Open AccessTechnical Note

Open-Source Software for Building-Integrated Photovoltaic Tiling for Novelty Architecture

Alexander W. H. Chin

Koami Soulemane Hayibo

and

Joshua M. Pearce

^1,2,*

Department of Electrical & Computer Engineering, Western University, London, ON N6A 3K7, Canada

Ivey School of Business, Western University, London, ON N6A 3K7, Canada

Author to whom correspondence should be addressed.

Designs 2024, 8(6), 118; https://doi.org/10.3390/designs8060118

Submission received: 8 September 2024 / Revised: 23 October 2024 / Accepted: 6 November 2024 / Published: 10 November 2024

(This article belongs to the Topic Net Zero Energy and Zero Emission Buildings)

Download

Browse Figures

Figure 1
STL models of the seven buildings architecture considered in this study. (a) The Thinker. (b) The Winged Victory of Samothrace. (c) The Colossus of Rhodes. (d) The Stanford Bunny. (e) The Tree. (f) The Inunnguaq. (g) The Pyramid. "> Figure 2
Rectangular overlay to determine the ground footprint (gray rectangle) of the novelty architecture BIPV, example of The Thinker model. (a) Side view. (b) Top view. "> Figure 3
Diagram of the PV layout used to estimate the ground footprint of the GPV. "> Figure 4
Rendering of The Thinker model design with tiled PV in different types of cities. (a) Fictional modern city. (b) Fictional futuristic city. "> Figure 5
Basic architectural render of some building models showing PV tiling in a location close to their cultural heritage. (a) Bunny in Stanford University, Stanford, CA, USA. (b) Inunnguaq near Roundhouse in Toronto, ON, Canada. (c) Colossus straddling the harbor in Rhodes, Greece. (d) Winged Angel of Samothrace on top of a hill near Athens, Greece. "> Figure 6
Rendering of the outside and inside of the Tree and Pyramid models tiled with semitransparent BIPV. (a) Tree model in downtown Vancouver, BC, Canada. (b) Pyramid model in Stanford University, Stanford, CA, USA. (c) Inside view of the pyramid model looking towards the Bunny. "> Figure 7
Sample simulation results showing the faces that generate the most energy per unit area (MWh/m2) in the optimal azimuth orientation of the Bunny model. (a) Treemap plot showing all faces with the energy density. (b) Front view of the Bunny model showing the faces with the energy density. (c) Side view of the Bunny model showing the faces with the energy density. "> Figure 8
Azimuth (°) dependency of the annual energy yield (MWh/MW) of the BIPVs simulated in London, ON, Canada. The azimuth of Face 0 of each building was used as the reference angle. "> Figure 9
Comparison of the polygon decimation impact on energy yield, the number of faces, and the appearance of The Thinker building model. (a) Energy yield (MWh/MW) plot for three different decimations with an azimuth optimization. On the right, the energy yield (MWh/MW) of each face is represented on a treemap plot with the corresponding building appearance. (b) Medium decimation (higher resolution). (c) Low decimation. (d) Extremely Low decimation (lower resolution). "> Figure 10
BIPV power density (MW/m2) as a function of building height (m) showing the comparison to real-world modern building heights. "> Figure 11
Representative Mitrex façade prints on PV modules: (a) marble, (b) sandstone, (c) granite, (d) slate, (e) wood, (f) brick, (g) metal, and (h) solid colors. "> Figure 12
Pyramid rendering using marble solar facing. "> Figure A1
The ray sections within the polygon start with an odd number of intersections and end with an even number of intersections [<a href="#B66-designs-08-00118" class="html-bibr">66</a>]. ">

Versions Notes

Abstract

Novelty architecture buildings can be tiled with conventional rectangular solar photovoltaic (PV) modules with both close-packed cells or partially transparent modules, vastly increasing renewable energy, reducing carbon emissions, and allowing for positive energy buildings. To enable this potential, in this study, for the first time, two open-source programs were developed and integrated to provide a foundation for designing and coating real-life novelty architecture buildings and objects with solar PV modules. First, a tiling algorithm was proposed and integrated into Blender that can generate solar PV modules on the face of any 3D model, and an augmented Python version of SAM was developed to simulate the performance of the resultant irregularly shaped PV systems. The integrated open-source software was used to analyze the energy performance of seven different novelty BIPVs located across the globe. The buildings’ energy performance was compared to conventional ground-based PV systems, and the results showed that the conventional arrays generate more energy per unit power than the BIPVs. The analysis reveals that the more complex the building model geometry, the less energy the building generates; however, the novelty BIPV power and energy densities far surpass conventional ground-based PV. The real estate savings observed were substantial, reaching 170% in one case where the BIPV reached 750 m in height. The BIPVs’ energy production is optimized by orienting the building via rotation and only needs to be carried out a single time for replication anywhere globally. The results show that the energy yield of the BIPV increases as the building becomes more detailed while the total power and energy decrease, indicating the need for the careful balancing of priorities in building design. Finally, the energy simulations demonstrate the potential for net-positive energy buildings and contribute to net-zero-emission cities. The findings indicate that BIPVs are not only appropriate for conventional residential houses and commercial buildings, but also for historical building replicas or monuments in the future. Further studies are needed to investigate the structural, electrical, and socio-economic aspects of novelty-architecture BIPVs.

Keywords:

novelty architecture; solar photovoltaic technology; building-integrated photovoltaic tiling; Blender; tiling

1. Introduction

The energy consumption in buildings has been increasing worldwide due to the increase in global average temperatures [1] and the rise of extreme weather events linked to climate change, notably the urban heat island effect [2]. To prevent the increase in mortality rate due to the increased frequency of heat waves [3], the energy consumption of buildings for cooling purposes has increased and now cities represent 75% of the global energy consumption [4]. This energy consumption is mainly supplied by fossil fuel energy sources because they remain predominant in the global energy mix, totaling 84.3% of all energy sources [5], causing cities and buildings to contribute up to 80% of global CO₂ emissions [2,6,7,8]. An increase in CO₂ emissions exacerbates the urban heat island effect, causing a positive feedback loop where the urban heat island effect and CO₂ emissions reinforce each other. One approach to breaking the cycle is for a building to transition to renewable energy prosumers by generating and sending renewable power to the grid, instead of only consuming carbon-emitting energy from the grid [9].

One way to realize the transition from simple consumer energy buildings to prosumer energy buildings is by merging solar photovoltaic (PV) technology into building materials. Building-integrated photovoltaics (BIPVs) are solar PV systems that seamlessly integrate into a building envelope [10,11]. BIPVs are not mere attachments to a building but displace the original building material in addition to generating electricity. Documented building components that integrate PV modules involve rooftops; façades including walls and windows; and external building attachments such as balconies and railing systems [12]. In addition to energy generation, BIPVs can be used for weather protection and thermal insulation [11], soundproofing [13], lighting adjustment [14], and improving building esthetics [15,16].

BIPVs are used for weather protection and thermal insulation by reducing the thermal load and the cooling load of a building. Moving BIPVs can for example, be adjusted to shade a building from sunlight and reduce cooling loads, or it can allow more light into a building and reduce heating loads [11,17]. The building energy consumption reduction using BIPVs is between 37% and 92%, depending on whether the BIPV is fixed or moveable and on the color of the PV modules [11,14]. Colored PVs can currently achieve an efficiency range from 24% to 75% of conventional PV modules [12]. The diversity in BIPV coloring enables the design of esthetically pleasing buildings, which increases the public acceptance of PV modules as building materials [18]. Furthermore, colored BIPVs can incentivize architects to integrate PV modules into unconventional building types, including heritage buildings, and in architecturally sensitive areas [15].

BIPVs, however, have historically been limited to systems intended to optimize solar energy generation and are generally limited in shape [12]. The optimal location of BIPV modules as part of a building envelope depends on the building design, orientation, and local weather conditions. Regardless of the location of the BIPV on the building, regular-shaped geometry (rectangular) and derived shapes are prioritized when performing BIPV envelope optimization for a building design [19]. Moreover, rooftops are the preferred location for BIPV installation as the percentage of rooftop BIPV is currently 80% of the global BIPV market [20,21].

Advanced materials and building techniques, however, have enabled far more interesting possibilities for building shapes called novelty architecture. Novelty architecture is a type of architecture where usable buildings take on unusual and out-of-the-ordinary forms, often for advertising or to attract attention. This includes constructing the building in the form of their market product or business venture, such as film shops shaped like a camera or lemonade stands in the form a lemon [22]. Prominent examples include the fish-shaped office building of the national fisheries development board in India and the Longaberger Company Headquarters in Newark, Ohio, in the shape of a basket [23]. Due to their novelty and exotic nature, many of these buildings serve as landmarks and tourist attractions.

Novelty architecture buildings can be tiled with conventional rectangular solar modules with both close packed cells or partially transparent modules to vastly increase renewable energy, reduce carbon emissions and allow for net-zero buildings and even building power surplus. Unfortunately, no design software exists to help architects design such buildings.

