Multi-Objective Optimization of a Folding Photovoltaic-Integrated Light Shelf Using Non-Dominated Sorting Genetic Algorithm III for Enhanced Daylighting and Energy Savings in Office Buildings

Abstract

This study developed a novel folding light shelf system that integrates reflectors, photovoltaic (PV) modules, and adaptive louvers that adjust based on solar altitude, aiming to improve daylight distribution, minimize glare, and reduce energy consumption in office buildings. The research employed an advanced optimization approach, utilizing Non-dominated Sorting Genetic Algorithm III (NSGA-III) and Latin Hypercube Sampling, a highly effective method suitable for managing complex multi-objective scenarios involving numerous variables, to efficiently identify high-performance configurations with increased precision. Key design variables across all three components of the system included angle, width, distance, and the number of folds in the light shelf, along with the number of louvers. The proposed method successfully integrates PV technology into light shelves without compromising their functionality, enabling both daylight control and energy generation. The optimization results demonstrate that the system achieved up to a 15% improvement in useful daylight illuminance (UDI) and a 16% reduction in cooling energy consumption. Furthermore, the PV modules generated 509.5 kWh/year, ensuring improved efficiency and sustainability in building performance.

1. Introduction

The global average temperature in 2024 exceeded the +1.5 °C threshold relative to pre-industrial baselines, reaching the highest levels ever recorded despite the increasing efforts to mitigate climate change [1]. In order to successfully limit global warming within safe boundaries, greenhouse gas (GHG) emissions in the atmosphere need to reach net-zero by 2050 as a central strategy [2]. Within this global context, energy conservation and emission reduction are crucial in international efforts to mitigate the impact of climate change [3,4]. The building sector is regarded as one of the largest consumers of energy and natural resources, thereby significantly impacting the environment [5]. According to the International Energy Agency (IEA), the building sector accounted for approximately 30% of global end-use energy consumption in 2021. In parallel, it was responsible for 15% of end-use sector CO₂ emissions [6]. Office buildings, in particular, play a major role, forming a substantial part of the commercial building category and contributing significantly to global energy use [7]. The high energy demand in office buildings arises mainly from lighting, heating, cooling, ventilation, and office equipment [8]. Electric lighting not only directly contributes to energy consumption but also increases cooling loads, accounting for up to 16% of the demand for cooling energy consumption. Cooling and lighting systems together can account for up to 50% of total electricity use in office buildings [9], highlighting the need to improve energy efficiency to reduce their environmental impacts [6,7]. In office environments, where occupants have limited control over their visual surroundings, achieving proper lighting while avoiding under- or over-lit conditions is challenging [10].

Unlike traditional single-objective optimization approaches, the design and optimization of facade shading systems must simultaneously consider both interior daylight and thermal performance, two factors that are often in conflict, where improving one may negatively impact the other [11]. For instance, a recent study demonstrated that increasing the window-to-wall ratio (WWR) to maximize daylight can significantly improve useful daylight illuminance (UDI) and reduce lighting energy consumption. However, this also resulted in an increase in cooling and heating energy due to increased solar heat gains and losses [12]. These findings confirm that increasing daylighting can lead to significant energy trade-offs, making it essential to balance daylight quality, glare control, and energy efficiency. Therefore, selecting an appropriate daylighting system is essential, not only for minimizing reliance on artificial lighting but also for controlling energy consumption [8,9]. To address these challenges, a wide range of daylighting systems have been developed, each offering distinct performance benefits. Systems like light shelves and light pipes are highly effective in improving daylight penetration and reducing electric lighting demand [10,13,14], while systems such as louvers and Venetian blinds are more beneficial for controlling glare, particularly under high solar exposure [11,15,16]. Fiber optic systems provide precise daylight redirection with minimal thermal gain, making them suitable for interior spaces with strict comfort requirements [17]. By aligning the choice of daylighting system with specific performance objectives, building designs can achieve greater overall efficiency.

Literature Review

Recent studies have increasingly focused on optimizing shading systems to improve the balance between daylighting, thermal comfort, and energy efficiency. For instance, Xu et al. discovered that even in extremely cold regions, incorporating horizontal shading in high-rise residential buildings could reduce overheating by 18.5%, directly impacting thermal comfort and energy consumption [16]. Further research has explored the optimization of shading elements to acquire the most efficient solutions. A study in Tallinn, Estonia, demonstrated that UDI improved by nearly 90%, while the Energy Use Intensity (EUI) reduced by 30% [18]. Similarly, in Japan, research on optimizing an expanded-metal shading device led to the complete removal of Annual Sun Exposure (ASE) and a 50% increase in UDI [19]. In another case, an optimization model achieved a 14% increase in energy savings while preserving 50% of daylight availability [11]. Additionally, a genetic algorithm-based study demonstrated that optimized shading configurations significantly enhanced both visual and thermal comfort while also reducing overall energy demand [20].

Nowadays, advanced shading solutions go beyond simple overhangs and include more sophisticated strategies, like light shelves. A light shelf is a daylighting system mounted either inside or outside a window designed to reflect natural light deep into a room [21], as illustrated in Figure 1a,b. This system simultaneously enhances both visual comfort and energy efficiency in indoor environments.

Functionally, the inner part of the light shelf reflects and redirects sunlight deeper into the interior, promoting a homogeneous light distribution across the workspace. Meanwhile, the external part shades the glazing and minimizes glare [22]. Light shelf performance relies on several factors, including their geometry, construction and material, dimensions, inclination angle of exterior surfaces, and climatic conditions [15]. A study conducted in Cheonan, South Korea, optimized different light shelf configurations by adjusting depth and angle on a monthly basis, resulting in significantly improved lighting, especially in areas over five meters from the window, throughout all seasons [23]. In a related study in Iran, the researchers found that the optimized light shelf configurations not only improved daylighting but also led to substantial energy savings [24]. However, regardless of the depth or angle of the light shelf, there is always a noticeable amount of glare near the window. Therefore, supplementary systems are employed to mitigate and control this glare, as depicted in Figure 1c.

