and 10 (for the MC1).  Tariff (MC10 and MC11), showed different results. Only MC11 showed a low value of TCE, which represents a good convenience of this business model ( Table 6). The energetic benefit in terms of SSR was not calculated because the energy produced with PV was all injected into the grid. In spite of the low value of TCE, it is necessary to underline that the FiT for the energy injected into the grid was considered 0.28 CHF/kW, as paid by the electric society in 2013 for the building, according to Swiss FiT skills. Nowadays, it is important to remark that the subsidies are no longer available in most EU countries. Self-consumption also shows feasible scenarios in a nonsubsidized situation. Table 6. FiT models.

Electrical Storage Implementation: Scenario 2
In this paragraph, the scenario concerning the storage implementation is discussed. Within the first part, we will provide an insight on the correlation between the batteries' capacity and the effect on building self-consumption and self-sufficiency, along with the main assumptions adopted for this scenario, with the goal to also define an optimal size of storage. 5% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Injected/Withdrawn energy (kWh/y)

Cbat % SSR % SCR % From Grid kWh/y To Grid kWh/y
This means that, for this size of PV system which included all building skin surfaces, the energy produced can be consumed almost totally, with the use of optimal accumulation systems, while selfsufficiency can reach a maximum percentage of about 58%.
Even the values of injection and withdrawal from the grid had an asymptotic profile towards a minimum limit, remaining almost unchanged from the achievement of 50% of storage capacity.
From this point of view, the capacity of the battery which optimized the values of SSR, SCR, energy input and energy taken from the grid is considered equal to 50% of the total daily electrical load, as already found in other studies in the literature [1,9,10,12,13,18,19,21]. In addition, in this case, the optimization method of the battery capacity took into account the TCE value, also becoming a cost-optimal sizing. For each scenario, in fact, Step 2 was developed to compare the different MCs according to the main parameters: PV energy production, final yield, PbP and TCE after specific years. The final yield stood for the ratio between the energy produced and the nominal power of the PV system (kWh/kWp). SSR SCR Figure 10. Comparison of the various storage capacity and the SSR and self-consumption rate (SCR) values (MC1).
In Figure 9 the correlation between the storage capacity and the resulting energy injected into the grid and withdrawn is reported. The graph of the SSR and SCR ( Figure 10) shows a non-linearity between the storage capacity and the SCR and SSR, resulting, instead, in an asymptotic trend towards a limit value as the storage capacity increases. In fact, when a capacity of about 50% of the load profile was reached, further capacity of storage led to negligible increases in the values of SSR and SCR. Oversized batteries, in fact, could not be fully charged or discharged. As storage capacity increased, the SCR varied from a minimum of 54% to a maximum of 90%, while the SSR varied from a minimum of 35% to a maximum of 58% (Table 7).  0  35  54  50 638  22 758  10  41  63  46 428  18 099  20  45  71  42 982  14 288  30  50  78  39 937  10 921  40  54  84  37 239  7 937  50  56  87  35 670  6 206  60  57  89  35 082  5 561  70  57  89  34 845  5 305  80  57  90  34 678  5 127  90  58  90  34 526  4 965  100  58  90  34 401  4 833 This means that, for this size of PV system which included all building skin surfaces, the energy produced can be consumed almost totally, with the use of optimal accumulation systems, while self-sufficiency can reach a maximum percentage of about 58%.
Even the values of injection and withdrawal from the grid had an asymptotic profile towards a minimum limit, remaining almost unchanged from the achievement of 50% of storage capacity.
From this point of view, the capacity of the battery which optimized the values of SSR, SCR, energy input and energy taken from the grid is considered equal to 50% of the total daily electrical load, as already found in other studies in the literature [1,9,10,12,13,18,19,21]. In addition, in this case, the optimization method of the battery capacity took into account the TCE value, also becoming a cost-optimal sizing. For each scenario, in fact, Step 2 was developed to compare the different MCs according to the main parameters: PV energy production, final yield, PbP and TCE after specific years. The final yield stood for the ratio between the energy produced and the nominal power of the PV system (kWh/kWp). The results of the analysis are reported in Table 8 that shows the comparative values of SSR, SCR and TCE, considering the period of 20 years calculated for the MC1. In conclusion, the optimal sizing of the battery, from an economic point of view (TCE) and for the MC1, would be assuming Cbat = 0.3 E load , as Figure 11 shows. The total costs over 20 years included  initial investment costs for the PV plant,  battery installation costs,  PV revenues for 20 years,  storage revenues for 14 years.
The results of the analysis are reported in Table 8 that shows the comparative values of SSR, SCR and TCE, considering the period of 20 years calculated for the MC1. In conclusion, the optimal sizing of the battery, from an economic point of view (TCE) and for the MC1, would be assuming Cbat = 0.3 Eload, as Figure 11 shows.   The analysis showed that for half of the cases analyzed, a storage system, with a capacity included between 10% and 30%, will be affordable by 2020 (MC1, MC4, MC6, MC7). These MCs represented the configurations with higher PV production. Within these cases, an improvement of the SCR permitted the self-consumption of a higher amount of energy in comparison with the MC where the energy production was low. As possible to notice from Figure 13, the MC1, MC4, MC6 and MC7 showed a low SCR at low capacity of the storage system, while, considering the high PV production, the SSR was the highest compared with the other analyzed MC. However, the installation of a storage system, also for the MC where batteries reduce the TCE, entailed low benefits considering the reduction in TCE in 20 years. A further reduction in the batteries' cost would permit the considerable reduction in the TCE. Although in these conditions and for this case study a storage system is not recommended, in a few years the batteries could be a solution that permits the reduction in the costs of a PV system.
The "energy" comparison in terms of SSR and SCR (Figures 13 and 14) showed that, except for the MC2 and MC3, where the SCR was already 100%, in all cases, the battery increased the values of self-sufficiency by about 30% at least, and the values of SSR by about 15%, averagely. The maximum percentage increase was in MC6, by about 38%, in terms of SCR, and by about 21% in terms of SSR. In MC2, MC3, MC4, MC5, MC7 and MC8, the SCR reached the value of 100%.  The analysis showed that for half of the cases analyzed, a storage system, with a capacity included between 10% and 30%, will be affordable by 2020 (MC1, MC4, MC6, MC7). These MCs represented the configurations with higher PV production. Within these cases, an improvement of the SCR permitted the self-consumption of a higher amount of energy in comparison with the MC where the energy production was low. As possible to notice from Figure 13, the MC1, MC4, MC6 and MC7 showed a low SCR at low capacity of the storage system, while, considering the high PV production, the SSR was the highest compared with the other analyzed MC. However, the installation of a storage system, also for the MC where batteries reduce the TCE, entailed low benefits considering the reduction in TCE in 20 years. A further reduction in the batteries' cost would permit the considerable reduction in the TCE. Although in these conditions and for this case study a storage system is not recommended, in a few years the batteries could be a solution that permits the reduction in the costs of a PV system. The analysis showed that for half of the cases analyzed, a storage system, with a capacity included between 10% and 30%, will be affordable by 2020 (MC1, MC4, MC6, MC7). These MCs represented the configurations with higher PV production. Within these cases, an improvement of the SCR permitted the self-consumption of a higher amount of energy in comparison with the MC where the energy production was low. As possible to notice from Figure 13, the MC1, MC4, MC6 and MC7 showed a low SCR at low capacity of the storage system, while, considering the high PV production, the SSR was the highest compared with the other analyzed MC. However, the installation of a storage system, also for the MC where batteries reduce the TCE, entailed low benefits considering the reduction in TCE in 20 years. A further reduction in the batteries' cost would permit the considerable reduction in the TCE. Although in these conditions and for this case study a storage system is not recommended, in a few years the batteries could be a solution that permits the reduction in the costs of a PV system.    Figure 13. Comparison between the capacity of the storage system and the SCR for the MC.
The "energy" comparison in terms of SSR and SCR (Figures 13 and 14) showed that, except for the MC2 and MC3, where the SCR was already 100%, in all cases, the battery increased the values of self-sufficiency by about 30% at least, and the values of SSR by about 15%, averagely. The maximum percentage increase was in MC6, by about 38%, in terms of SCR, and by about 21% in terms of SSR. In MC2, MC3, MC4, MC5, MC7 and MC8, the SCR reached the value of 100%.

