Design for structural and energy performance of long span buildings using geometric multi-objective optimization
Introduction
In the conceptual design of buildings, many traditional optimization methods have seen only limited application, despite the emphasis contemporary designers place on building performance. This is largely due to the complex requirements of contemporary architecture, and the fact that human intuition and judgment are still central to the design process. Even as early as the conceptual design phase, architects must simultaneously consider and prioritize a multitude of interrelated design objectives, and while an increasing number of these objectives are quantitatively measureable, many are not. Two of the most important objectives related to building performance are the embodied and operational energy used in a building’s materials and operations, respectively. Often, the goals of reducing each of these quantities trade off with one another, as well as with other qualitative design goals, in unexpected ways.
In contemporary design, a high-performance, sustainable building has been identified as one that minimizes energy consumption throughout the four main stages of a building’s lifetime: materials manufacturing, construction, use and maintenance, and end of life [1]. The International Energy Agency estimates that buildings accounted for nearly a third of total final energy consumption globally in 2013 [2], a number which includes the substantial embodied energy of building materials as well as the operational energy used to keep buildings lit, heated, and cooled. The need for a reduction in energy consumption and carbon emissions due to buildings has been well documented. In light of this, a conceptual designer could simply convert every aspect of the design to a unit of emissions and run a traditional optimization to find a single solution. However, in practice this would hamper the ability of designers to express preference, and also ignore financial constraints and other architectural complications that influence the development of a real building. As such, architects often find a pure performance optimization approach reductive, overly simplified, and deterministic, which can lead to resistance towards the adoption of optimization methods in design [3].
Furthermore, for long span roofs and other large structures with specific spatial requirements, consideration of form can dominate the conceptual design phase. This is especially true because structural efficiency depends more on the geometry of a building than on material, sizing, and other building characteristics developed in later stages, and because structural material makes up a more sizable portion of the overall embodied energy. For example, De Wolf [4] found that the amount of embodied energy within a building’s 50-year lifecycle usage can range from 4 to 22% of the total. When considering carbon emissions instead of energy consumption, the embodied portion can rise to as high as 80% of total emissions, depending on which exact source is consulted. As shown by Kaethner and Burridge [5], the material contained in the substructure and superstructure can be responsible for over half of these embodied emissions.
With more emphasis on cutting operational energy usage while pushing towards net zero buildings, the embodied energy of future buildings will make up an increasingly larger portion of total energy usage over their lifetimes (Fig. 1) [6]. Consequently, this paper focuses on multi-objective optimization (MOO) with structural efficiency and operational energy efficiency as the two measurable objectives, since a MOO approach gives the designer flexibility but encompasses the most significant quantitative performance goals of contemporary architecture.
Although MOO has demonstrated a greater potential than traditional optimization to assist conceptual designers in generating and deciding between high-performing, early-stage designs, it too has seen only limited use in practice. This lack of application can be attributed to the complicated model translation that must occur between design and analysis software, the often linear process in which members of a design team are only given small latitude to ‘optimize’ for their own performance goals without reference to other disciplines, and the difficulty of using optimization within a process that includes subjective preferences and design goals that are difficult to formulate numerically [7]. Related fields such as aerospace, mechanical, and other pure engineering design disciplines have been more successful than architecture in overcoming some of these obstacles [8]. The differences in scale, production, and customization of buildings when compared with airplanes or cars have contributed to a building industry that is more fragmented and mostly unexposed to optimization workflows. For a conceptual MOO procedure to become popular with building designers looking for original, expressive forms, researchers must overcome difficulties of non-quantitative objectives and disconnected disciplines while showing how an integrated process can lead to a diverse range of design outcomes that meet a variety of aesthetic preferences.
Many academic researchers have addressed the limitations of multi-optimization for use in conceptual design, but few have studied the strong link between architectural form and different performance metrics simultaneously in a way that demonstrates the significant potential of MOO to influence the leading edge of architectural design. A large number of the major contributions in the field, which are described in detail and cited in the next section, have been restricted to geometries that are primarily made of rectangular boxes, which are easy to model in terms of energy usage. However, even within the typology of the long span roof, there are a wide variety of architectural forms and corresponding building shapes and structural systems that could be optimized for performance. In addition, researchers have been largely unable to define a clear way for architects to interact with MOO data, which can include performance feedback from multiple engineering disciplines in different units and scales, in a way that leads to good design decisions. In order to have greater impact on innovative and creative architectural practices, it is important to develop methodologies that effectively navigate meaningful tradeoffs and produce design examples that are applicable to a wider range of building geometries.
In response, this paper demonstrates how MOO can be used to generate geometrically diverse architectural design solutions in different climate regions through three case studies of buildings with long span roofs. Since selecting the right building shape and form in the early stages has a large effect on the overall success of a building with demanding structural and spatial requirements, the case studies focus on these large-scale conceptual design decisions.
The optimization method in this paper uses simulation to focus simultaneously on both the embodied energy found in structural material and the operational energy of the building. The case study results are presented in terms of overall energy requirements, but the embodied and operational components are kept independent since the two are not always equal when time, financing, and other practical realities of construction are taken into account. These results demonstrate the utility of separating structural efficiency (primarily upfront emissions and cost) and operational energy efficiency (emissions and cost over time) in optimization for conceptual design, as well as showing the effect of the two separate objectives on architectural form. Overall, this paper illustrates how the application of MOO can yield a wide range of expressive, high-performance designs, provides increased awareness of how architects might respond to particular climates while designing long span buildings, and contextualizes structural efficiency within the broad goals of sustainable design.