Recent research has also explored innovative advancements in light shelf design to improve indoor environmental quality. For instance, Zhao et al. developed a folding light shelf that reduced lighting energy consumption by an average of 49.5% and improved indoor light uniformity by 20.1% compared to traditional light shelves [25]. Light shelves have also been combined with photovoltaic (PV) modules, serving the dual function of shading the building and generating electricity [26]. A related study employed a multi-objective optimization approach to assess a PV-integrated light shelf system in a classroom, revealing a significant improvement in daylight, energy performance, and occupant satisfaction [27]. However, integrating PV modules with light shelves presents certain challenges. Since both components share the same angle when mounted, this configuration may not be ideal, as they require various optimal orientations [28]. For instance, Lee et al. evaluated different light shelf angles and PV module attachments under diverse lighting and thermal conditions, finding that indoor comfort is strongly influenced by the selection of angles that balance solar heat gain, daylighting efficiency, and energy generation [21]. In this regard, researchers enhanced both energy generation and daylighting by employing different structures like folding and curves with distinct angles for PV modules and light shelf reflectors [25,26]. A summary of recent studies on shading systems and their impact on building performance is presented in

Table 1. A literature summary of previous studies on shading devices and their building performance outcomes.

Reference	Year	Climate	Building Type	Shading Type	Simulation Tool	Performance Aspect
[29]	2016	hot and dry	Office	External Venetian blind	EnergyPlus	Useful Daylight Illuminance, Lighting Energy, HVAC, Energy Consumption
[11]	2019	Mediterranean	Office	Amorphous external shading	EnergyPlus	Total Energy Consumption, Useful Daylight Illuminance
[18]	2022	humid continental	Classroom	Exterior static Shadings	Rhinoceros and Grasshopper	Visual Comfort, Glare, Daylight, View Out, Energy Saving
[19]	2022	warm and temperate	Mock-up room	Exterior expanded-metal shading	Rhinoceros and Grasshopper	Annual Sunlight Exposure, Spatial Daylight Autonomy, Useful Daylight Illuminance, View
[30]	2015	humid subtropical	Office	Exterior louver	Rhinoceros and Grasshopper	Visual Comfort, Energy Efficiency
[31]	2023	hot semi-arid	Office	Exterior louver	Rhinoceros and Grasshopper	Building Energy Consumption, Thermal Comfort
[20]	2024	dry and cold semi-desert	Office	Exterior static Shadings	Rhinoceros and Grasshopper	Energy Use Intensity, Spatial Daylight Autonomy, Glare, Thermal Comfort
[14]	2020	humid subtropical	Mock-up office room	Internal–external light shelf	Experimental	Improving Depth of Daylight Penetration, Uniformity Ratio
[24]	2020	cold semi-arid	Residential tower	Internal light shelf	Rhinoceros and Grasshopper	Energy Consumption, Thermal Comfort
[25]	2024	hot summer, humid continental climates	Full-scale testbed	Light shelf with folding reflector	Experimental	Daylighting Performance, Lighting Energy Consumption
[26]	2019	humid continental	Full-scale test bed	Light shelf with photovoltaic	Experimental	Lighting Energy, Power Generation, Uniformity Ratio
[27]	2021	dry and cold semi-desert	Classroom	Light shelf with photovoltaic	Rhinoceros and Grasshopper	Daylight, Energy, Occupant Satisfaction
[28]	2022	humid continental	Full-scale testbed	Light shelves with photovoltaic	Experimental	Daylighting Energy, Indoor Uniformity Ratio, Glare
[15]	2022	humid continental	Full-scale testbed	Light shelves with photovoltaic	Experimental	Building Energy, Uniformity Ratio
[32]	2022	humid continental	Full-scale testbed	Light shelves with photovoltaic	Experimental	Improve Daylighting and Concentration Efficiency
[33]	2024	very hot	Office conference room	Trapezoidal profile louver shading	Rhinoceros and Grasshopper	Energy Saving, Daylighting Improvement
This Study	2025	hot and arid	Office	Light shelves with photovoltaic and louver	Rhinoceros and Grasshopper 1.0.0007 + Python Programming 3.13.5	Daylighting Performance, Glare, Energy Consumption, and Power Generation

.

Extensive research has addressed louver configurations, indicating that parameters such as slat depth, spacing, and quantity significantly influence their effectiveness, though their optimal values are dependent on contextual factors including latitude, building geometry, and façade orientation [13,32,33,34]. Nevertheless, despite progress in developing advanced daylighting and PV-integrated shading systems, a critical gap remains in the integrated optimization of dynamic, folding PV-integrated light shelves combined with lower-level glare-reducing louvers, particularly under realistic environmental conditions and multiple performance metrics. Most current studies have also been focused on static or limited multi-objective systems, typically optimizing up to two or three objectives [35,36,37,38,39].

Despite the effectiveness of light shelves in redirecting daylight through upper window parts, they often leave the lower window areas exposed, which can lead to occupant glare [40], as illustrated in Figure 1b. To address this issue, additional shading strategies are necessary for the lower sections of the building façade [41]. The present study introduces a novel, dynamic folding light shelf system that is composed of two primary modules, a lower one incorporating a glare-reducing louver system and an upper module featuring an innovative dynamic folding light shelf that integrates PV modules on the exterior-facing slats and highly reflective interior surfaces on the inner slats. To the best of our knowledge, the integration of NSGA-III and LHS has not yet been applied in the context of a dynamic daylighting system that simultaneously integrates PV generation. Furthermore, the proposed system offers a scalable solution for reducing building energy demand while improving indoor environmental quality. The system supports broader energy conservation goals and presents strong potential for integration into national energy codes, regulatory standards, and green building certification frameworks.

2. Materials and Methods

2.1. Research Workflow

This study systematically investigates a comprehensive overview of the proposed archetype and its underlying design methodology through a six-step approach (Figure 2). In the first step, the process begins with the development of the base geometry, followed by the parametric modeling of the folding light shelf and louver system. Key design variables are defined at this stage to enable flexibility in simulation and optimization. The study area and weather data (e.g., the EPW file) are retrieved in the second step and incorporated to ensure specific simulation results. Then, in the third step, building performance simulation parameters are established using Honeybee’s software, including occupancy patterns, HVAC loads, equipment and lighting densities, infiltration rates, and ventilation. In the fourth step, a multi-objective optimization process is then developed using the integration of the Non-dominated Sorting Genetic Algorithm III (NSGA-III) and Latin Hypercube Sampling (LHS) to establish the base static configuration of the system. After that, the optimal configurations from the first-stage optimization are evaluated using three decision-making methods: Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), Pareto front analysis, and Design explorer. Finally, the selected static configuration is further refined into a dynamic adaptive system through a second optimization phase conducted with brute-force optimization, focusing exclusively on the rotation angles of the louver and light shelf components while maintaining the same set of four objectives.

