1876-6102 © 2016 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).Peer-review under responsibility of the organizing committee EENVIRO 2015doi: 10.1016/j.egypro.2015.12.269

Energy Procedia 85 (2016) 9 – 16

ScienceDirect

Sustainable Solutions for Energy and Environment, EENVIRO - YRC 2015, 18-20 November 2015, Bucharest, Romania

Informed Geometries. Parametric Modelling and Energy Analysis in Early Stages of Design

Ionuţ Antona* and Daniela Tănasea* aUAUIM - University for Architecture and Urbanism Ion Mincu,18-20 Academiei Street, Bucharest 010014, Romania

Abstract

Recent developments in computational design tools have bridged a gap between a well-established parametric building modeling[1] and analysis or simulation software such as EnergyPlus[2], Radiance[3], Daysim[4] and OpenStudio[5], opening up the possibility for architects to use the computational power to model and simulate real environmental behavior of the architectural artefact and its components. Now architects are able to evaluate the behavior of a project, whether it is a building, a city, a landscape or infrastructure and a new road towards an architecture based on performance is opened[6]. Therefore we can put the idea of performance as a precedent to shape development and the architectural form becomes informed by the performative aspects. We can use various computational tools in order to gather qualitative and quantitative aspects of the architectural artefacts performance in the early stages of design, and we can go further from just optimizing a form after it has been defined. This paper will discuss the implications of parametric modeling and energy analysis in architectural design and will present the authors research in developing architectural forms using the above described tools. © 2015 The Authors. Published by Elsevier Ltd. Peer-review under responsibility of the organizing committee EENVIRO 2015.

Keywords: Computational design; parametric design; energy analysis; grasshopper

1. Introduction

The idea of performance is a generic one, dependent on the context of the specific project and can be understood in a very broad sense, reaching fields like economy, spatial planning, society, culture or technology[7]. Today’s computational environment allows architects to determine with great precision and to visualize that which remained

E-mail address: ionut.anton@idz.ro

Available online at www.sciencedirect.com

© 2016 Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).Peer-review under responsibility of the organizing committee EENVIRO 2015

10 Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16

unseen and a series of digital analysis tools can inform the design process on certain aspects of performance pertinent to a project. This shift and focus on performance in architecture is driven by the need for more resilience in architecture and design, a central theme of contemporary architecture.

Traditionally, architects design a building or a space and shape it according to their will. After the space is designed the project is delivered to an engineering team that tries to analyze the environmental performance of the project and develops various energy scenarios and selects one of the best scenarios available[8]. The problem in this workflow is that the energy analysis rarely changes the form of the architectural artefact, and the change is implemented with a lot of effort.

Although engineers use advanced computational tools for performance analysis (and here we include energy analysis), they are employed without affecting the architectural form and without being used directly as morphogenetic agents in a process of form generation. The new computational tools available to architects and engineers can be used as more than optimization tools for already established architectural forms. Combined with parametric modeling, these tools can lead to an active integration of the analysis means in the development of architectural form in the early stages of design [6,9].

Architects like Brank Kolarevic and Rivka Oxman advocate for an approach in which architectural form is tied to a result of a performance analysis that induces a feedback loop into the design of the architectural form. Kolarevic points out that in these cases the result, based on performance criteria may lead to a performance based optimized solution, but it might not be the best solution in respect to its esthetic criteria, and architects need to find a way to manage the two. A solution is to include in a parametric model all the relations that lead to the development of a form under the emergent actions of performance driven forces simulated in a digital environment[10]. Rivka Oxman goes further and develops a morphogenetic algorithm based on performance that entails an interactive shape development combined with optimization and performance analysis[6].

By combining performance with formal generation, performance becomes just another parameter in an algorithmic development of architecture, where shape is negotiated by several criteria. By using a parametric model in which formal generation rules are transformed into an algorithm, one can easily implement analysis tools that provide precise feedback into the generative process. Here analysis can contribute actively in the generative process of the architectural form.

Performance shouldn’t be understood as a mere conditioning of the project based on a generic solution applied as a solution to a great number of practical problems. This reductive approach based on performance and efficiency can lead to a very functionalist design. The domain of performance criteria must be extended and seen as an opportunity to develop the generative design process based on local information pertinent to each project, such as cultural, technical and environmental conditions. More over this approach must find a way to incorporate performance criteria into a project leveling creativity with efficiency[11].