To overcome this challenge, this study reports on a new open-source software stack consisting of a novel tiling program within Blender that tiles the surface of buildings using an algorithm, where the module dimensions can be adjusted to tile most novelty building types. A supporting frame from the non-tiled surface is generated to make the buildings technically viable. Then, a python-based derivative of the Solar Advisory Model (SAM) is used to calculate annual energy yield to optimize novelty building BIPV orientation. This novel approach is tested on seven case-study novelty building structures, in which simulations are used to probe the impact of low-polygon simplification of object geometry on energy yield as well as orientation and geographic location. Finally, conclusions are drawn about the promise of this first program and algorithm to tile objects with solar panels to enable novelty architecture BIPVs. This is the first software approach that has attempted to make automatic tiling for BIPV novelty architecture a reality.

2. Methods

The open-source Blender script named BIPV-frame-generator.blend [24] is run on the open-source physics-based rendering engine in Blender version 3.5 [25] and can be divided into three stages:

Modification Stage;
Tiling Algorithm Stage;
Support Generation and Energy Output Stage.

2.1. Modification Stage

To prepare 3D models (e.g., STLs) for running, a model was first imported into Blender and modified. This normally includes scaling with the scale modifier to building size (as most open-source models are for small-scale 3D printing). It is possible to tile any STL as is, but a higher solar power density can be achieved by reducing the polygon count using the decimate modifier, thus creating more PV modules per unit surface area. The STL mesh may also not be suitable for tiling if it has too many thin faces. In this case, the limited dissolve modifier simplifies the mesh by dissolving edges and vertices that separate flat regions, forming a single face on the same plane. Finally, faces that should not be tiled (e.g., the bottom of the model on the ground) are selected and deleted.

2.2. Tiling Algorithm Stage

This stage is where the script runs the tiling algorithm (detailed in Appendix A) that fits as many rectangular PV modules as possible on each face of the model. Before running the script, the solar module length, width, and thickness are input. Running the script (select the model and press Alt + P.) then performs the algorithm on each face.

2.3. Support Generation and Energy Output Stage

Support generation provides for the building surface that is not covered by PVs, for which several methods of building 3D printing can be used in the manufacturing process. The solar panels areas are negatively extruded into the face they lie on, which leaves a hole, and the remaining faces are extruded by user-selected thickness to form a printable 3D model.

The intersection and knife tool provide the best cutting method for extracting solar panels in faces with little to no artifacts, which is referred to as the “cookie cutter” method. Essentially, the face is a sheet of cookie dough, and each solar panel is a rectangular cookie cutter. This rectangular cookie cutter makes rectangular cuts along the dough and removes them. The remaining excess is then extruded.

A similar process can happen in Blender in faces that form a cookie cutter in the form of the solar panel. These cutters then intersect the plane, and the knife intersection tool is run, separating the solar panels from the excess. These solar panels are then removed and what is left are the excess faces.

Once this is performed for each face, the model is left with a frame structure that outlines all the excess areas that could not be tiled with solar panels. By extruding each of these faces individually, a complete 3D model is formed for the non-PV parts of the building exterior. There will be gaps, however, in between these faces when extruding. A 3D wireframe of the model with a specified thickness needs to be generated to fill these gaps.

2.4. Output

For energy generation, the number of solar modules is recorded along with their tilt and orientation angles by the algorithm. The tilt angle is the angle from the normal vector of the solar panel to the positive Z axis and the orientation angle is the angle between a reference axis, which could be north, south, east, or west, and the projection that the normal vector makes on the XY plane. Since the model is made up of faces that have normal vectors, the script creates a large array calculating the tilt and orientation angle of each face and counts the number of solar PV modules that will fit on each face using the tiling algorithm mentioned above. This results in an array with the following format:

[[tilt angle, orientation angle, number of solar panels on face 1], …, [tilt angle, orientation angle, number of solar panels of face n]]

This array is then exported as a text file in the directory of the user’s choice. The surface area coverage,

C

(%), is a percentage calculated using the following equation:

C = \frac{n (w l)}{S} \times 100 %

(1)

where

n

is the number of tiled PV modules,

w

(m) is the width of the PV module,

l

(m) is the length of the PV module, and

S

(m²) is the surface area of the model. The dimensions of the solar modules used are width of 1.048 m, length of 2.108 m, and thickness of 0.04 m, which is common PV module dimension. Finally, for the image rendering, a cropped commercial solar module image was used as a texture [26].

2.5. Selected Example Models

Seven models were selected to demonstrate the capabilities of the software, and the basic STL models are shown in Figure 1. For each model, all surfaces except those with ground contact were tiled. All models used are available in the Open Science Framework (OSF) [24].

2.5.1. The Thinker

“The Thinker”, originally called “The Poet”, is a well-known sculpture made by French artist Auguste Rodin. It depicts a well-built man in deep thought, leaning forward, his chin resting on his hand and elbow resting on his knee. Such a profound pose conveys to the viewer that thinking is a powerful and concentrated exercise [27]. Originally, The Thinker was to be part of a pair of large bronze doors in which the figure, representing Dante Alighieri, sits atop, reflecting about heaven, hell, and the fate of all humankind in his epic poem The Divine Comedy. The Thinker was selected because it has become a ubiquitous symbol for all who use their imagination to create [27]. An open-source model was used [28].

2.5.2. The Winged Victory of Samothrace

The Winged Victory of Samothrace is a marble statue of the Greek goddess Nike who is seen as the Goddess of Victory. She is an important and popular figure in the Greek world. Nike’s ability to decide upon a victory was crucial for the ancient Greeks who were consistently participating in hometown competitions and battles within the Greek mainland and abroad [29]. An open-source model was selected as an example of revitalizing ancient art [30].

2.5.3. The Colossus of Rhodes

The Colossus of Rhodes, famously known as one of the Seven Wonders of the Ancient World, is a statue of the Greek sun-God Helios erected in commemoration of the raising of the siege of Rhodes. The sculpture was said to be 32 m high and took a staggering 12 years to build from 292 to 280 B.C. It was unfortunately brought down by an earthquake where the remains were left in place for nearly 900 years, which continued to serve as an attraction of Rhodes. Afterward, the statue was broken up and sold by the Saracens who conquered the island in 653 A.D. Therefore, no original fragments remain [31]. An open-source model was used [32], which was selected because of its ties to solar power, with some modifications (the original model was decimated by a factor of 0.001).

2.5.4. The Stanford Bunny (Low-Poly Version)

The Stanford Bunny is an iconic bunny model originating from the first 3D scan. It is a terra cotta bunny garden decoration. This new 3D scanning technological breakthrough was pioneered by Stanford professor Marc Levoy and his postdoctoral fellow Greg Turk when they created the world’s first seamless 3D computer model of a complex object using a range-finding laser scanner in the early 1990s. Since then, the Stanford Bunny has been used in many 3D computer graphics papers and is a standard model for computer graphic researchers to practice and experiment on [33]. An open-source, low-poly version was chosen [34].

2.5.5. The Tree (Low-Poly Version)

Both PVs and trees derive energy from the sun and contribute to the mitigation of climate change, which is why it is fitting to select a model of a tree tiled with PVs modules. An open-source low-poly blocky version was chosen [35]. The branches inside the tree foliage were removed and only the top surfaces of the tree were subject to tiling to maximize the PV energy output.

2.5.6. The Inunnguaq

The Inunnguaq is a well-known symbol in Canada that is a stone arrangement representing a human figure originally created by the Inuit people in the Arctic [36]. In Canada, they are present in urban cities like Toronto and Montreal. Oftentimes, an Inunnguaq is confused with an Inuksuk or Inukshuk, which is instead a marker made from stones used to convey navigational messages. The Inunnguaq only conveys a feeling of arrival to the Arctic and is not intended to contain information of the surroundings [37]. The model was created in OpenSCAD 2021.01 using a parametric design.

2.5.7. The Pyramid

The Pyramid is one of the only models that does not self-shade during the day due to its geometry; therefore, there are no losses from the modeled energy output of the BIPVs. The modeling developed in this study will provide accurate energy estimates regardless of the Pyramid’s geographic location or orientation as long as adjacent structures do not shade it. Pyramidal structures have been prominent in human history; remnants of these architectural artifacts still exist today and are part of the world heritage. The structures are spread out worldwide and can be found in places like Egypt, Sudan, Mexico, and Guatemala [38]. The existing pyramidal structures across the world do not have smooth faces; therefore, the Pyramid model was made more intricate by introducing extra faces to create bumps and ridges on the four main triangular surfaces.

2.5.8. Solar Energy Simulation

The energy generated by the PV-tiled buildings is modeled by summing up the electricity generation of each face. All modules on a specific building face have the same azimuth and tilt angles and are connected to the same inverter to limit the mismatch losses that could arise from the difference in azimuth angles. Shading losses are neglected to simplify the calculation. The energy analysis provided in this study focuses on the DC energy generated by the PV modules before the inverter connection. To find the total energy generated by a building face, the energy produced by a single PV module is calculated and scaled by the number of modules on that face, and the generation of an entire building is the sum of the energy generated by each face.