2.2. Base Case Model of Proposed System

In this study, a single-zone office space commonly used in performance-based simulation studies is employed as a standard model. The model geometry consists of a rectangular room measuring 3.60 m in width, 8.20 m in length, and 2.80 m in height, with a WWR of 45% on the south façade. This configuration is adapted from the widely referenced Reinhart reference office model [42], chosen for its relevance to typical office layouts and suitability for daylighting and energy analyses. The south-facing window, where the system is integrated, measures 3.00 m (width) × 1.50 m (height).

A key innovation of this system is the light shelf’s folding structure, which integrates the PV system (see Figure 3). The interior-facing surfaces of the light shelf are equipped with multiple reflectors that function as sequential mirrors to boost the light shelf’s efficiency by redirecting natural light to varying depths within the space. On the exterior-facing surfaces, PV modules are integrated and positioned perpendicularly to the sun’s rays to maximize solar energy capture through direct radiation. This dual functionality enables the system to improve the interior daylight levels while simultaneously generating electricity, which contributes to a reduction in the building’s overall energy consumption. Multi-objective optimization was conducted to identify the best angular configuration. This innovative approach, a combination of NSGA-III and Latin Hypercube Sampling, considered a wide range of design parameters to identify the optimal configuration across an annual cycle.

Moreover, to facilitate dynamic optimization, a brute-force method was employed, keeping all variables constant except the angular parameters. To enhance computational efficiency, the simulation year was strategically segmented and divided into four representative seasonal periods, while daily occupancy was further categorized into three key time slots. This careful approach allowed for the assignment of optimal light shelf and PV module angles tailored to specific climatic and usage conditions, resulting in a dynamic annual profile. Subsequently, detailed simulation analyses were conducted to assess the system’s performance, focusing on daylight utilization and overall energy efficiency. This is detailed in the following sub-sections.

2.3. Description of Simulation Metrics

2.3.1. Daylight Autonomy (DA)

Daylight Autonomy (DA) is a critical metric that quantifies the amount of natural daylight available in an indoor space over a year. DA is the percentage of occupied hours throughout the year during which the illuminance level at each sensor meets or exceeds a specified threshold [43]. Typically, the standard illuminance threshold used for DA is 300 lux [44], although this value can be adjusted depending on the specific visual needs of the space under evaluation. Here is the equation used to calculate DA:

$$\mathrm{D}\mathrm{A}={\displaystyle \frac{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{h}\mathrm{o}\mathrm{u}\mathrm{r}\mathrm{s} \mathrm{w}\mathrm{i}\mathrm{t}\mathrm{h} \mathrm{d}\mathrm{a}\mathrm{y}\mathrm{l}\mathrm{i}\mathrm{g}\mathrm{h}\mathrm{t} \mathrm{i}\mathrm{l}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{n}\mathrm{c}\mathrm{e}\ge \mathrm{t}\mathrm{h}\mathrm{r}\mathrm{e}\mathrm{s}\mathrm{h}\mathrm{o}\mathrm{l}\mathrm{d}}{\mathrm{T}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l} \mathrm{n}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r} \mathrm{o}\mathrm{f} \mathrm{o}\mathrm{c}\mathrm{c}\mathrm{u}\mathrm{p}\mathrm{i}\mathrm{e}\mathrm{d} \mathrm{h}\mathrm{o}\mathrm{u}\mathrm{r}\mathrm{s}}} \times 100$$

(1)

2.3.2. Useful Daylight Autonomy (UDI)

The Useful Daylight Autonomy (UDI) was specifically designed to address the limitations of simpler daylight metrics like DA, which does not account for the negative impacts of excessive daylight, such as overheating or visual discomfort [45]. UDI measures the percentage of time throughout the year that indoor illuminance stays within the optimal range, not too bright to cause glare, yet not too dim to be ineffective [46]. The calculation formula for the ${UDI}_{avg}$ is provided below:

$${UDI}_{avg}={\displaystyle \frac{t_{300Lux}\le E\le t_{3000Lux}}{\mathrm{T}}}\times 100$$

(2)

The useful daylight illuminance (${UDI}_{avg}$) metric measures the percentage of total occupied hours during which daylight levels remain within a “useful” range of 300 to 3000 lux. In this context, E represents the total hours when daylight illuminance falls between 300 and 3000 lux [44], while $\mathrm{T}$ denotes the overall number of occupied hours in a year.

2.3.3. Daylight Glare Probability (DGP)

Daylight Glare Probability (DGP) is a numerical metric used to assess discomfort glare caused by direct daylight within a space. DGP is calculated using a formula that incorporates both the vertical illuminance at eye level and the luminance of specific light and glare sources visible in the observer’s field of view [47].

$$DGP=5.87\times {10}^{-5}\times E_v+9.18\times {10}^{-2}\times \log \left(1+{\sum }_i{\displaystyle \frac{L_{si}^2{W}_{si}}{E_v^{1.87}\times P_i^2}}\right)+0.16$$

(3)

which $E_v$ is the vertical illuminance at eye level [lux], $L_{si}$ is the luminance of the $i$-th light source in the space [cd/m²], $W_{si}$ is the solid angle of the $i$-th space light source, and $P_i$ is the Guth position [48].

DGP calculations extend across time and space, using a climate-based daylight metric called spatial Glare Autonomy ($sGA$). $sGA$ quantifies the percentage of test points that maintain a glare-free environment for a minimum acceptable portion of daylight hours [49]. This approach offers a more accurate understanding of visual comfort by analyzing how light interacts with different areas continuously throughout the day and year [50]. The $sGA$ formula is structured as follows:

$$sGA={\displaystyle \frac{{\sum }_{hour=1}^{hour=N}{GA}_{\left(i\right)}}{N}}\times 100 GA\left(i\right)=\left\{\begin{array}{c}1, {DGP}_{threshold}\ge {ga}_i\\ 0, {DGP}_{threshold}\le {ga}_i\end{array}\right.$$

(4)

According to this equation, a design with a high $sGA$ value is expected to provide consistent visual comfort and minimize glare, while lower values could indicate regions susceptible to discomfort and glare problems.

2.3.4. Annual Sunlight Exposure (ASE)

Annual Sunlight Exposure (ASE) specifically focuses on restricting excessive daylight exposure to prevent issues like glare and maintaining a balance between natural light utilization and visual comfort. It quantifies how much of the space receives excessive direct sunlight (in this study, it is 1000 lux) for more than a certain number of hours (in this study, 250 h).