Conclusions
The paper presented and discussed the results of a comparative analysis on a multifamily building, focusing on alternative scenarios for energy and economic effectiveness of building skin retrofit with PV parts. By assuming real monitored data for the case study, we analyzed and compared, in a life-cycle standpoint, the main construction and energy related design strategies affecting cost and energy effectiveness of the whole building. The main parameters relevant for BIPV skin design, including both construction aspects (orientation, surface typology) and energy strategies (such as self-consumption/sufficiency, storage and combination with BAPV) were highlighted and discussed in dedicated scenarios. This analysis permitted us to define both approaches and conclusions that are scalable for other case studies to correlate in detail the construction/architectural choices, typical of the architectural stage, with the energy and cost results which are crucial in order to establish the intervention feasibility for BIPV systems.
In Scenario 1 (current building load profile without storage) the following conclusions emerged: Affordable values of TCE can also be reached with only BIPV or BAPV (MC2 and MC5, respectively). It shows, once again, that cost effectiveness is reached through a balanced combination of final yield, investment cost, SSR, SCR and revenues.  Energy and cost-effectiveness: The most effective solar solutions from energetic and economic perspective are represented by building skin configurations in which the accessories (mainly trackers and balconies) are not used. This can be explained due to the high investment cost and the low irradiation.  The cost-effectiveness is higher by about 25% in BAPV solutions MC5, MC7 and MC8 (on the rooftop and garage) due to the lower investment cost in comparison with the building state of the art (MC10). However, it has to be noticed that self-sufficiency is very low due to the limited rooftop area so that the scenario is not relevant from a nearly-zero energy building perspective.  The solution without any PV surface in the building skin (MC9), even though it does not correspond to the target of nearly-zero energy building with on-site energy production, is about 15% more convenient than MC10 (state of the art) but not convenient in comparison with most of the optimized solutions. This marks the need for optimizing building skin design with costs and building energy needs.  Solutions with only BIPV surfaces on the façade (MC2 and MC3) are average solutions in terms of cost effectiveness. Although the final yield of these solutions is low, the investment cost (ECBIPV) of the PV installation is also low, therefore supporting their cost convenience.