2. Post-optimization and Pre-rationalization

Recent developments in the field of computation can lead to the possibility to use the digital environment to simulate the real behavior of architectural artefacts and its components. This analysis must be made in all stages of design, construction and usage, from conceptual design to the building decommission[12].

A traditional design workflow used simulation after the formal development of the architecture of a building. In this situation the project would be very rigid and could implement only small changes in the formal arrangement of the project. If the performance criteria analysed was indeed a decisive factor in the project development, the optimization process would have been a very laborious. Therefore after each change in the project the analysis model must be rebuilt and a new series of analyses would be made. The new result lead to new changes and the process was reiterated until an optimum was reached.

The focus in design now lies with the capacity of the digital environment to synthetize, not only to calculate. Today’s analysis software, although specialized on different types of analyses, are developed for interoperability. Performance issues are addressed no more in an isolated mode, in a linear progression, but all at the same time. The aim is to get an overall performance analysis of the whole building, linking all criteria, in order to be able to draw conclusions and develop solutions at a holistic level[5].

Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16 11

The integration and control of several levels of information is one of the most attractive promises made by recent developments of the digital tools. The final architectural artefact is no more an imposition of the architects will, but becomes a response based on performance, on analysis and collaboration between a great number of variables and real data.

Greg Lynn defines design as an abstract space where active modelling forces, informed by real data act as form generators[13]. The aim is to obtain a formal complexity that fuels construction, usage and selection processes. Through this approach the final form of the architectural artefact is determined by parameters based on performance. This formal generation should not be interpreted only through formal aspects[11], but must include also subjective and objective criteria pertinent to the architectural concept.

Here we must make a division between two processes, a Post-optimization and a Pre-rationalization process. Even though both use digital analysis tools they are fundamentally different. Both aim at the optimization of geometry, but they differ in the way that performance criteria are integrated in the project. Post-optimization is an approach in which one tries to find a solution for a shape that has been derived from purely architectural esthetic reasoning[14]. The computational tools here serve a purpose to remodel the design in order to optimize. Frank Gehry states that here the ”computer is a tool not a partner, an instrument for catching a curve, not for inventing it”[15]. In such a case, analysis software are used after the design is finalized in order to validate the architectural artefacts performance. This optimization method seeks to maximize the efficiency of a project with an already established and fixed geometry[16].

Pre-rationalization, on the other hand, refers to a design process that uses a series of geometric rules and methods that generate a viable solution. The starting point is no longer a geometry derived solely from an architect’s will, but from a series of rules that incorporate several parameters based on performance, material and fabrication and an interval of minimum and maximum bounds in which the parameters can vary. In most cases, the architectural form is a result of an evolutive and geometrical transformation process based on trial and error. This Pre-rationalization process contains all the constraints and assembly logic and the geometric rules for design generation.

Present analysis software combined with this kind of parametric modeling aim to develop a system for automatization of change in design. Any change in the results of the analysis is immediately fed back to a loop in the generative parameters of the architectural form and the process is reiterated. The use of computational tools make this process less cumbersome and sometimes instantaneous, with direct visualization of the results on the architectural geometry. This process aims to make a direct link between simulation and formal generation, not only an optimization of a predefined shape.

In contemporary architectural practice the digital environment is no longer used as a mere tool for visualization, but as a formal generator. Parametric design establishes relationships between different agents that influence the final form, which no longer is arbitrary. The interest shifts from form making to form finding and architectural practice transforms from a predefined fixed design to a process design. Therefore both the final result and the becoming process of the form are as important to the current design practice[17].

One of the approaches used is one in which genetic algorithms are combined with analysis and intermediate visualization methods[12]. The genetic (evolutive) approach is a design principle based on a series of iterations that uses trial-error method. In a genetic (evolutive) design, performance criteria are presented as parameters with constraints. The process outputs only the solutions that reach a certain optimum based on the input parameters. This process is executed recursively and follows several iterations until the results achieve the desired performance objectives. Using this process a series of viable solutions are generated automatically and the designer has a wider solution space to explore.