2.3.5. Energy Use Intensity (EUI)

In this study, we considered cooling, heating, and lighting energy as the primary contributors to Energy Use Intensity (EUI). Since cooling energy consumption significantly outweighs heating and lighting in our analysis, we developed a weighted function to prioritize reducing the overall EUI with a particular emphasis on cooling. An equal importance factor of 0.5 to the total EUI and cooling energy was assigned to achieve this balance. This normalization ensures that both aspects are given equal priority in the assessment, demonstrating the fairness and thoroughness of our research. We termed this metric the Weighted Energy Index (WEI).

$$\mathrm{W}\mathrm{E}\mathrm{I}=\left(\left(Heating+Cooling+Lighting\right)\times 0.5 \right)+(Cooling\times 0.5)$$

(5)

For the optimization process, each objective was solved sequentially, meaning the outcome of one objective influences the stage for the next objective, while all objectives are from one simulation. Thus, first, we evaluated the system’s daylighting performance using the DA metric. Next, the $sGA$ metric was calculated to prevent excessive daylight that could cause glare. Once these daylighting objectives were established, we assessed PV electricity generation by analyzing total incident radiation. Since Ladybug did not offer a dedicated component for calculating electricity generation, we used the incident radiation component to approximate the best solutions. In the final step, we calculated the total energy consumption of the building by measuring the WEI. Considering the objectives of daylighting, glare control, and PV generation, we can optimize the system to find the most efficient solution for meeting the building’s energy and daylighting demands.

2.4. Study Area and Weather Data

Tehran was selected as the case study location due to its distinctive climate, urban complexity, and environmental challenges [51]. Tehran is the capital and most populous city of Iran [52], representing one of the most densely populated urban areas on the globe [53]. Geographically, it is located at 35.72° N latitude and 51.33° E longitude, and experiences a semi-arid climate (BSk) under the Köppen classification, with hot, dry summers and cold, wet winters [54]. In addition, the city is affected by severe air pollution and a substantial stock of aging buildings with thermally inefficient envelopes [51], thereby presenting a highly relevant context for testing optimization strategies aimed at improving energy performance, daylighting, and solar utilization in real-world applications.

To provide accurate climatic input for the simulations, weather data from the Mehrabad synoptic station were utilized. This station was selected due to its long-term data availability, high reliability, and representativeness of typical urban conditions in Tehran [55]. The EnergyPlus Weather (EPW) file generated from Mehrabad’s records was used to simulate real-world climate conditions. This file includes hourly data on temperature, solar radiation, humidity, wind, and sky conditions and was employed to generate a comprehensive climatic dataset that served as the environmental basis for all energy and daylighting simulations conducted in this study.

2.5. Description of Optimization Algorithm

This study developed NSGA-III with the LHS method as an innovative approach to multi-objective optimization.

NSGA-III is an advanced evolutionary algorithm designed specifically to solve complex multi-objective optimization problems, particularly when dealing with more than three objectives [56]. NSGA-III addresses this limitation by incorporating predefined reference points and niche-preserving techniques, ensuring a more uniform distribution of solutions along the Pareto front. This enhancement enables NSGA-III to effectively balance convergence and diversity, making it particularly suitable for solving high-dimensional multi-objective optimization problems at a faster and more precise search speed [48,50]. The NSGA-III parameters, which were used in this study, are shown in

Table 2. Thermal properties of building model and building equipment.

Parameter	Value
Population	100
Generation count	150
Crossover probability	0.5
Mutation probability	0.35
Crossover distribution index	5
Mutation distribution index	10
Random state	42
Number of partitions	15

.

To complement the strengths of NSGA-III, the LHS method was employed as an advanced initial sampling strategy due to its proven ability to generate a uniformly distributed set of samples within a multidimensional parameter space. This uniformity minimizes sampling redundancy and ensures comprehensive coverage [57]. The synergy between LHS and NSGA-III enhances both the efficiency of the search process and the diversity of the solutions. Specifically, by initiating the algorithm with a highly diverse initial population, LHS enables NSGA-III to investigate a broader spectrum of the Pareto front and reduces the likelihood of premature convergence to suboptimal solutions [57].

As illustrated in Figure 4, several key parameters were identified as primary indicators in the light shelf system, including light shelf width, its distance from the ceiling, the number of slats, and the angle of both the exterior and interior light shelves. To model the number of slats, a rule-based approach was adopted based on the overall depth of the light shelf. Specifically, the system ensures that shallower light shelves contain fewer slats, while deeper configurations incorporate more divisions. Based on this approach, the conditional modeling logic is as follows: for light shelves with a depth less than or equal to 0.45 m, the number of slats is up to 4; for depths ranging from 0.45 m to 0.75 m, the maximum number of slats is 5; and for depths ranging from 0.75 m to 0.9 m, up to 6 slats are permitted.

In parallel, key design parameters were also considered for the louvers, including the vertical distance from the light shelf, their number, depth, and angle. Louvers are an integral part of the building’s design, serving to moderate direct solar gain, thereby reducing cooling loads and improving occupant comfort. By analyzing these factors, this study aims to develop a highly efficient hybrid shading system. The detailed parameters can be found in

Table 3. Variable parameters for the static configuration of the system.

Exterior Light Shelf Variable
Design Parameters	Unit	Increments	Range
Depth	Meter	0.15	[0.45–0.90]
Distance from roof	Meter	0.15	[−0.60–−0.30]
Number of slats (Set I = if condition for each depth)	Number	1	If Depth == 0.45 → [2,3,4] If Depth == 0.60 or 0.75 → [3,4,5] If Depth == 0.90 → [4,5,6]
Angle of light shelf	Degree	1	[0–30]
Angle of photovoltaic	Degree	1	[0–40]
Angle of light shelf (Clockwise (+) and anticlockwise (–) tilt angles)	Degree	1	[−10–10]
Interior Light Shelf Variable
Depth	Meter	0.05	[0.10–0.30]
Angle of light shelf (Clockwise (+) and anticlockwise (–) tilt angles)	Degree	1	[−10–10]
Louver Variable
Distance from light shelf	Meter	0.1	[−0.5–−0.3]
Number	Number	1	[3,4,5,6,7,8]
Depth	Meter	0.05	[0.10–0.30]
Angle	Degree	1	[0–45]

.