The solar energy model for the building energy simulation is performed using a Python version of the System Advisor Model (SAM). The PV SAM modeling package is preferred because it is open-source, its results are validated against real-world PV system performances, the solar panels and inverter database are maintained and updated with the most recent technology in the market, and the software package is frequently revised to incorporate the most recent PV energy calculation methods [39,40]. The SAM, however, can only handle a maximum of four PV arrays with different tilt and azimuth angles in a single simulation [41]. Complicated novelty architecture can have hundreds or thousands of faces depending on their rendering resolution and the building size. Consequently, a recently released Python version of the SAM PV modeling package was used to simulate the building energy performance in this study [42]. The model uses the same equations as the SAM 2023.12.17 Revision 2 software package and has results with a 99% accuracy compared to the SAM [42]. The flexibility of the Python version of the SAM is fit to the complex nature of the energy performance simulation of a building with multiple faces, each with a different azimuth and orientation angle. The parameters used in the PV model simulation are reported in the OSF.

The Python PV performance model was used to simulate the energy generated by the example building models described above. For each model, the energy generation was estimated for London, ON, Canada, which was considered the reference location in this study. A 10° incremental sensitivity analysis was performed to determine the optimum azimuth angle that generates the maximum annual energy in each building model. After the performance comparison in the reference location, the energy generation of each building was simulated in a location close to its cultural heritage. The Tree building model has no associated cultural heritage; therefore, it was placed in a country near the equator where trees are usually part of the urban environment. Finally, the rendering resolution and building size impact have been investigated by comparing the energy metrics of lower-polygon objects against higher-polygon objects, as well as models with different heights applied to The Thinker in its cultural heritage location. This sensitivity analysis has been performed on the Thinker model by varying the decimation of the design of the building polygon blocks between 4176, 829, and 328 polygons and the scale of the building between 50 m, 100 m, 300 m, 500 m, and 750 m. Below 50 m the structure could not be tiled with the current BIPV dimensions (2.108 by 1.048 m) as the faces are too small.

In each of the previously listed scenarios, the performance of the azimuth-optimized novelty architecture BIPV was compared to an optimized fixed-tilt conventional ground-based PV (GPV) of the same DC power size for each cultural heritage location. The metrics used to compare the performance of the different building models are the annual energy yield (MWh/MW), energy density (MWh/m²), and power density (MW/m²). This comparison provides information on the trade-off between the generated PV system power and the PV land occupation footprint.

The ground footprints of the BIPV and the GPV are required to compare the energy density of both systems. The BIPV footprint was determined in Blender by overlaying a rectangular plane onto each model from a top-down view where the plane outlines the match the model shape. The area of this rectangular plane was then recorded as the footprint. Figure 2 shows an example of The Thinker from a side view and a top view showing the footprint. It should be emphasized that the footprint is overestimated because using a rectangular plane introduces some extra space that is physically not occupied by the BIPV.

The ground footprint of the conventional GPV was calculated using analytical geometry combined with the PV layout assumptions. Assumptions are necessary for conventional GPV footprint calculation because the layout depends on the type of racking, the inverter used, the maintenance equipment required, and the shape of the land where the system is installed. In this study, the land was considered rectangular in shape. The type of racking considered was a 4 × 15 modules with landscape orientation [43]. The total power of the GPV is divided by that of 60 PV modules in a single row to determine the number of PV rows in the rectangular shape. The rows are considered spaced by the minimum distance possible to prevent shading in each location, as illustrated in Figure 3.

Equation (2) is used to calculate the ground footprint of the GPV where

A_{m i n}

(m²) is the minimum area occupied,

N_{R}

is the number of rows,

H

(m) is the projection of a row height on the ground,

D

(m) is the distance between two rows, and

L

(m) is the length of a row. In Equations (3) and (4),

w

(m) is the width of a module,

β

is the optimal tilt angle,

φ

(°) is the latitude coordinate of the location, and

δ

(°) is the sun declination angle during the winter solstice in the northern hemisphere (−23.45°) [44].

A_{m i n} = (N_{R} \times H + (N_{R} - 1) \times D) \times L (m^{2})

(2)

H = 4 \times w \times \cos β (m)

(3)

D = \frac{4 \times w \times \sin β}{\tan (90 - φ + δ)} (m)

(4)

3. Results

3.1. Designs and PV Power Capacity

The results of the analysis for the seven models are shown in Table 1. The number of surfaces is the number of faces that are subject to tiling from Blender. The total surface area is the sum of the surface area of all faces that are subject to tiling, which is calculated with the 3D print addon in Blender [25]. The number of tiled PV modules is counted in the code and annual energy output is calculated with the SAM solar simulation.

The rendered building models showing the position of the solar modules are shown in Figure 4 and Figure 5. In Figure 4, The Thinker building design is rendered in a fictional contemporary city (Figure 4a) and a fictional futuristic city (Figure 4b). The rendered scenes show that erecting a building in the shape of The Thinker statue is feasible and would fit inside a city at scales that are becoming common for buildings in large cities.

Figure 5 shows four building model designs rendered in a location close to their cultural heritage. The Bunny is placed in Stanford University because the polygonal method used in this study originated from there, and the first model created was that of a bunny. The Inunnguaq, the Colossus, and the Angel are rendered in Toronto, Rhodes Islands, and Athens, respectively. These cities are locations close to the buildings’ historical origins. It is important to note that when building a replica of historical or cultural monuments, it is recommended to consult with the local population for an accurate representation of any culturally sensitive aspect of the monuments.

Figure 6 illustrates the Tree and Pyramid model along with an inside view of the pyramid, showing a closer view of the semitransparent BIPV modules.

3.2. Energy Simulation Results

A treemap plot was used to visualize the energy performance of the novelty architecture building as the plot shows each face of the building model with the performance metric (Figure 7 with the example of the Bunny model). The color scheme used in the treemap is transposed on top of the 3D model of the design to show the physical location of the faces and their energy production value range.

The energy performance analysis shows that installing PV modules on novelty architecture buildings is feasible. The results of the base case scenario, with all the buildings located in London, ON, Canada, show that the dependence of the energy yield on the azimuth of the BIPV does not follow the sinusoidal pattern of conventional GPV. There was no identified generalized pattern between the energy yield and the azimuth in the BIPV models, as shown in Figure 8, because of the geometric variety of the chosen models. The analysis of Figure 8 reveals that the more complex the building model geometry, the less energy the building generates. Clustering of the energy yield is dependent on the geometry in Figure 8. The Pyramid and the Tree (see Figure 1), which are buildings consisting of simple geometries facing upwards, generated more than 950 MWh/MW regardless of their orientation. At the same time, the remaining models, which have more complex geometries, yielded less than 800 MWh/MW.

The optimized Face 0 azimuth across all designs varies from 65.54° for the Bunny to 318.75° for the Colossus, as shown in Table 2. The non-sinusoidal pattern and the discrepancy between the optimal azimuths stem from the difference in the model’s shape and an arbitrary face choice for Face 0 of the buildings. This discovery emphasizes the need for developing new energy performance and optimization methods tailored to novelty BIPVs. Table 2 also shows the performance comparison of the BIPV models and conventional GPVs in London, ON, Canada. The data in Table 2 show that the BIPV models outperform the GPVs in terms of energy density and power density, despite the GPVs having higher energy yields than the BIPVs. For example, the GPV equivalents of the Tree and the Colossus models have 1.28 and 1.92 times more energy yields than the BIPV models. The Tree model, however, surpasses its GPV equivalent by a factor of 1.73 and 1.35 in power and energy density, respectively. Similarly, the Colossus packs 11.46 times more PV power per square meter and generates 5.97 times more energy per square meter of ground than its GPV counterpart. These trends are consistent across all models.

Table 3 shows the results of each building in a location close to its cultural heritage and compares each BIPV performance to a conventional GPV in an identical location with an optimal tilt angle. The optimal azimuth of Face 0 obtained in different locations matches the azimuth in the base case scenario where all buildings were in a unique location. This finding establishes that the optimization of the building azimuth needs to be performed only once. When the buildings moved to different locations with different latitudes and conventional GPV optimal tilt angles, the GPV performance metrics were not the same. The overall trend showing that the BIPVs have better energy and power density than the GPVs was preserved nonetheless.

When the decimation of the polygons used to generate the building faces is increased, the number of active faces of the PV increases accordingly, and the opposite effect is observed when the decimation of the polygons decreases, as shown in Figure 9. The more faces the model has, the more realistic the building appears. The results show that the energy yield of the BIPV increases as the building becomes more detailed. This outcome could incentivize building designers to make novelty architecture buildings look as realistic as possible if energy yield is the primary metric.

On the other hand, when the resolution of the polygons increases, the PV active area decreases despite the ground footprint remaining unchanged, as shown in Table 4. A lower active area implies a lower energy density generated by the building. The energy density of the extremely low decimation (low resolution) is 635 W/m² while the energy density of the medium decimation (high resolution) is 362 W/m². Thus, building owners simply interested in the most power and energy would favor low resolution structures. In addition to the previous finding, Table 4 shows that the optimal azimuth of the BIPV changes with the decimation level. Despite using the same reference face for the three decimation models of The Thinker, the simulation shows the optimal azimuth of Face 0 is, respectively, 110.3°, 154.9°, and 133.7°, for the extremely low, the low, and the medium decimation models.