Section snippets
Optimization for structure or energy
This research builds on a wide body of existing scholarship concerning the integration of visual criteria into optimization algorithms, geometry optimization for building performance, and multi-objective optimization in architectural design. A brief overview of major contributions is given here, beginning with research that focuses exclusively on either structure or energy usage. Mueller and Ochsendorf [9] created structureFIT, which is a browser-based conceptual design tool that allows users
Methodology
To demonstrate the capabilities of MOO in the conceptual design of long span structures, three case study models were developed and analyzed. The procedure for testing the case studies included the development of a parametric design space, using a MOO algorithm to find the Pareto front of best performing designs, visualizing the results in a way that designers can easily compare between samples, and analyzing the overall data across different forms and contexts (Fig. 2). The design lessons
Shape and location of Pareto fronts
The performance and optimization results for all three case studies are presented in this section. Fig. 5 gives the Pareto front found by the algorithm between the two objectives for each structural configuration and location. In each of these plots, a single design generated at any point during the algorithm is represented as a grey circle, while the Pareto optimal points are colored black. In a few cases, these plots display one or two additional Pareto points than were not returned in the
Discussion and summary of contributions
This paper makes three key contributions to the broad goal of encouraging optimization in integrated conceptual design for high-performance buildings.
First is the demonstration of how a multi-objective optimization methodology for structural efficiency and operating energy efficiency can be used to generate a geometrically diverse range of high-performing designs. The results of the various case studies presented in this paper show that although finding expressive architectural forms and
Future work and concluding remarks
This research could be extended in a number of ways. One next clear step is to look at tradeoffs in terms of monetary cost in addition to energy usage. Although cost and energy are largely correlated, the nature of specific tradeoffs could fundamentally change with this different perspective. Analyzing cost in addition to energy could also allow for more specific study of the effects of upfront costs and price changes over time on optimized architectural form. In addition, there are many
Acknowledgements
The authors are grateful to Professor John Ochsendorf for his assistance in selecting and formulating the case studies. The authors would also like to thank J. I. F. de Oliveira for his contributions related to the development of a multi-objective optimization tool for Grasshopper.
References (53)
- et al.
Energy use in the life cycle of conventional and low-energy buildings: a review article
Energy Build.
(2007) - et al.
Combining structural performance and designer preferences in evolutionary design space exploration
Autom. Constr.
(2015) - et al.
Low-energy design: combining computer-based optimisation and human judgement
Build. Environ.
(2002) - et al.
Applying multi-objective genetic algorithms in green building design optimization
Build. Environ.
(2005) - et al.
Multiobjective optimization of building design using TRNSYS simulations, genetic algorithm, and Artificial Neural Network
Build. Environ.
(2010) - et al.
Optimizing tradeoffs among housing sustainability objectives
Autom. Constr.
(2015) Multicriteria optimisation energy-saving buildings
Build. Environ.
(1997)- et al.
Structural analysis of Pareto-optimal solution sets for multi-objective optimization: an application to outer window design problems using Multiple Objective Genetic Algorithms
Build. Environ.
(2010) - et al.
Genetic-algorithm based approach to optimize building envelope design for residential buildings
Build. Environ.
(2010) - et al.
Towards a multi-objective optimization approach for improving energy efficiency in buildings
Energy Build.
(2008)
A multi-objective feedback approach for evaluating sequential conceptual building design decisions
Autom. Constr.
Building energy optimization in the early design stages: a simplified method
Energy Build.
Balancing energy efficiency and structural performance through multi-objective shape optimization: case study of a rapidly deployable origami-inspired shelter
Energy Build.
Ancient vernacular architecture: characteristics categorization and energy performance evaluation
Energy Build.
Optimum overhang dimensions for energy saving
Build. Environ.
EnergyPlus: creating a new-generation building energy simulation program
Energy Build.
jMetal: a Java framework for multi-objective optimization
Adv. Eng. Softw.
Adding advanced behavioural models in whole building energy simulation: a study on the total energy impact of manual and automated lighting control
Energy Build.
AIA Guide to Building Life Cycle Assessment in Practice
Material quantities in building structures and their environmental impact
S.M. Thesis
Embodied CO2 of structural frames
Struct. Eng.
MDO: assessment and direction for advancement—an opinion of one international group
Struct. Multidiscip. Optim.
Genetic algorithms for optimization of building envelopes and the design and control of HVAC systems
J. Sol. Energy Eng.
Cited by (77)
Multi-Objective optimization of energy consumption pattern in order to provide thermal comfort and reduce costs in a residential building
2024, Energy Conversion and ManagementLow-carbon design: Building optimization considering carbon emission, material utilization, and daylighting
2024, Journal of Cleaner ProductionIntegrating early assessment of robotic constructability into design optimization of a standalone classroom
2024, Automation in ConstructionReducing embodied carbon in structural systems: A review of early-stage design strategies
2023, Journal of Building Engineering