In this study, four objectives were established for optimization, including minimizing the WEI due to the high amount of cooling load (kWh/m²), maximizing DA, maximizing the power output from PV systems (kWh/m²), and maximizing $sGA$. These objectives balance energy use and indoor comfort. The results of this study were selected using three methods, including TOPOSIS, Pareto front, and Design explorer. This triple use does not imply simultaneous application; rather, it indicates that multiple complementary methods were employed to identify optimal solutions. TOPOSIS is simple and provides clear rankings, but becomes more sensitive to subjective weighting and normalization as the number of criteria grows [58]. In contrast, the Pareto front effectively captures trade-offs without requiring weights, avoiding bias, but tends to produce larger, more complex solutions that are harder to visualize and choose from with four objectives [1]. Design explorer facilitates interactive exploration and visualization, yet it does not perform any analytical processing. Instead, it allows the researcher to approximate and select an optimal solution through the iterative constraining of objectives [59].

2.5.1. TOPSIS Method

To select the optimal solution among various alternatives by considering different criteria, this study used TOPSIS, a widely recognized method for addressing multi-criteria decision-making (MCDM) problems [44,45,46]. This method uses a normalized decision matrix to calculate the distance from the positive ideal solution (PIS) and negative ideal solution (NIS), using a cosine-based metric. In the present study, equal weights were assigned to all objectives due to the absence of prior preferences or expert evaluation. This is a common approach in multi-criteria decision-making when all criteria are considered equally important and avoids introducing subjective bias [60]. The optimal solutions are detailed in

Table A1. The optimal TOPOSIS value of objectives.

DPvLsh	DPvLshR	NS	ALshS	APvS	APvLsh	DLvPvLsh	NLv	DLv	ALv	DInLsh	AInLsh	WEI	DA	$\mathit{s}\mathit{G}\mathit{A}$	PV
3	1	0	1	2	10	0	2	3	26	4	-10	52.6	44.1	0.95	551.0
3	1	0	1	0	10	0	2	2	45	4	-9	53.2	43.9	0.95	564.3
3	1	0	1	0	10	0	2	3	26	4	1	52.5	45.6	0.95	564.2
3	1	0	1	0	9	0	2	2	45	4	6	53.2	44.7	0.95	557.7
3	1	0	1	2	8	0	2	3	26	4	3	52.7	45.2	0.95	540.8

in Appendix A.

2.5.2. Pareto Front Method

In order to balance daylight availability and energy efficiency in building design, multi-objective optimization methods have become crucial [61]. In this study, the Pareto front was utilized to select optimal non-dominated solutions, meaning a solution can only be improved in terms of one objective by reducing the quality of at least one other objective. From the 38 non-dominated solutions generated, the 5 top-performing configurations were selected from the Pareto optimal set, as shown in

Table A2. The optimal Pareto front value of objectives.

DPvLsh	DPvLshR	NS	ALshS	APvS	APvLsh	DLvPvLsh	NLv	DLv	ALv	DInLsh	AInLsh	WEI	DA	$\mathit{s}\mathit{G}\mathit{A}$	PV
3	1	1	24	0	10	1	3	1	26	4	-10	52.2	47.4	0.93	310.2
3	2	1	17	0	9	0	3	0	25	4	7	53.2	51.8	0.91	391.9
3	2	0	0	0	10	0	3	0	12	4	6	53.1	51.7	0.90	575.3
3	1	1	24	10	8	0	3	1	7	4	-10	50.6	47.3	0.92	332.4
3	1	1	22	2	10	0	3	3	44	4	4	53.7	43	0.96	379.9

in Appendix A.

Figure 5 displays the Pareto front for four different optimization goals. Due to the difficulty of visualizing four objectives in a single graph and to better understand the trade-offs and optimal points, the Pareto front is shown in four 3D plots, each of which captures a combination of three objectives. Each plot within is dedicated to a unique combination of three objectives, thereby enabling a closer examination of the relationships and compromises between them.

2.5.3. Design Explorer Method

Moreover, the optimal solutions using the Design explorer method are shown in Figure 6, and the five most significant optimal responses are presented in

Table A3. The optimal design explorer value of objectives.

DPvLsh	DPvLshR	NS	ALshS	APvS	APvLsh	DLvPvLsh	NLv	DLv	ALv	DInLsh	AInLsh	WEI	DA	$\mathit{s}\mathit{G}\mathit{A}$	PV
2	2	1	17	1	10	0	3	0	4	4	7	52.7	51.9	0.90	412.8
3	2	0	15	0	9	0	3	0	6	4	8	52.8	51.5	0.90	510.8
3	2	0	17	0	9	0	3	0	7	3	4	52.8	52.3	0.91	509.2
3	2	0	20	1	8	0	3	0	5	3	5	52.7	51.4	0.90	496.3
3	1	1	21	0	10	0	3	1	9	4	5	52.4	51.1	0.92	389.3

in Appendix A.

2.6. Selection of Static System

Once the office and its integrated shading system were modeled and optimized, the next step was to identify the optimal solution based on the predefined priorities. To this end, a comparative analysis was conducted, employing the three-selection method outlined in Section 2.5.1, Section 2.5.2 and Section 2.5.3. A detailed comparison table, presented in Figure 7, was utilized to evaluate the performance of these approaches and determine the most effective solution. Given the prioritization of maximizing DA, the DA values across the solutions were systematically compared. The analysis revealed that the third solution, derived from the Design explorer method, achieved the highest DA value of 52.3%, surpassing the performance of the alternative solutions. Additionally, this solution exhibited the lowest WEI among the high-performing DA solutions, about 52.8, while demonstrating a negligible difference in its $sGA$. Consequently, based on these findings, the third solution from the Design explorer method was selected as the optimal configuration.

Upon comparing the values of the variables and parameters among the three methods, we observed that the angle of the light shelf slats, louvers, and interior light shelf varied the most. At the same time, the remaining factors remained within a relatively consistent range. For instance, the external light shelf maintained a depth of 90 cm in most of the optimum cases. The angle of the PvLsh system was approximately 10 degrees. Furthermore, there were three louvers, with the distance between the louvers and the light shelf being 30 cm. Through a comparative analysis of these cases and evaluating our priorities and façade design quality, we concluded that alternative number 3 of the Design explorer method represented the optimal case, achieving the best objectives among all methods. The selected configuration is summarized in

Table 4. The variables of the selective case.