The Thinker building model’s height sensitivity simulation results are displayed in Table 5. The optimal azimuth shows that when the building is scaled up, there is a slight change in the building’s optimal orientation. This orientation change, however, is not abrupt, as the models with heights of 100 m, 300 m, and 500 m have the same azimuth (154.8°). The data in Table 5 confirms that scaling up the height of novelty architecture BIPVs results in better energy and power density than GPVs. The power density and energy density of GPVs have minimum variability because the PV size and the ground footprint are scaled up linearly. The BIPV model is scaled up in three dimensions, making the energy and power density scale non-linear. The three-dimensional scaling behavior allows the BIPVs to use a smaller ground footprint for higher performance, as shown in Figure 10. Figure 10 also compares the chosen heights of The Thinker model to real-world buildings and monuments. The power packing factor of novelty architecture buildings increases with height, while the GPV packing factor remains constant. The non-linearity observed in both the energy and power density, however, shows there could be a limit where the BIPV height increase will result in slower power density gain. It is interesting to note that in Paris, at least, some of the scaled novelty buildings are within the height limit set by the Eiffel tower.

4. Discussion

4.1. Significance of the Results and Advantages of the Proposed Approach

The simulation of all buildings in one unique location reveals that the optimal orientation of a novelty-architecture BIPV is different compared to a conventional BIPV or a GPV. The optimal orientation for GPV is well known and documented in the literature as being due-south in the Northern Hemisphere or due-north in the Southern Hemisphere [45]. On the other hand, when retrofitting solar panels on an existing building facade, BIPV system installers usually have only four different orientations available because of the current modern building rectangular shapes. They only have other options for rooftops or building off of the structures’ surfaces. In the novelty BIPV analyzed in this study, the unconventional shapes of the buildings have highlighted the need to analyze each building individually for the best azimuth. There is a clear advantage to combining the SAM software capabilities and the flexibility of Python, so architects and civil engineers can account for building orientation in future projects to create cleaner cities and reduce their carbon footprint.

It is clearly shown that lower polygon models make for higher annual energy density, as more PV modules can be tiled on larger flat surfaces. Lower-poly models, however, exhibit a lower building energy yield and reduce model detail, which can affect its novelty and overall recognition by the public. Therefore, a balance is needed between PV tiling coverage and discernability/esthetic appeal, where the model should ideally be just recognizable to maximize PV coverage. This balance will optimize the annual energy yield, energy density, and power density while maintaining novelty and esthetic appearance.

One of the most interesting findings from the results is the performance of the novelty BIPVs compared to a conventional GPV. These results mirror those in the greater BIPV literature. In all simulations performed in this study, the BIPVs unquestionably outperform their ground counterparts in power density and energy density despite the GPVs having a better energy yield than the BIPVs. One interesting case is the building height simulation analysis. While examining the height sensitivity results, it was discovered that scaling The Thinker model design emphasizes the potential of novelty BIPVs to address the land occupation challenge GPV modules are facing [46]. This discovery confirms that BIPVs, especially novelty BIPVs, are a viable solution for the current PV industry real estate challenge. This is due to the approach shown here related to the better utilization of surface area on the ground. It is clear that the energy yield (kWh/kW) of the GPVs is higher than the BIPVs regardless of the building heights because GPVs have optimal orientation, tilt angle and spacing to minimize shading losses. BIPVs and novelty BIPVs have none of these advantages because entire facades can be facing the wrong direction and suffer from heavy shading losses (e.g., when one ear on the rabbit shades the other). It is, however, crucial to consider more than just the maximum energy output of a PV infrastructure when assessing its sustainability. BIPVs and novelty BIPVs with large heights have much greater power per unit area (kW/acre) and energy per unit area (kWh/acre) potential than GPVs. For a specified PV power rating, the results show that novelty BIPVs consistently require 2.6 times less real estate than GPVs. This outcome allows the proposed BIPV designs to pack more power than their ground counterparts per unit real estate area, making up for the energy yield deficit. It should also be pointed out that this is true only in low-building-packing-density arrangements. All cases here only consider unshaded novelty architectures. GPV does not shade any area outside of its footprint, while novelty BIPV could shade areas far past its footprint. There is already substantial literature on solar access rights [47,48,49], and BIPV developers need to consider solar access laws [50,51]. To minimize the off-property shading of other buildings that could have BIPV, such novelty architecture could be strategically located in public parks or near the northern shores of water bodies in the Northern Hemisphere and southern shores in the Southern Hemisphere.

Installing PV modules on novelty architecture offers several potential benefits. In addition to performing conventional building functions, such as housing office spaces, parking lots, and apartments, these buildings generate energy that could cover their self-energy consumption and beyond. For example, the height comparison data in Table 5, show that The Thinker building could deliver 121 GWh/year if its height is 500 m. With this height, Figure 10 shows that The Thinker would be between the height of the Empire State Building (ESB) and the CN Tower (CNT). The energy consumption values of those two buildings are 63.41 GWh/year and 16.3 GWh/year for the ESB [52], and the CNT [53] respectively, suggesting that the energy consumption of high-rise buildings is not proportional to their height as it depends on a large extent on its use. Instead, this energy consumption depends on parameters, such as design, planning, operation decisions, and cooling/heating requirements. By interpolating the data in Table 5, The Thinker would generate 36.15 GWh/year and 60.74 GWh/year if it was the height of the ESB and the CNT, respectively. This comparison indicates that The Thinker building could generate more than 50% of its energy if it operated as the ESB, thus offsetting energy imports, and more than 300% if it performed as the CN Tower, thus becoming a net-positive energy building. Net-positive energy buildings are of growing interest [54,55] because the concept views the role of a building as one that adds value to its context, community, or system of which it is a part [56]. This allows for system level optimization so that the PV-powered novelty architecture building would be both a symbolic as well as literal keystone to a given community.

Another significant benefit that could arise from novelty-architecture BIPVs is the socio-economic improvement of the city or country where the BIPV is located. The design of the buildings is undoubtedly peculiar and eye-catching. For this reason, having a novelty BIPV could spark tourists’ interest, which is sometimes one of the economic pillars of a country [57,58]. The building could host museums or libraries where visitors could go and learn about the accurate historical facts of the monument represented by the building design. In addition to the social and cultural values such a building could bring to a city, it could also supply electricity to surrounding neighborhoods. For instance, if The Thinker was installed in Paris and had the consumption pattern of the CN Tower, it would generate 44.44 GWh more energy than it needs to operate. In France, the average electricity consumption per household is approximately 5.7 MWh/year [59]. This consumption pattern suggests that, if located inside Paris, The Thinker building model, with the height and consumption patterns of the CN Tower, would generate enough excess energy to power 7796 households in Paris. By powering almost 8000 homes in Paris, this novelty architecture building is not only offsetting its self-consumption, but also allowing other households to use clean energy and fostering the global movement towards net-zero-emission cities [60,61].

4.2. Current Limitations and Future Work

There are a few limitations with the software that could be addressed in future studies. Depending on the novel architecture design, certain surfaces of the model may not cut the frame out properly, leaving artifacts and unwanted faces. This effect is often caused by either thin faces or a poor mesh. Currently, the solution is to separate the uncut faces by selecting and rerunning the script for those faces. If that fails, the depth parameter can be increased to a higher value used for the panel cutouts. In the worst-case scenario, the faces must be redesigned, and the script rerun. To prevent these issues, the ideal model should have well-defined faces with a simple geometry, such as triangles, completely flat faces, no overlapping faces, and a clean mesh with low-to-medium detail. More work can also be performed to improve the tiling algorithm for better surface coverage, as there are some surfaces throughout the example models where more PV modules could physically fit. This, however, will require more computation time and research. Another approach could enable PV modules of different shapes and dimensions to fill those areas, but this would increase the production costs. Due to the nature of the algorithm, most frame leftovers contain triangle shapes. In the future, non-standard PV module shapes, such as triangles of different shapes and sizes, can replace some parts of the frame and tremendously increase PV tiling coverage and annual energy output. Future research should investigate the viability, benefits, and drawbacks of new PV module shapes, such as trapezoids, not only in thin-film technologies, but also in crystalline silicon PV modules for use in novelty architecture and other applications. In addition, companies like Mitrex have already commercialized BIPVs that allow for different visual images to be printed on functional PV modules [62]. Mitrex facings [62] include marble, limestone, granite, slate, wood, brick, and metal, as well as any other color or texture that architects desire (see Figure 11). The solar facades cost the same as conventional facades and the efficiency penalty is generally 10% compared to a conventional module with the same cells. This can make a solar building look like it is made of stone, as demonstrated in Figure 12.