Along the process intermediate results are also revealed and can become relevant to the project by stimulating the creativity of the architect. The aim is not to obtain a perfect optimum, a maximum possible of performance, but rather to negotiate criteria along the process and fine tune the objectives. Therefore the ability to develop new models is direct dependent on the architect’s ability to use and control these tools. The architect edits the rules of the morphogenetic process and the emergent shapes are determined by quantifiable performance criteria and the creativity of the architect. The generative capacity of the computational environment is directly linked to the way that the architect interprets and the results and modifies the performance criteria in a dynamic process.

12 Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16

3. Parametric design and energy analysis tools

This paper will focus next on the use of parametric modeling tools, such as Grasshopper[18,19] and Dynamo[20,21] along with energy analysis software such as Ecotect[22], EnergyPlus[2], Radiance[3], Daysim[4] and OpenStudio[5]. Grasshopper is a visual algorithm editor that builds parametric and algorithmic models on the freeform modelling platform Rhinoceros[23]. Grasshopper allows architects and engineers to develop algorithms for design by building on parametric and rule based systems. The programming capabilities and interoperability with other analysis software of Grasshopper is extended by a series of plugins dedicated to energy analysis, such as GECO[24], Diva for Rhino[4] and Ladybug[25]. These plugins provide the necessary tools to integrate the above mentioned analysis software into a parametric modeling environment. Our research has implemented such analysis for solar radiation calculated on faces of a freeform architectural artefact, solar access analysis in a complex urban environment and daylight analysis for freeform skylight design.

Fig. 1. (a) Solar radiation calculated on faces of a freeform architectural artefact, (b) daylight analysis for freeform skylight design, (c) solar access analysis in a complex urban environment

Furthermore, by implementing a genetic algorithm through Galapagos[26], we were able to devise an algorithm that developed a variation in geometry that searched for the optimum configuration for a canopy with several performance criteria. These criteria were the least solar energy absorbed on its faces and at the same time the most shadow area. Both objectives were used as criteria for a genetic algorithm that searched through a series of parameters that established the general shape of the canopy.

The algorithm started from a surface composed of large triangular subdivision. The position of each vertex in a grid on the vertical axis is determined by a series of parameters that are also genomes for the Galapagos search. A solar analysis done on the large faces of the canopy determines values that influence the aperture of each face. The less solar energy received, the larger the aperture. The sum of the solar energy absorbed is one of the search criteria. The algorithm also calculates the shadow area on the ground plane of the canopy which is used as the search objective criteria. The evolutionary solver Galapagos determines the optimum genome (sum of parameters) that influence the geometry in order to obtain the least solar energy absorbed and the maximum shadow area.

Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16 13

Fig. 2. Genetic algorithm variation for a canopy with the least solar energy absorbed and the maximum shadow area

4. INformed Geometries workshop

To further explore the capabilities of parametric modeling combined with performance analysis the authors have developed and lead a workshop as a research method. The aim of the workshop was to study the interaction between digital tools and practicing architects. Although the workshop has a highly detail theme and predefined tools the results are not predefined. Using the tools at their disposal, participants have the opportunity to explore and test their own vision. Another aim of the workshop was to fabricate the developed projects into physical artefacts. The fabrication component is crucial to the development of parametric design in order to establish a full design workflow, from concept to physical materialization. Thus the workshop becomes a research through design method that tests the way in which digital tools can be applied to architectural practice. The research transcends a new expression of architecture and its dependence on digital tools and goes into substance researching the methods to think and make architecture and design using digital tools.

The INformed Geometries workshop objective was to use digital form finding tools and generate geometries influenced by real data and performance analysis and simulation of physical forces. The theme of the workshop emphasized the link between the physical and digital environments. Furthermore a central part of the workshop and a generator in the design was the use of 3d printing technology that allows the development at a small scale of objects with a high degree of complexity. The fabrication logic was integrated in the design of the objects by exploring soft geometries developed as thin surfaces with very complex details.

The models resulted in the workshop are imagined as pertaining to virtual worlds, but are the participants own interpretations of architectural artefacts. The proposed image is willingly pushed to the limits of reality in order to emphasize the capabilities of digital tools, both as control over a complex geometry informed by real data, but also as a result of a 3d printing process. The resulting geometry is determined by relating several performance criteria that negotiate influence over the architectural shape.