DPvLsh	DPvLshR	NS	ALshS	APvS	APvLsh	DLvPvLsh	NLv	DLv	ALv	DInLsh	AInLsh	WEI	DA	$\mathit{s}\mathit{G}\mathit{A}$	PV
3	2	0	17	0	9	0	3	0	7	3	4	52.8	52.3	0.91	335.5

, and its corresponding model is illustrated in Figure 8.

2.7. Selection of Dynamic System

To enhance solar radiation control and ensure optimal performance throughout the year, a dynamic and adjustable shading system was developed, building upon the previously optimized static configuration. After determining the optimal fixed state, an advanced scheduling strategy was devised to adjust the system’s angular settings according to the solar altitude angle during occupied hours. Rather than relying on real-time sensors, the system operates based on pre-calculated optimal angles derived from the interval-based optimization process. For each defined daily and seasonal interval, the angle that yields the best balance between daylight provision and thermal load reduction was identified. These discrete settings are then programmed into the control system, enabling the shading device to automatically adjust its position at predefined times without the need for continuous sensor feedback. At this stage, all non-angular variables derived from the annual optimization were held constant, while the optimization process concentrated solely on angular parameters, as outlined in

Table 5. Variable parameters for dynamic configuration.

Exterior Light Shelf Variable
Design Parameters	Unit	Increments	Range
Angle of reflector slats	Degree	10	[0–30]
Angle of photovoltaic slats	Degree	10	[0–40]
Interior Light Shelf Variable
Angle of light shelf (Clockwise (+) and anticlockwise (−) tilt angles)	Degree	10	[−10–10]
Louver Variable
Angle	Degree	15	[0–45]

.

The implementation of this dynamic system was grounded in a solar geometry rationale to ensure precise responsiveness to significant seasonal and daily variations in solar altitude angles. The annual cycle was divided into four key seasonal intervals: the two solstices (21 June and 21 December) and the two equinoxes (21 March and 21 September), which represent the extremes and transitional points of solar altitude and azimuth angle throughout the year. This division was not only a response to software constraints in modeling dynamic behaviors but also aligned the system’s configuration with the most critical shifts in solar path.

Additionally, to optimize computational efficiency, occupancy hours were subdivided into three distinct periods: 8:00 a.m.–11:00 a.m., 11:00 a.m.–2:00 p.m., and 2:00 p.m.–5:00 p.m. These time slots reflect changes in solar position and intensity over the occupied hours, capturing the morning low-angle sun, the near-zenith midday condition, and the declining afternoon sun. For each daily interval, an optimal shading configuration was determined and uniformly applied across all hours within the day, balancing precision with computational feasibility. This method produced a daily adjustment schedule, which was then extrapolated across months based on the outcomes of the optimization process. This approach ensures that the dynamic system provides consistent daylight control while maintaining practical simulation and control requirements.

Given the limited number of variables (240 possible scenarios), a brute-force algorithm was applied to exhaustively evaluate all configurations within the defined range. Simulations were conducted for each seasonal period during occupied hours, ensuring a comprehensive performance analysis. During non-occupied hours, the system angles reverted to the optimal configuration identified through annual optimization. Finally, to present the outputs in the suggested dynamic system, where the proposed system varies according to the hour and month, the system was defined using the HB Dynamic Aperture Group and HB Dynamic State. The scheduling framework for adjustments was applied to annual daylight simulations through the HB Aperture Group Schedule component, as shown in

Table A4. Schedule of angular adjustments.

Schedule for Adjustment of Reflector Slat Angle
Hour/Month	Jan–Mar	Apr–Jun	Jul–Sep	Oct–Dec
8:00 a.m.–11:00 a.m.	0°	20°	20°	0°
11:00 p.m.–2:00 p.m.	0°	30°	10°	0°
2:00 p.m.–5:00 p.m.	0°	20°	0°	0°
5:00 p.m.–00:00 a.m.	17°	17°	17°	17°
00:00 a.m.–8:00 a.m.	17°	17°	17°	17°
Schedule for Adjustment of Photovoltaic Slat Angle
8:00 a.m.–11:00 a..m	0°	0°	0°	0°
11:00 p.m.–2:00 p.m.	0°	0°	10°	0°
5:00 p.m.–5:00 p.m.	0°	0°	0°	0°
2:00 p.m.–00:00 a.m.	0°	0°	0°	0°
00:00 a.m.–8:00 a.m.	0°	0°	0°	0°
Schedule for Adjustment of Louver Angle
8:00 a.m.–11:00 a.m.	0°	0°	0°	0°
11:00 p.m.–2:00 p.m.	15°	0°	0°	0°
2:00 p.m.–5:00 p.m.	0°	0°	0°	0°
5:00 p.m.–00:00 a.m.	7°	7°	7°	7°
00:00 a.m.–8:00 a.m.	7°	7°	7°	7°
Schedule for Adjustment of Interior Light Shelf Angle
8:00 a.m.–11:00 a.m.	10°	0°	10°	10°
11:00 p.m.–2:00 p.m.	10°	10°	10°	10°
2:00 p.m.–5:00 p.m.	10°	10°	10°	0°
5:00 p.m.–00:00 a.m.	3°	3°	3°	3°
00:00 a.m.–8:00 a.m.	3°	3°	3°	3°

in Appendix A.

3. Results

3.1. Correlation Between Objectives

The correlation heatmap in Figure 9 illustrates the linear relationships among the four primary objectives in this multi-objective optimization study. The most notable finding is the strong negative correlation (−0.95) between DA and $sGA$, indicating a significant trade-off between these objectives (for example, as DA increases, $sGA$ decreases due to heightened glare and distracting light levels in the space). Additionally, a moderate negative correlation (−0.65) is observed between the WEI and DA, suggesting that higher DA levels contribute to a reduced WEI by lowering both heating and artificial lighting energy demands. As expected, greater light penetration leads to lower $sGA$, showing a moderate positive correlation (0.67) between the WEI and $sGA$. Notably, PV demonstrates minimal correlation with the other objectives, indicating that PV output can be optimized independently, allowing for greater flexibility in integrating renewable energy without significantly affecting daylighting or energy efficiency.