Conclusions
The paper presented and discussed the results of a comparative analysis on a multifamily building, focusing on alternative scenarios for energy and economic effectiveness of building skin retrofit with PV parts. By assuming real monitored data for the case study, we analyzed and compared, in a life-cycle standpoint, the main construction and energy related design strategies affecting cost and energy effectiveness of the whole building. The main parameters relevant for BIPV skin design, including both construction aspects (orientation, surface typology) and energy strategies (such as self-consumption/sufficiency, storage and combination with BAPV) were highlighted and discussed in dedicated scenarios. This analysis permitted us to define both approaches and conclusions that are scalable for other case studies to correlate in detail the construction/architectural choices, typical of the architectural stage, with the energy and cost results which are crucial in order to establish the intervention feasibility for BIPV systems.
In Scenario 1 (current building load profile without storage) the following conclusions emerged: • Comparison between BAPV and BIPV: the most affordable solution in term of TCE is represented by the MC8, the solution combining BAPV on the roof and BIPV on facades, where the high value of final yield in roof is combined with the low extra cost in façade.

•
Affordable values of TCE can also be reached with only BIPV or BAPV (MC2 and MC5, respectively). It shows, once again, that cost effectiveness is reached through a balanced combination of final yield, investment cost, SSR, SCR and revenues. • Energy and cost-effectiveness: The most effective solar solutions from energetic and economic perspective are represented by building skin configurations in which the accessories (mainly trackers and balconies) are not used. This can be explained due to the high investment cost and the low irradiation.

•
The cost-effectiveness is higher by about 25% in BAPV solutions MC5, MC7 and MC8 (on the rooftop and garage) due to the lower investment cost in comparison with the building state of the art (MC10). However, it has to be noticed that self-sufficiency is very low due to the limited rooftop area so that the scenario is not relevant from a nearly-zero energy building perspective.

•
The solution without any PV surface in the building skin (MC9), even though it does not correspond to the target of nearly-zero energy building with on-site energy production, is about 15% more convenient than MC10 (state of the art) but not convenient in comparison with most of the optimized solutions. This marks the need for optimizing building skin design with costs and building energy needs.

•
Solutions with only BIPV surfaces on the façade (MC2 and MC3) are average solutions in terms of cost effectiveness. Although the final yield of these solutions is low, the investment cost (EC BIPV ) of the PV installation is also low, therefore supporting their cost convenience.
• Business models with FiT: if compared with the self-consumption business model, subsidies with FiT, that are no longer available in most EU countries, would support a more convenient scheme (MC1 vs MC10).
In Scenario 2 (including storage) the main conclusions are • Optimal storage size: a correct sizing of the battery capacity can make significant increases to the values of self-consumption and self-sufficiency, as well as minimizing the values of power flows between the household and the grid.

•
However, an excessive increase in the battery capacity corresponds to insignificant increases in both self-consumption and self-sufficiency, and reductions in the values of energy injected and withdrawn from the public grid reaching for both horizontally asymptotic values. In addition, an oversized battery, where a large part of its capacity would remain unused, leads exclusively to an inconvenient increase in installation costs, without energy benefits, as we have seen in the figure that shows the comparison between the capacity of the storage system and the TCE for the MC.

•
Oversized solar systems benefit from the installation of a storage system with a capacity included between 10% and 30%. The installation of a battery permits a considerable increase in the amount of electricity self-consumed reducing the TCE (MC1, MC4, MC6 and MC7). However, the benefits induced by a storage system are not worthwhile within a retrofitting effort. A further reduction in the batteries' cost would permit a considerable reduction in the TCE. Although in these conditions and for this case study a storage system is not recommended, in a few years the batteries could be a solution that permits the reduction in the costs of a PV system. No further links with the building construction in comparison with those found within Scenario 1 have been found.
This study provides a practical reference for researchers and professionals in finding the technical approaches and strategies supporting appropriate business models or at least a clearer value proposition project-related to BIPV systems. Outcomes permitted us to show, in a real case study, how BIPV design choices related to the building skin (including different building design options, surface orientations, technologies, price levels, etc.) are closely related to the specific building typology and energy system (HVAC, load profile, etc.) by directly affecting the total life cost, self-consumption/sufficiency and cost-effectiveness. It emerged as the two major components in view of the full decarbonization of the built environment are closely related. Retrofitting existing buildings to increase the energy efficiency is not independent from a transition to energy positive buildings producing electricity, covering their heating and cooling needs and contributing to the grid stability with sustainable, renewable energy technologies as part of the building skin design.
Further development will permit the extension of this methodology to other BIPV case studies, in order to perform a sensitivity analysis of the main parameters that influence the integration of solar systems in building skin and finding other patterns of the main technical approaches supporting appropriate value proposition BIPV systems.