14 Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16

Fig. 3. INformed Geometries, the resulting geometry and the correlation of performance criteria

The research used the digital environment for performance analysis and simulation of physical forces in order to implement the results of these simulations and analysis in the process of formal generation. We created feedback loops in order to develop architectural artefacts that are influenced by physical conditions. The resulting form is an informed response that archives the performance criteria established by the generating algorithm.

Global shapes are generated by the interaction of physical forces interior or exterior to the system. We assume that the initial geometry has a series of elasticity relation defined between its components and what are the fixed elements and the forces acting on the system. A process of dynamic relaxation[27] transformed the geometry until it reaches an equilibrium. We overlaid several performance criteria, such as structural and solar analysis on the same model, by using Grasshopper and several plugins that communicate in real time and exchange data with external analysis software. The data resulted in the analysis was used as form generator for each of the starting surface components.

Data from the structural and environmental analysis was implemented at component level over the analyzed surface and each component was variably influenced by the resulting data. The aim was to influence each of the geometric components on the surface with different data. We obtained therefore a series of patterns that could be applied on each of the five category of shapes defined as starting geometries.

If the geometry was influenced by only one set of data, pyramids with variable height were constructed or variable width apertures were made in each component. If a component was influenced by two sets of data truncated pyramids were constructed where the upper aperture was determined by one criteria and the height was dependent on the other criteria. All the models were constructed using quad surfaces that were then passed through a Catmull-Clark surface subdivision[28] in order to soften the resulting form.

Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16 15

The fabrication constraint of 3d printing was satisfied by offsetting the resulted geometry and thus creating a volume. The correctly constructed volumes were also watertight and adjustments to the parameters and the initial geometry had to be made to achieve this constraint. The resulted objects also needed to be inscribed in a 7x7x7 centimeters volume and the number of initial components and details was also influenced by the capabilities of the 3d printer.

Fig. 4. INformed Geometries, the 3d printed architectural artefacts.

5. Conclusions

Architects are no longer bound to Post-optimization methods, but can use parametric modeling combined with various performance analysis software in order to influence the architectural form in the early stages of design. This opens up new opportunities in architectural practice where interdisciplinary teams can work together for more resilient architectural developments. This method can lead to a synergy between performance and the architects creativity that leads to more informed, efficient and meaningful architectural artifacts.

References

[1] R. Woodbury, Elements of Parametric Design, 1st ed., Routledge, 2010. [2] D.B. Crawley, L.K. Lawrie, F.C. Winkelmann, W.F. Buhl, Y.J. Huang, C.O. Pedersen, et al., EnergyPlus: creating a new-generation

building energy simulation program, Energy and Buildings. 33 (2001) 319–331. [3] K. Lagios, J. Niemasz, C.F. Reinhart, Animated building performance simulation (abps)-linking rhinoceros/grasshopper with

radiance/daysim, Proceedings of SimBuild. (2010). http://trafficlight.bitdefender.com/info?url=http%3A//www.ibpsa.us/sites/default/files/publications/SB10-DOC-TS06A-03-Lagios.pdf&language=en_US (accessed September 29, 2015).

[4] J.A. Jakubiec, C.F. Reinhart, DIVA 2.0: Integrating daylight and thermal simulations using Rhinoceros 3D, Daysim and EnergyPlus, in: Building Simulation-12th Conference of International Building Performance Simulation Association, 2011. http://trafficlight.bitdefender.com/info?url=http%3A//www.researchgate.net/profile/John_Jakubiec/publication/264353052_DIVA_2.0_integrating_daylight_and_thermal_simulations_using_Rhinoceros_3D_DAYSIM_and_EnergyPlus/links/53da0f460cf2a19eee881e4f.pdf&language=en_US (accessed September 29, 2015).

[5] R. Guglielmetti, D. Macumber, N. Long, OpenStudio: an open source integrated analysis platform, in: Proceedings of the 12th Conference of International Building Performance Simulation Association, 2011. http://trafficlight.bitdefender.com/info?url=http%3A//www.nrel.gov/docs/fy12osti/51836.pdf&language=en_US (accessed September 29, 2015).