Concerning the relationships between the design variables and objectives, Figure 10 offers valuable insight into the correlation features. Using values ranging from −1 (i.e., indicating a strong negative correlation) to 1 (i.e., indicating a strong positive correlation), each cell in the heatmap shows the correlation between a specific input parameter and an output metric. As shown in Figure 10, the significant finding is the positive correlation between PVLSH depth and solar electricity generation, with a correlation coefficient of 0.59. This implies that increasing the depth of PVLSH slats enhances PV electricity output, as larger slats offer more surface area for solar panels. On the other hand, the number of louvers is strongly correlated with heating energy demand (0.61) and UDI-Low (0.60), while showing a negative correlation with lighting metrics like sDA, UDI, and UDI-Up. This suggests that a higher number of louvers limits daylight and solar energy penetration, resulting in increased heating demands and more areas with insufficient daylight levels. These findings help inform practical design decisions.

Furthermore, the depth of louvers shows a positive correlation with UDI-Low (0.35), heating energy use (0.37), and GA (0.39). This indicates that deeper louvers reduce natural light availability, lowering daylight and solar energy levels while increasing UDI-Low value and heating energy demand. However, they also enhance visual comfort by minimizing GA. In contrast, the depth of louvers is negatively correlated with ASE, −0.38, and UDI-Up (−0.39), meaning that deeper louvers reduce direct sun exposure and excessive daylight within the space, contributing to a more controlled and visually comfortable environment.

3.2. Comparison of Scenarios

This study involves four distinct scenarios for comparative analysis: (i) The first scenario represents the baseline state of the building, operating with no shading system. (ii) In the second scenario, we examine the impact of an internal shading device, specifically a curtain. The characteristics of this curtain were selected to represent typical conditions in administrative offices in Iran. It features a diffuse reflectance value of 40% [62] and a transmitted diffuse value of 30% [31] to effectively capture its capacity to diffuse and transmit light. (iii) The third scenario involves implementing the optimized shading system, which was selected based on performance optimization criteria to enhance the building’s energy efficiency and daylight quality. (iv) The final scenario incorporates the dynamic state of the previously optimized system, which adjusts its movement and orientation based on the solar altitude angle. By comparing these four configurations, we aimed to gain insights into the influence of different shading strategies on indoor environmental conditions and energy performance.

Table 6. Daylight metric comparison across three scenarios, including no shading, internal shading, and optimized static shading.

	DA	UDI	UDI-Low	UDI-High	$\mathit{s}\mathit{G}\mathit{A}$
No Shading
	54.0%	57.2%	30.5%	12.3%	74.5
Interior Shading (curtain)
	38.0%	45.4%	49.2%	5.4%	92.0
Suggested Static System
	52.3%	61.6%	31.2%	7.2%	78.8

presents a comparative summary of daylight performance across the three scenarios.

The findings reveal a better daylight distribution and energy efficiency achieved by the suggested dynamic system compared to the three alternative scenarios. Figure 11 illustrates the comparison of lighting metrics across the three scenarios, highlighting the substantial enhancements offered by the optimized approach. These significant improvements in UDI underscore the effectiveness of the optimized design in improving lighting quality and energy efficiency, demonstrating improved system performance in key metrics.

The no-shading scenario demonstrates a high level of DA at approximately 54.0% and UDI-high at 12.3%, reflecting significant daylight penetration across the space. While this scenario ensures ample natural light availability, it causes glare, overheating, and a lack of sufficient illuminance at the back of the room, as evidenced by UDI-Low being 30.5%, ASE being the highest at 28.6%, and $sGA$ being the lowest at 74.0% across the three scenarios. These factors can negatively impact both visual comfort and overall energy efficiency.

The interior shading scenario, incorporating a curtain, achieves more moderated daylight performance. Compared to the no-shading scenario, the curtain scenario substantially reduces UDI-High to 5.4%, a decrease of 56.3%, effectively mitigating issues such as glare and excessive sunlight. This indicates that the curtain scenario enhances visual comfort by limiting the overwhelming impact of direct daylight. However, it is essential to note that the curtain scenario exhibits the highest $sGA$ value (92%) due to its ability to block a substantial amount of natural light and prevent direct sunlight penetration, resulting in an ASE value of 0. This trade-off is further highlighted by the decrease in the DA level to 38.0%, a drop of 29.6%, and the reduction in the UDI level to 45.4%, a decrease of 20.7%, compared to the no-shading scenario. This information equips us with a comprehensive understanding of the curtain scenario’s limitations, ensuring that we can make informed decisions about lighting design.

In contrast, the suggested system in its static form effectively balances daylight access and visual comfort. It substantially reduces UDI-high to 7.2%, a 41.6% decrease compared to the no-shading scenario, while maintaining favorable UDI and DA levels. Specifically, DA for the suggested system reaches 52.3% and increases by 3.1% compared to the no-shading scenario. Additionally, UDI reaches 61.58%, representing a 7.6% improvement over the no-shading scenario and a significant 35.7% increase compared to the curtain scenario. The system also improves $sGA$ to 79%, making a 5.76% enhancement over the no-shading scenario. Also, it decreases ASE to 14.3%, a 50.0% decrease compared to the no-shading scenario. These findings indicate that the proposed system allows for adequate natural light penetration while effectively mitigating glare and discomfort caused by excessive daylight. Consequently, it provides a well-balanced lighting environment that optimizes daylight utilization without compromising occupant comfort.

When the system operates in a dynamic configuration, as shown in

Table 7. Daylight metric for optimized dynamic shading scenario.

	DA	UDI	UDI-Low	UDI-High
Suggested Dynamic System
	61.2%	65.9%	23.2%	8.2%

and Figure 12, a significant improvement in daylighting metrics is observed. Specifically, DA reaches 61.2%, representing an increase of 17.1% compared to the static configuration and 13.4% compared to the no-shading scenario. Similarly, UDI rises to 65.9%, indicating a 6.9% improvement over the static configuration and 15.1% over the no-shading scenario. This increase in UDI under dynamic conditions not only demonstrates the current benefits but also hints at the potential for further improvement. Furthermore, UDI-Low decreases to 23.2%, indicating a significant reduction of approximately 23.9% compared to the no-shading scenario and 25.7% relative to the static configuration, and almost all of the room receives sufficient illuminance. This decline highlights an overall improvement in daylight distribution and less under-illumination. However, the UDI-High value reaches 8.2%, representing a 13.5% rise compared to the static configuration, suggesting a higher occurrence of excessive daylight penetration under dynamic conditions. This increase is attributed to temporal mismatches between the hourly adaptive adjustments of the dynamic shading system and the annual aggregation method used for UDI calculation. While the system frequently optimizes DA and UDI, transient overexposures during certain hours cumulatively raise the UDI-High value compared to the static system.