Another limitation of the PV modeling software is that it does not account for direct and self-shading of the modules. For these designs, neither were the structural calculations performed nor were the real-world electrical systems analyzed (e.g., inverter connection and the series and parallel connections). These calculations, including the impact of all types of shading, are beyond the scope of this initial study and are left for future research. More specifically, future research is needed to quantify the self-shading effect of the proposed novelty architecture building design. One recent method proposed in the literature uses geometric calculations to track self-shading in vertical swinging PV in agrivoltaics farm [42]. This method could be integrated into the proposed software in the future to evaluate the self-shading losses in complex novelty architecture, improving the accuracy of the energy model. As shown here the energy yield for any façade is correct if it is not self-shaded. For buildings like those shown in Figure 12, there is no error for self-shading. For buildings like those shown in Figure 5d, the wings will partially shade parts of the structure during certain times of the day and the energy yields will be lower. The corrections for self-shading are needed to eliminate these overestimates.

In addition, as novelty BIPV buildings can be quite large, as shown in the figures, a new method that uses height maps to estimate the shading of adjacent buildings is needed. As novelty BIPVs enables solar capture far in excess of its footprint, further studies are needed to investigate the policies and laws regarding solar access rights and suggest design pathways that ensure equitable access to solar resources for neighborhoods surrounding novel buildings.

Other BIPV modeling tools take a different approach and are summarized in Table 6.

As can be seen from Table 6, there are several functionalities that are new and provided by the software described in this study. Other attributes, like those of providing all solar energy for a given building like self-consumption rates, techno-economic feasibility, and regulatory constraints are extremely important for actual implementation, but are left for future work. In addition, these applications were initially designed to be stand alone; however, architecture firms generally have a Building Information Modeling (BIM). Future research is needed to create application programming interfaces (API)s for the software tools developed here to make them easier to integrate into BIMs. Specifically, the API could help with integrating into multi-objective optimization tools like SAMA [64] that address energy performance, capital costs, and operating costs to determine economic feasibility. For novelty architecture to be practical, it is critical to ensure optimal temperature regulation in novelty architecture buildings. Future studies are also thus required to model the thermal behavior of the BIPV and explore the heating, ventilation, and air conditioning needs. The thermal behavior study could be coupled with the structural envelope stability analysis through computational fluid dynamics and finite element analysis to reveal BIPV compliance with existing high-rise building codes, especially building height limitations and wind loading constraints. This analysis would enhance the understanding of the building’s energy consumption and contribute to a more accurate evaluation of the building’s net-zero emission potential [54,65,66] (e.g., using PV energy to run heat pumps [67]). The electrical wiring between individual PV modules and the array configuration must be optimized to improve the building energy generation. For instance, future studies could compare string inverters against microinverters based on the system cost and energy performance. Further work is also needed to optimize electrical integration. For systems with large faces all pointing in the same direction they could be operated on a single inverter (e.g., one inverter for each face—like the pyramid structure). For a more complex system like The Thinker, it is recommended to use microinverters so the current mismatch does not become an issue. Additionally, the building’s energy production could be boosted by tracking the sun through the building’s foundational platform rotation [68]. This rotation could improve the building’s esthetic appeal and increase its energy performance.

There are also several other aspects to explore before practically implementing a novelty BIPV. Future work should also look at any safety, environmental or economic considerations for a specific location. In addition, a next realistic milestone would be a 3D printable version of the building [69]. The tiling software can be used on small objects, such as human-sized models with smaller PV modules. Advancements in 3D printing and PV allow for a physical prototype to be built where the supporting frame can be 3D printed, smaller PV modules can be tiled, and future research could analyze the performance of such a 3D-printed design. This analysis would undoubtedly help understand the electrical performance of the BIPV and ultimately pave the way for possible construction methods using construction 3D printing. Three-dimensional printing models would also promote the exploration of more design shapes. For example, a design of a human holding a shield at an optimal tile angle could improve the energy harvesting surface.

5. Conclusions

In this study, for the first time, two open-source programs were developed and integrated to provide a foundation for designing and coating real-life novelty architecture buildings and objects with solar photovoltaic modules. First, a tiling algorithm was proposed and integrated into Blender that can generate solar PV modules on the face of any 3D model, and a Python replica of NREL SAM software was developed to simulate the performance of the resultant irregularly shaped PV systems. The integrated open-source software was used to analyze the energy performance of seven different novelty BIPVs. The building’s energy performance was compared to conventional ground-based PV systems, and the results showed that the GPVs generate more energy per unit power than the BIPVs. The analysis reveals that the more complex the building model geometry, the less energy the building generates; however, the novelty BIPV power and energy densities far surpass GPV. This has the clear potential of saving real estate for large-scale PV energy generation and showing the value of tiling high-rise buildings with PV modules that do not shade other BIPV. The real estate savings observed were substantial, reaching 170% in one case where the BIPV was 750 m. The BIPVs’ energy production was compared to similar buildings’ energy consumption, and the BIPVs have the potential to be net-positive energy buildings that contribute to the potential of net-zero emissions cities. This study is the first to combine a tiling algorithm and an energy performance model to design and simulate the performance of PVs clad on novelty architecture buildings. The findings could propel BIPVs from residential houses and commercial buildings to historical building replicas or monuments in the future. These would present challenges related to zoning and local regulations. Further studies are needed to investigate the structural, electrical, and socio-economic aspects of novelty architecture BIPVs. In addition, the software can be enhanced further by providing an API to integrate into other software and integrating multi-objective optimization features that address both energy performance, capital costs, and operating costs to determine economic feasibility.

Author Contributions

Conceptualization, J.M.P.; methodology, A.W.H.C., K.S.H. and J.M.P.; software, A.W.H.C. and K.S.H.; validation, A.W.H.C., K.S.H. and J.M.P.; formal analysis, A.W.H.C., K.S.H. and J.M.P.; investigation, A.W.H.C., K.S.H. and J.M.P.; resources, J.M.P.; data curation, A.W.H.C., K.S.H. and J.M.P.; writing—original draft preparation, A.W.H.C., K.S.H. and J.M.P.; writing—review and editing, A.W.H.C., K.S.H. and J.M.P.; visualization, A.W.H.C., K.S.H. and J.M.P.; supervision, J.M.P.; project administration, J.M.P.; funding acquisition, J.M.P. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Thompson Endowment and the Natural Sciences and Engineering Research Council of Canada (NSERC).

Data Availability Statement

All data are available at https://osf.io/erbyp/ (accessed on 7 November 2024).

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

Appendix A

The algorithm divides each face of the building into as many uniform rectangles with the same orientation, like traditional rooftop solar panels.

The algorithm begins by aligning the face along the XY plane with its normal face facing upwards towards the positive Z axis. Then, the longest edge of the face is rotated to be in line with the negative X axis. This way, the longest edge of the face has the lowest Y value. Afterwards, the minimum and maximum Y coordinate from the face is used to divide the face into rectangular rows with a width of the solar panel width. The code will then loop over each row.

For each row, the code places a rectangle on the row until no more can be placed. This is carried out through Ray casting, where two infinitely long rays, a solar panel width apart, form the upper and lower boundaries of the row. The rays then determine where along the X axis they intersect. By doing so, the rays determine the area enclosed and check to see how many rectangles can fit inside. If a face contains only convex angles within, each ray will only intersect the face twice.

If the face has concave angles within, the ray may intersect more than twice, but the number of intersections will always be an even number due to the polygon being enclosed. Hence, rays within the face will always start with an odd number of intersections and end with an even number of intersections. Therefore, the algorithm avoids placing rectangles with intersections that start with an even number of intersections and end with an odd number of intersections, because those areas will not be enclosed by the face.

Figure A1. The ray sections within the polygon start with an odd number of intersections and end with an even number of intersections [66].

Algorithm Pseudo Code

For every face in the model:

Align along the XY plane

Align the longest edge along the negative X axis

Find the lowest and highest Y coordinates (yMin, yMax)

yRay is equal to yMin

While yRay is less than yMax:

Get intersections of top horizontal ray

Get intersections of bottom horizontal ray

Find available areas in between the intersections

Place as many rectangles that can fit in those areas

yRay increases by panelWidth amount

References

Staffell, I.; Pfenninger, S.; Johnson, N. A Global Model of Hourly Space Heating and Cooling Demand at Multiple Spatial Scales. Nat. Energy 2023, 8, 1328–1344. [Google Scholar] [CrossRef]
Perera, A.T.D.; Javanroodi, K.; Mauree, D.; Nik, V.M.; Florio, P.; Hong, T.; Chen, D. Challenges Resulting from Urban Density and Climate Change for the EU Energy Transition. Nat. Energy 2023, 8, 397–412. [Google Scholar] [CrossRef]
Tong, S.; Prior, J.; McGregor, G.; Shi, X.; Kinney, P. Urban Heat: An Increasing Threat to Global Health. BMJ 2021, 375, n2467. [Google Scholar] [CrossRef] [PubMed]
UN-Habitat. Urban Energy|UN-Habitat. Available online: https://unhabitat.org/topic/urban-energy (accessed on 31 August 2023).
Ritchie, H.; Roser, M.; Rosado, P. Energy Mix. Our World in Data 2022. Available online: https://ourworldindata.org/energy-mix (accessed on 7 November 2024).
OECD Decarbonising Buildings in Cities and Regions; OECD Urban Studies; OECD: Paris, France, 2022; ISBN 978-92-64-63968-3.
UNEP Cities and Climate Change. Available online: http://www.unep.org/explore-topics/resource-efficiency/what-we-do/cities/cities-and-climate-change (accessed on 8 September 2023).
IEA Buildings—Energy System. Available online: https://www.iea.org/energy-system/buildings (accessed on 8 September 2023).
Parag, Y.; Sovacool, B.K. Electricity Market Design for the Prosumer Era. Nat. Energy 2016, 1, 16032. [Google Scholar] [CrossRef]
Ballif, C.; Perret-Aebi, L.-E.; Lufkin, S.; Rey, E. Integrated Thinking for Photovoltaics in Buildings. Nat. Energy 2018, 3, 438–442. [Google Scholar] [CrossRef]
Svetozarevic, B.; Begle, M.; Jayathissa, P.; Caranovic, S.; Shepherd, R.F.; Nagy, Z.; Hischier, I.; Hofer, J.; Schlueter, A. Dynamic Photovoltaic Building Envelopes for Adaptive Energy and Comfort Management. Nat. Energy 2019, 4, 671–682. [Google Scholar] [CrossRef]
Hamzah, A.H.; Go, Y.I. Design and Assessment of Building Integrated PV (BIPV) System towards Net Zero Energy Building for Tropical Climate. e-Prime—Adv. Electr. Eng. Electron. Energy 2023, 3, 100105. [Google Scholar] [CrossRef]
Chen, T.; Tai, K.F.; Raharjo, G.P.; Heng, C.K.; Leow, S.W. A Novel Design Approach to Prefabricated BIPV Walls for Multi-Storey Buildings. J. Build. Eng. 2023, 63, 105469. [Google Scholar] [CrossRef]
Saw, M.H.; Singh, J.P.; Wang, Y.; Birgersson, K.E.; Khoo, Y.S. Electrical Performance Study of Colored C-Si Building-Integrated PV Modules. IEEE J. Photovolt. 2020, 10, 1027–1034. [Google Scholar] [CrossRef]
Pelle, M.; Lucchi, E.; Maturi, L.; Astigarraga, A.; Causone, F. Coloured BIPV Technologies: Methodological and Experimental Assessment for Architecturally Sensitive Areas. Energies 2020, 13, 4506. [Google Scholar] [CrossRef]
Sun, H.; Heng, C.K.; Reindl, T.; Lau, S.S.Y. Visual Impact Assessment of Coloured Building-Integrated Photovoltaics on Retrofitted Building Facades Using Saliency Mapping. Sol. Energy 2021, 228, 643–658. [Google Scholar] [CrossRef]
Akbari Paydar, M. Optimum Design of Building Integrated PV Module as a Movable Shading Device. Sustain. Cities Soc. 2020, 62, 102368. [Google Scholar] [CrossRef]
Awuku, S.A.; Bennadji, A.; Muhammad-Sukki, F.; Sellami, N. Myth or Gold? The Power of Aesthetics in the Adoption of Building Integrated Photovoltaics (BIPVs). Energy Nexus 2021, 4, 100021. [Google Scholar] [CrossRef]
Youssef, A.M.A.; Zhai, Z.; Reffat, R.M. Generating Proper Building Envelopes for Photovoltaics Integration with Shape Grammar Theory. Energy Build. 2018, 158, 326–341. [Google Scholar] [CrossRef]
Kaji Esfahani, S.; Tenorio, R.; Karrech, A.; Defendi, K.; Jerez, F. Analysing the Role of Roof Mounted BIPV System Optimization on Decreasing the Effect of Duck Curve in Perth, Western Australia: An Experimental Case Study. Sustain. Energy Technol. Assess. 2021, 47, 101328. [Google Scholar] [CrossRef]
Shukla, A.K.; Sudhakar, K.; Baredar, P. Recent Advancement in BIPV Product Technologies: A Review. Energy Build. 2017, 140, 188–195. [Google Scholar] [CrossRef]
Munsey, C. Bottles in Mimetic Architecture. 2003. Available online: https://www.angelfire.com/zine2/thesodafizz/BottleArchitecture_July2003.pdf (accessed on 7 November 2024).
Adams, C. Mimetic Architecture: Why Does This Building Look like a Fish? Available online: https://www.cnn.com/style/article/mimetic-architecture-osm/index.html (accessed on 2 August 2023).
Chin, X.; Hayibo, K.S.; Pearce, J.M. BIPV Frame Generator and Associated Files. 2023. Available online: https://osf.io/erbyp/ (accessed on 7 November 2024).
Foundation, B. Blender Version 3.5. blender.org. Available online: https://www.blender.org/download/releases/3-5/ (accessed on 2 August 2023).
Buy Solar Module HT-SAAE HT54-18X-405W FULL BLACK|pvXchange.Com. Available online: https://www.pvxchange.com/Solar-Modules/HT-SAAE/HT54-18X-405W-FULL-BLACK_1-996000360 (accessed on 30 July 2023).
Rodin, A. The Thinker; Art, Architecture and Engineering Library: Ann Arbor, MI, USA, 1879. [Google Scholar]
3D Printable. The Thinker at the Musée Rodin, France by Musée Rodin. Available online: https://www.myminifactory.com/object/3d-print-the-thinker-at-the-muse-rodin-france-2127 (accessed on 2 August 2023).
Mayfield, M. The Winged Victory: Nike in Ancient Greece. Stud. Antiq. 2021, 20, 47–58. [Google Scholar]
3D Printable. Winged Victory of Samothrace at The Louvre, Paris by Scan the World. Available online: https://www.myminifactory.com/object/3d-print-winged-victory-of-samothrace-2073 (accessed on 2 August 2023).
Maryon, H. The Colossus of Rhodes. J. Hell. Stud. 1956, 76, 68–86. [Google Scholar] [CrossRef]
The Colossus of Rhodes Free 3D Model—Download stl File. Available online: https://3dmag.org/en/market/item/2464/ (accessed on 2 August 2023).
The Stanford Bunny. Available online: https://faculty.cc.gatech.edu/~turk/bunny/bunny.html (accessed on 2 August 2023).
Thingiverse.com. Low Poly Stanford Bunny by Johnny6. Available online: https://www.thingiverse.com/thing:151081 (accessed on 30 July 2023).
Low Poly Tree Pack 3d Model. Free Download|Reazilla. Available online: https://creazilla.com/nodes/1432904-low-poly-tree-pack-3d-model (accessed on 30 July 2023).
Hallendy, N. Inuksuit: Silent Messengers of the Arctic, 1st ed.; Douglas & McIntyre: Vancouver, BC, Canada, 2001; ISBN 978-1-55054-874-7. [Google Scholar]
Heyes, S. Protecting the Authenticity and Integrity of Inuksuit within the Arctic Milieu. Études Inuit Stud. 2002, 26, 133–156. [Google Scholar] [CrossRef]
Magazine, S. Ancient Pyramids Around the World. Smithsonian Magazine, 20 November 2009. [Google Scholar]
Gilman, P.; DiOrio, N.A.; Freeman, J.M.; Janzou, S.; Dobos, A.; Ryberg, D. SAM Photovoltaic Model Technical Reference 2016 Update; National Renewable Energy Lab. (NREL): Golden, CO, USA, 2018; p. 93. [Google Scholar]
Milosavljević, D.D.; Kevkić, T.S.; Jovanović, S.J. Review and Validation of Photovoltaic Solar Simulation Tools/Software Based on Case Study. Open Phys. 2022, 20, 431–451. [Google Scholar] [CrossRef]
Freeman, J.M.; DiOrio, N.A.; Blair, N.J.; Neises, T.W.; Wagner, M.J.; Gilman, P.; Janzou, S. System Advisor Model (SAM) General Description (Version 2017.9.5); National Renewable Energy Laboratory (NREL): Golden, CO, USA, 2018; p. NREL/TP--6A20-70414. [Google Scholar]
Hayibo, K.S.; Pearce, J.M. Vertical Free-Swinging Photovoltaic Racking Energy Modeling: A Novel Approach to Agrivoltaics. Renew. Energy 2023, 218, 119343. [Google Scholar] [CrossRef]
Kopecek, R.; Libal, J. Towards Large-Scale Deployment of Bifacial Photovoltaics. Nat. Energy 2018, 3, 443–446. [Google Scholar] [CrossRef]
Appelbaum, J.; Aronescu, A. Inter-Row Spacing Calculation in Photovoltaic Fields—A New Approach. Renew. Energy 2022, 200, 387–394. [Google Scholar] [CrossRef]
Duffie, J.A.; Beckman, W.A. Solar Engineering of Thermal Processes; Wiley: Hoboken, NJ, USA, 2013; ISBN 978-1-118-41812-3. [Google Scholar]
van de Ven, D.-J.; Capellan-Peréz, I.; Arto, I.; Cazcarro, I.; de Castro, C.; Patel, P.; Gonzalez-Eguino, M. The Potential Land Requirements and Related Land Use Change Emissions of Solar Energy. Sci. Rep. 2021, 11, 2907. [Google Scholar] [CrossRef] [PubMed]
Gergacz, J.W. Legal Aspects of Solar Energy: Statutory Approaches for Access to Sunlight. Boston Coll. Environ. Aff. Law Rev. 1982, 10, 1. [Google Scholar]
Bronin, S.C. Solar Rights. Boston Univ. Law Rev. 2009, 89, 1217. [Google Scholar]
Clarke, P. The Legal Right to Solar Access. Environment 2019, 1–15. Available online: https://acumen.architecture.com.au/environment/energy/energy-saving-technologies/the-legal-right-to-solar-access/ (accessed on 30 July 2023).
Bowden, M.-A. Protecting Solar Access in Canada: The Common Law Approach. Dalhous. Law J. 1984, 9, 261. [Google Scholar]
Limiting Off Property Shading of Solar Energy Systems—Sustainable Development Code. Available online: https://sustainablecitycode.org/brief/off-property-shading-of-solar-energy-systems/ (accessed on 7 November 2024).
Empire State Building. Empire State Building Fact Sheet; Empire State Building: New York, NY, USA, 2023; p. 2. [Google Scholar]
The CN Tower. Sustainability at the CN Tower; The CN Tower: Toronto, ON, Canada, 2022; p. 4. [Google Scholar]
Kolokotsa, D.; Rovas, D.; Kosmatopoulos, E.; Kalaitzakis, K. A Roadmap towards Intelligent Net Zero- and Positive-Energy Buildings. Sol. Energy 2011, 85, 3067–3084. [Google Scholar] [CrossRef]
Mousavi, S.; Villarreal-Marroquín, M.G.; Hajiaghaei-Keshteli, M.; Smith, N.R. Data-Driven Prediction and Optimization toward Net-Zero and Positive-Energy Buildings: A Systematic Review. Build. Environ. 2023, 242, 110578. [Google Scholar] [CrossRef]
Cole, R.J.; Fedoruk, L. Shifting from Net-Zero to Net-Positive Energy Buildings. Build. Res. Inf. 2015, 43, 111–120. [Google Scholar] [CrossRef]
Dabeedooal, Y.J.; Dindoyal, V.; Allam, Z.; Jones, D.S. Smart Tourism as a Pillar for Sustainable Urban Development: An Alternate Smart City Strategy from Mauritius. Smart Cities 2019, 2, 153–162. [Google Scholar] [CrossRef]
Salinas Fernández, J.A.; Guaita Martínez, J.M.; Martín Martín, J.M. An Analysis of the Competitiveness of the Tourism Industry in a Context of Economic Recovery Following the COVID19 Pandemic. Technol. Forecast. Soc. Chang. 2022, 174, 121301. [Google Scholar] [CrossRef] [PubMed]
Tostivint, F. Montant de la Facture Moyenne Électricité Par Mois + Consommation; Selectra: Paris, France, 2017. [Google Scholar]
Seto, K.C.; Churkina, G.; Hsu, A.; Keller, M.; Newman, P.W.G.; Qin, B.; Ramaswami, A. From Low- to Net-Zero Carbon Cities: The Next Global Agenda. Annu. Rev. Environ. Resour. 2021, 46, 377–415. [Google Scholar] [CrossRef]
O’Regan, A.C.; Nyhan, M.M. Towards Sustainable and Net-Zero Cities: A Review of Environmental Modelling and Monitoring Tools for Optimizing Emissions Reduction Strategies for Improved Air Quality in Urban Areas. Environ. Res. 2023, 231, 116242. [Google Scholar] [CrossRef]
Mitrex Architectural Solar Facades, Reimagined—Mitrex. Available online: https://www.mitrex.com/solar-facade (accessed on 9 July 2024).
Pearce, J.M.; Sommerfeldt, N. Economics of Grid-Tied Solar Photovoltaic Systems Coupled to Heat Pumps: The Case of Northern Climates of the U.S. and Canada. Energies 2021, 14, 834. [Google Scholar] [CrossRef]
Chou, K.C.; Shih, S.G.; Liu, S.Y. Research of the Main Building Characteristics for Rotating Buildings. Appl. Mech. Mater. 2014, 638–640, 1660–1665. [Google Scholar] [CrossRef]
Yin, H.; Qu, M.; Zhang, H.; Lim, Y. 3D Printing and Buildings: A Technology Review and Future Outlook. Technol. Archit. Des. 2018, 2, 94–111. [Google Scholar] [CrossRef]
Mecki Answer to “How Can I Determine Whether a 2D Point Is Within a Polygon?” Stack Overflow 2008. Available online: https://stackoverflow.com/questions/217578/how-can-i-determine-whether-a-2d-point-is-within-a-polygon (accessed on 7 November 2024).
Rana, S.; Jamil, U.; Asgari, N.; Hayibo, K.S.; Groza, J.; Pearce, J.M. Residential Sizing of Solar Photovoltaic Systems and Heat Pumps for Net Zero Sustainable Thermal Building Energy. Computation 2024, 12, 126. [Google Scholar] [CrossRef]
Seme, S.; Štumberger, B.; Hadžiselimović, M.; Sredenšek, K. Solar Photovoltaic Tracking Systems for Electricity Generation: A Review. Energies 2020, 13, 4224. [Google Scholar] [CrossRef]
El-Sayegh, S.; Romdhane, L.; Manjikian, S. A Critical Review of 3D Printing in Construction: Benefits, Challenges, and Risks. Archiv. Civ. Mech. Eng. 2020, 20, 34. [Google Scholar] [CrossRef]