[6] R. Oxman, Performance-based design: current practices and research issues, International Journal of Architectural Computing. 6 (2008) 1–17. [7] B. Kolarevic, Digital morphogenesis, in: Architecture in the Digital Age: Designing and Manufacturing, Taylor and Francis, 2003: pp. 12–28. [8] M. Moradi, Making-Intelligent Place of Buildings in Parametric (Algorithmic) Architecture, (n.d.).

http://trafficlight.bitdefender.com/info?url=http%3A//www.ojceu.ir/main/attachments/article/31/J.%2520Civil%2520Eng.%2520Urban.%2C%25204%2520%282%29%2520103-109%2C%25202014.pdf&language=en_US (accessed September 29, 2015).

[9] B. Kolarevic, Architecture in the Digital Age: Design and Manufacturing, 1st ed., Taylor & Francis, 2003. [10] B. Kolarevic, A. Malkawi, Performative architecture: beyond instrumentality, Spon Press, 2005. [11] B. Kolarevic, Computing the Performative, in: B. Kolarevic, A. Malkawi (Eds.), Performative Architecture: Beyond Instrumentality,

Routledge, 2005: pp. 195–202. [12] A. Malkawi, Performance Simulation. Research and Tools, in: B. Kolarevic, A. Malkawi (Eds.), Performative Architecture Beyond

Instrumentality, Routledge, New York, 2005: pp. 87–95. [13] G. Lynn, Animate Form, Princeton Architectural Press, New York, 1999.

16 Ionuţ Anton and Daniela Tănase / Energy Procedia 85 (2016) 9 – 16

[14] C. Ceccato, Computing and Constructing Continous Differentiation, in: A. Menges (Ed.), Material Computation: Higher Integration in Morphogenetic Design, John Wiley & Sons, Chichester, 2012: pp. 96–103.

[15] V. Bodden, F.O. Gehry, Frank Gehry, The Creative Company, 2008. [16] A. Rahim, Performativity: Beyond Efficeincy and Optimization in Architecture, in: B. Kolarevic, A. Malkawi (Eds.), Performative

Architecture: Beyond Instrumentality, Routledge, New York, 2005: pp. 179–192. [17] S. Marcus, Paradigma formei (frumoase), in: Paradigme Universale, Paralela 45, Piteşti, 2011: pp. 48–50. [18] M. Day, Rhino Grasshopper, AEC Magazine. (2009). http://aecmag.com/index.php?option=com_content&task=view&id=293&Itemid=32

(accessed July 20, 2011). [19] Robert McNeel & Associates, Grasshopper, (n.d.). http://www.grasshopper3d.com/. [20] Autodesk, Dynamo, (n.d.). http://dynamobim.com/. [21] I. Keough, Dynamo: designing a visual scripting interface for the Revit API (notes), Also Se E Https://github. com/ikeough/Dynamo/wiki

for More Information about Coding in Dynamo. (2011). [22] Autodesk, Ecotect Analysis - Sustainable Building Design Software, Autodesk. (2011).

http://usa.autodesk.com/adsk/servlet/pc/index?id=12602821&siteID=123112 (accessed July 27, 2011). [23] Robert McNeel & Associates, Rhinoceros 3d, (n.d.). https://www.rhino3d.com/. [24] U. Frick, T. Grabner, example files geco - Grasshopper, (2010). http://www.grasshopper3d.com/group/geco/forum/topics/example-files-geco

(accessed July 27, 2011). [25] M. Roudsari, M. Pak, A. Smith, Ladybug: A Parametric Environmental Plugin for Grasshopper to Help Designers Create an

Environmentally-Conscious Design, (2014). http://trafficlight.bitdefender.com/info?url=http%3A//www.ibpsa.org/proceedings/BS2013/p_2499.pdf&language=en_US (accessed September 29, 2015).

[26] D. Rutten, Galapagos: On the logic and limitations of generic solvers, Architectural Design. 83 (2013) 132–135. [27] D. Piker, Kangaroo - Grasshopper, (2011). http://www.grasshopper3d.com/group/kangaroo (accessed July 28, 2011). [28] G. Piacentino, Weaverbird: Topological Mesh Editing for Architects, Architectural Design. 83 (2013) 140–141.