It is also important to note that, due to limitations in the Honeybee simulation platform, $sGA$ and ASE could not be computed dynamically, as the plugin currently does not support hourly adaptive shading states for these specific metrics.

Turning to energy performance metrics, as illustrated in Figure 13, the comparison highlights critical differences among the three scenarios, excluding the dynamic scenario, as it could not be measured due to the simulation limitations in modeling time-dependent thermal interactions. This limitation may lead to an underestimation of potential efficiency gains achievable with a dynamic control strategy.

The no-shading and curtain scenarios demonstrate elevated cooling energy consumption, accounting for approximately 38.39% and 39.3%, respectively. This high demand stems from the lack of control of solar heat gains, which results in increased indoor cooling loads and consequently higher energy consumption for air conditioning.

In contrast, the proposed static system demonstrates significantly improved performance, reducing cooling energy consumption to 32.1%, representing a decrease of 16.64% compared to the no-shading scenario and 18.6% relative to the curtain scenario. Additionally, it lowers the WEI to 52.8%, marking a reduction of 7.6% compared to the no-shading scenario and 10.3% relative to the curtain scenario. As a result, heating energy consumption rises to 38.3%, marking an increase of 9.4% compared to the no-shading scenario and 9.5% relative to the curtain scenario. Despite the rise in heating demand, the system performs well due to the substantially reduced WEI and cooling energy use. Specifically, cooling energy accounts for roughly 1.8 times the heating demand, underscoring the importance of reducing cooling loads as a priority.

In this study, cooling energy demand far exceeds heating requirements over the year, with the primary objective being the reduction in cooling loads. The implementation of this configuration successfully achieves a substantial reduction in cooling energy demand. While the system does lead to a slight increase in heating and lighting energy consumption, these increments are so small that they are almost negligible when compared to the improvements in daylighting quality and cooling efficiency.

Furthermore, the optimized shading system is designed as a multifunctional building-integrated PB solution, employing advanced technologies such as thin-film copper indium gallium selenide (CIGS) modules to enable seamless on-site renewable energy generation without compromising shading performance. The PV unit is dimensioned at 0.25 m in width and 3.0 m in length, with multiple units installed sequentially along the length of the shading device to achieve the targeted annual production capacity of 509.5 kWh. This installation strategy leverages the inherent flexibility and lightweight characteristics of thin-film CIGS technology, making it particularly suitable for integration within dynamic or non-planar shading geometries. Given the room’s measured annual EUI of 3322.21 kWh, this integrated PV configuration offsets approximately 15.3% of the total annual energy demand, accounting for realistic conversion losses, inverter inefficiencies, and local solar irradiation conditions. By integrating energy generation directly into the adaptive shading infrastructure, this approach not only reduces cooling loads and regulates indoor daylight levels within recommended thresholds but also contributes a measurable share of on-site renewable electricity. This dual functionality effectively diminishes the building’s reliance on external energy sources, supports energy self-sufficiency, and aligns with progressive sustainability standards. These findings underscore the efficacy of the optimized system in improving daylight quality and minimizing energy consumption, making it a responsible and superior choice for energy-efficient architectural design.

4. Discussion

This study introduced an integrated and dynamic optimization system of PV-equipped folding light shelves combined with glare-mitigating louvers. The proposed system aimed to optimize visual comfort and energy use and was designed with a comprehensive multi-objective approach. The optimization process for the base configuration of the system utilized a hybrid approach combining NSGA-III and Latin Hypercube Sampling, which is suitable for managing complex multi-objective scenarios involving numerous variables. These variables included the depth and positioning of the light shelf, their angles, several slats, the placement and angle of louvers, and several other geometric parameters.

The results of this study reveal that configurations with a light shelf depth of approximately 0.9 m, a roof distance of 0.6 m, and four to five slats offered the best daylighting performance. The optimal louver arrangement typically involved three louvers with a 0.3 m distance from the light shelf, while angles of various components remained highly adaptive to the changing solar altitude. Subsequently, brute-force optimization based on the static configuration was conducted to extract the system’s angular schedule while maintaining the same objectives. This led to an annual orientation schedule that enabled dynamic performance optimization.

The findings demonstrate strong performance, showcasing a substantial enhancement in daylighting conditions. The system reduced glare by 33.7% and cooling energy use by 16.6%, underscoring its overall performance. Moreover, the results show that UDI increased significantly by 15.1% and resulted in a 7.5% reduction in the WEI. The PV panels, based solely on incident solar irradiance and excluding system efficiency losses, were estimated to generate 509.5 kWh/year of energy, a testament to the effectiveness of the integrated approach. This study also confirms the system’s efficiency in bringing about daylighting and energy saving in buildings, thereby supporting future developments in energy-efficient and adaptive architectural design. In particular, future studies could further expand on this work by incorporating real-time control algorithms in diverse climatic regions.

Against the optimization results tabulated by [20], where shading design optimization yielded a 43.6% improvement in sDA, a 52.77% reduction in DGP, and a 13.9% reduction in EUI, our system achieved lower daylight improvement ratio values but added the additional advantage of harvesting local renewable energy (509.5 kWh/year). This added functionality is a significant difference in hybrid system performance. Therefore, compared to the study by [31] where the UDI gains reached 59.6% and the EUI reductions were as much as 31.29% for static shading devices, our findings indicate slightly reduced performance increments in these aspects; however, the flexibility of our design supports dynamic optimization throughout the year, which is not available for static systems. Furthermore, compared to the uncertainty and sensitivity analysis outcomes in [29], our results continue to concur on the high impacts of shading geometry and slat angle, but our PV-integrated approach offers an additional energy-saving choice not investigated in that work.

5. Conclusions

This study presents a designed folding PV-integrated light shelf system with high scalability for ASHRAE Climate Zone 3B buildings, such as that of the city of Tehran, as a sample of similar warm, arid climates. Its uniformity makes it suitable on a large scale in regions under identical climatic conditions, i.e., Middle East regions, Southern Europe regions, or other regions of Climate Zone 3B, where it can perform across the board to reduce Energy Use Intensity (EUI) and maximize daylighting and occupant comfort.

The integration of dynamic shading with PV energy harvesting offers a significant functional advantage over conventional static shading systems. The findings support the potential of adaptive architectural elements in promoting energy efficiency and occupant well-being.