Figure 1. STL models of the seven buildings architecture considered in this study. (a) The Thinker. (b) The Winged Victory of Samothrace. (c) The Colossus of Rhodes. (d) The Stanford Bunny. (e) The Tree. (f) The Inunnguaq. (g) The Pyramid.

Figure 2. Rectangular overlay to determine the ground footprint (gray rectangle) of the novelty architecture BIPV, example of The Thinker model. (a) Side view. (b) Top view.

Figure 3. Diagram of the PV layout used to estimate the ground footprint of the GPV.

Figure 4. Rendering of The Thinker model design with tiled PV in different types of cities. (a) Fictional modern city. (b) Fictional futuristic city.

Figure 5. Basic architectural render of some building models showing PV tiling in a location close to their cultural heritage. (a) Bunny in Stanford University, Stanford, CA, USA. (b) Inunnguaq near Roundhouse in Toronto, ON, Canada. (c) Colossus straddling the harbor in Rhodes, Greece. (d) Winged Angel of Samothrace on top of a hill near Athens, Greece.

Figure 6. Rendering of the outside and inside of the Tree and Pyramid models tiled with semitransparent BIPV. (a) Tree model in downtown Vancouver, BC, Canada. (b) Pyramid model in Stanford University, Stanford, CA, USA. (c) Inside view of the pyramid model looking towards the Bunny.

Figure 7. Sample simulation results showing the faces that generate the most energy per unit area (MWh/m²) in the optimal azimuth orientation of the Bunny model. (a) Treemap plot showing all faces with the energy density. (b) Front view of the Bunny model showing the faces with the energy density. (c) Side view of the Bunny model showing the faces with the energy density.

Figure 8. Azimuth (°) dependency of the annual energy yield (MWh/MW) of the BIPVs simulated in London, ON, Canada. The azimuth of Face 0 of each building was used as the reference angle.

Figure 9. Comparison of the polygon decimation impact on energy yield, the number of faces, and the appearance of The Thinker building model. (a) Energy yield (MWh/MW) plot for three different decimations with an azimuth optimization. On the right, the energy yield (MWh/MW) of each face is represented on a treemap plot with the corresponding building appearance. (b) Medium decimation (higher resolution). (c) Low decimation. (d) Extremely Low decimation (lower resolution).

Figure 10. BIPV power density (MW/m²) as a function of building height (m) showing the comparison to real-world modern building heights.

Figure 11. Representative Mitrex façade prints on PV modules: (a) marble, (b) sandstone, (c) granite, (d) slate, (e) wood, (f) brick, (g) metal, and (h) solid colors.

Figure 12. Pyramid rendering using marble solar facing.

Table 1. Information and data on the selected BIPV model designs and simulation locations.

Building Design	Angel	Bunny	Colossus	Inunnguaq	Pyramid	The Thinker	Tree
Culture of Origin	Greek	U.S.	Greek	Canada	Various Cultures	French	None
Total Outer Surface (m²)	23,287	11,661	53,117	13,088	9066	32,012	5371
PV Module Surface Area Coverage	0.86	0.70	0.68	0.79	0.91	0.70	0.76
Real Estate Ground Footprint (m²)	8406	4145	6674	1549	6029	7839	4902
Model Height (m)	136.0	71.2	292.0	84.5	38.2	130.0	52.9
Number of PV Modules	7286	3671	16,382	4702	3741	10,202	1854
Total PV Power (MW)	3.42	1.73	7.70	2.21	1.76	4.79	0.87
Default Simulation Location	London, ON, Canada
Cultural Heritage Simulation Location	Athens	Palo Aalto	Athens	Toronto	Cairo	Paris	Lomé
Optimal GPV Tilt Angle at Location	30	30	30	35	26	30	10

Table 2. Energy performance of the buildings models with an optimal azimuth compared to optimized-tilt, ground-based PVs (GPVs) in London, ON, Canada.

Building Model Design	Angel	Bunny	Colossus	Inunnguaq	Pyramid	The Thinker	Tree
Total PV Size (MW)	3.42	1.73	7.70	2.21	1.76	4.79	0.87
Optimal Building Face 0 Azimuth (°)	255.20	65.54	318.75	72.00	278.96	144.78	225.62
Building Annual Energy (MWh)	2584.14	1318.56	5366.83	1580.16	1809.91	3729.08	909.64
Optimized GPV Annual Energy (MWh)	4582.57	2308.90	10,303.55	2957.35	2352.92	6416.61	1166.08
Total PV Active Area (m²)	14,572	7342	32,764	9404	7482	20,404	3708
Building Ground Footprint (m²)	8406	4145	6674	1549	6029	7839	4902
Conventional PV Area (m²)	33,954.68	17,172.14	76,470.44	21,927.19	17,451.85	47,660.42	8501.16
Building Energy Yield (MWh/MW)	754.62	764.22	697.03	715.03	1029.37	777.71	1043.91
GPV Energy Yield (MWh/MW)	1338	1338	1338	1338	1338	1338	1338
Building Energy Density (MWh/m²)	0.31	0.32	0.80	1.02	0.30	0.48	0.19
GPV Energy Density (MWh/m²)	0.13	0.13	0.13	0.13	0.13	0.13	0.14
Building Power Density (W/m²)	407.38	416.25	1153.66	1426.69	291.64	611.68	177.76
GPV Power Density (W/m²)	100.85	100.47	100.69	100.79	100.75	100.61	102.50

Table 3. Energy performance of the buildings models with an optimal azimuth compared to optimized-tilt ground-based PV (GPV) in a location close to their cultural heritage.

Building Model Design	Angel	Bunny	Colossus	Inunnguaq	Pyramid	The Thinker	Tree
Cultural Heritage Location	Athens	Stanford	Athens	Toronto	Cairo	Paris	Lomé
Total PV Size (MW)	3.42	1.73	7.70	2.21	1.76	4.79	0.87
Optimal Building Face 0 Azimuth	255.20	65.54	318.75	72.00	278.96	154.78	225.62
Building Annual Energy (MWh)	2655.03	1506.49	5416.32	1609.41	2379.15	2438.91	1014.03
Optimized GPV Annual Energy (MWh)	8406	4145	6674	1549	6029	7839	4902
Total PV Active Area (m²)	4953.70	3027.47	11,138.02	3040.16	3141.53	4186.42	1206.77
Building Ground Footprint (m²)	14,572	7342	32,764	9404	7482	20,404	3708
Conventional PV Area (m²)	28,713.20	14,395.57	64,638.47	22,616.52	12,374.22	45,442.86	4438.61
Building Energy Yield (MWh/MW)	775.32	873.14	703.46	728.26	1353.12	508.64	1163.71
GPV Energy Yield (MWh/MW)	1446.58	1754.68	1446.58	1375.68	1786.71	873.09	1384.90
Building Energy Density (MWh/m²)	0.32	0.36	0.81	1.04	0.39	0.31	0.21
GPV Energy Density (MWh/m²)	0.17	0.21	0.17	0.13	0.25	0.09	0.27
Building Power Density (W/m²)	407.38	416.25	1153.66	1426.69	291.64	611.68	177.76
GPV Power Density (W/m²)	119.26	119.85	119.12	97.71	142.09	105.52	196.32

Table 4. Energy performance of different decimation levels of The Thinker building with an optimal azimuth in Paris.

Decimation Level	Extremely Low (Low Resolution)	Low (Base Case)	Medium (High Resolution)
Total PV Size (MW)	4.98	4.79	2.84
Optimal Building Face 0 Azimuth (°)	110.27	154.78	133.70
Building Annual Energy (MWh)	2424.21	2438.91	1474.72
Building Ground Footprint (m²)	7839	7839	7839
Total PV Active Area (m²)	21,184	20,404	12,072
Building Energy Yield (MWh/MW)	486.96	508.64	519.83
Building Energy Density (MWh/m²)	0.31	0.31	0.19
Building Power Density (W/m²)	635.06	611.68	361.90

Table 5. Building height sensitivity simulation results for The Thinker model design in Paris with an optimal azimuth and a comparison with conventional GPVs.

Building Height (m)	50	100	300	500	750
Total PV Size (MW)	0.36	2.65	32.83	96.82	224.11
Optimal Building Face 0 Azimuth (°)	164.78	154.78	154.78	154.78	144.78
Building Annual Energy (MWh)	187.43	1354.69	16,608.22	48,911.45	113,130.28
Optimized GPV Annual Energy (MWh)	1212	4831	43,592	121,158	272,490
Total PV Active Area (m²)	313.92	2314.80	28,662.30	84,529.73	195,664.65
Building Ground Footprint (m²)	1530	11,282	139,696	411,986	953,642
Conventional PV Area (1000 m²)	3.98	30.43	375.46	1107.12	2562.71
Building Energy Yield (MWh/MW)	521.28	510.96	505.91	505.20	504.81
GPV Energy Yield (MWh/MW)	873.09	873.09	873.09	873.09	873.09
Building Energy Density (MWh/m²)	0.15	0.28	0.38	0.40	0.42
GPV Energy Density (MWh/m²)	0.08	0.08	0.08	0.08	0.08
Building Power Density (W/m²)	296.66	548.80	753.09	799.09	822.44
GPV Power Density (W/m²)	90.24	87.14	87.44	87.45	87.45

Table 6. Comparison of the proposed software and existing BIPV modeling tools based on the IEA-PVPS investigation [63].

System	Tiling Complex Geometry	Building Geometry Modelling	Weather Data Inputs	PV Modules and Inverter Data	System Layout and Array Configuration	POA Irradiance	Shading Evaluation	PV Energy Conversion Simulation	PV System Losses
Skelion	Not implemented	Create 3D model	Input from Meteonorm 8.1	Manual input PV module power rating	Reposition but not define array configuration	Perez model	Shading factor analysis based on building geometry	Built-in empirical model	Manual input based on PVsyst results
SAM	Not implemented	Create simplified 3D model	Built-in Meteonorm 8.1	Manual input detailed specifications	Reposition and reconfigure façade system array	Perez model	Shading calculator based on simplify geometry	Built-in equivalent circuit model	Simulation
PVsyst	Not implemented	Import 3D model in COLLADA format	Built-in Meteonorm 8.2	Input detailed specifications via PAN/OND files	Reposition and reconfigure façade system array	Perez model	Shading factor analysis based on building geometry	Built-in equivalent circuit model	Simulation
BIMsolar	Not implemented	Import 3D model in Skp format	Input from Meteonorm 8.1	Manual input detailed specifications	Reposition and configure case system array	Ray tracing	Ray tracing	Built-in equivalent circuit model	Simulation
Ladybug Tools	Not implemented	Create 3D model	Input from Meteonorm 8.2	No Input	No array configuration defined	Ray tracing	Ray tracing	Calculation based on formula	Manual input based on PVsyst results
*PVSOL**	Not implemented	Import 3D model in COLLADA format	Built-in Meteonorm 8.2	Input detailed specifications via PAN files and inverter template	Reposition and reconfigure façade system array	Hay and Davies model	Near shade calculation based on building geometry	Built-in equivalent circuit model	Simulation
Solarius PV	Not implemented	Import 3D model in IFC format	Built EOI Meteonorm 7.1	Manual input detailed specifications	Reposition and configure case system array	Perez model	Manual input shading factor	Built-in empirical model	Manual input based on PVsyst results
INSIGHT	Not implemented	Create 3D model	Built-in Autodesk Climate Server	No input	No array configuration defined	Ray tracing	Ray tracing	Calculation based on formula	Manual input based on PVsyst results
Proposed Software	3D model input into Blender	Import 3D model into Blender and export text files	Input from NSRDB weather database	Input from CEC database or user specified inputs	Multiple orientation and tilt angle modeling	Perez model	Manual input shading factor	Equivalent circuit model based on SAM	Simulation

Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

© 2024 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Share and Cite

MDPI and ACS Style

Chin, A.W.H.; Hayibo, K.S.; Pearce, J.M. Open-Source Software for Building-Integrated Photovoltaic Tiling for Novelty Architecture. Designs 2024, 8, 118. https://doi.org/10.3390/designs8060118

AMA Style

Chin AWH, Hayibo KS, Pearce JM. Open-Source Software for Building-Integrated Photovoltaic Tiling for Novelty Architecture. Designs. 2024; 8(6):118. https://doi.org/10.3390/designs8060118

Chicago/Turabian Style

Chin, Alexander W. H., Koami Soulemane Hayibo, and Joshua M. Pearce. 2024. "Open-Source Software for Building-Integrated Photovoltaic Tiling for Novelty Architecture" Designs 8, no. 6: 118. https://doi.org/10.3390/designs8060118

APA Style

Chin, A. W. H., Hayibo, K. S., & Pearce, J. M. (2024). Open-Source Software for Building-Integrated Photovoltaic Tiling for Novelty Architecture. Designs, 8(6), 118. https://doi.org/10.3390/designs8060118

Article Menu