The digital transformation of the AEC industry through BIM has improved productivity during detailed design and construction planning phases. Early design choices influence a project's success but ...have yet to benefit from BIM-based approaches. This paper investigates the feasibility and acceptance of employing Generative Design (GD) to optimize early Mechanical, Electrical, and Plumbing (MEP) designs in residential real estate for space efficiency. Interviews indicate the main issue is acceptance due to the belief that a GD approach needs to be more robust. BIM is integrated with GD, utilizing the architectural layout (IFC) as input to generate design variants tailored to minimize technical space while ensuring installation feasibility. Robustness is assessed via Monte Carlo Simulation, revealing an estimated success rate of 99% (81% with 95% confidence). These results quantify the robustness of the approach, paving the way to broader acceptance of GD in the early phases of AEC projects.
•Optimized technical spaces in early-stage residential building projects.•Generative Design in BIM using a tuned Genetic Algorithm for optimization.•Identified algorithmic robustness as crucial for acceptance of Generative Design.•Generated multiple acceptable designs representing the Pareto Front in optimization.•Genetic Algorithm robustness proved with 99% success via Monte Carlo Simulation.
High-rise residential building façades (HRBFs), given their size and abundant façade information, pose a challenge for conventional parsing methods. This paper presents FaçadeGraph, an approach for ...parsing the information of HRBFs into hierarchical attributed graphs. The method decomposes HRBF information into five hierarchical layers: ternary, floor, unit, space, and component. The façade elements are identified as semantics information with geometric attributes. The topological relationships between the elements are classified into affiliation, connection, aggregation, and decoration. The efficacy of FaçadeGraph was evaluated through the analysis of 36 HRBFs in China. The result showed that FaçadeGraph is effective in transforming diverse façade designs into consolidated graphs for automated syntax analyses. The paper contributes to the knowledge body of façade design by serving as an analytical tool for design feature analysis and underlying the development of generative HRBF design.
•An attributed graph-based façade parsing approach is proposed.•A hierarchical structure is used to indicate façade elements for semantic analysis.•Universally applicable rules are defined for façade representation.•The framework is validated by a dataset of façades on 36 high-rise buildings.
Abstract
The benefits of additive manufacturing (AM) extend beyond the attributes of physical products and production processes they enable. Experience with AM can augment the way design is ...approached and can increase opportunities to pivot toward less familiar design tasks. We begin this qualitative study with a natural experiment made possible by an exogenous shock: the COVID‐19 pandemic. Through a three‐stage case study approach using a grounded theory‐building method, we contrast AM usage among a set of firms, half of which pivoted their resources away from their traditional production and toward a response to this shock. We engage in an abductive reasoning approach to consider common threads in AM capabilities that facilitated this pivoting. Our analyses suggest that the advanced use of generative design (GD), a category of computational technologies enabling novel and optimized design, is a critical attribute of these firms that ended up pivoting to make COVID‐related products. Specifically, firms with experience applying this capability demonstrated a unique ability to pivot during this shock and emphasized their valuation of AM‐enabled agility. We revisited these firms 2 years after initial contact and found that GD was associated with higher levels of innovation and was largely viewed by designers as a mechanism driving double‐loop learning. Overall, our study provides insights into the symbiosis between human and artificially intelligent GD, and the role of such symbiosis in advancing AM capabilities.
This work describes a generative method for the exploration of product shapes in the conceptual design stage. The method is based on three concepts: the notion of grammars to capture product ...appearance, the implementation of sketching transformation rules to produce design variations and the use of a parametric modeller to build shapes. We represent product solutions as 3D sketches using combinations of basic shapes arranged in simple and schematic product structures. This procedure allows creating many varied configurations with a minimal number of shapes, and facilitates the adaptation of the generative model to different products. The performance of the method is demonstrated through several examples from the literature.
•A generative method for aesthetic shape exploration in product design is proposed.•The model generates a high number of variations and may adapt to different products.•The method focuses on the visual appearance of products.•It includes several sketch-related creative strategies (reformulation, analogy …).
Long computational time poses a significant obstacle to the practical utilization of performance-based generative design (PGD) in the early design stage. This study proposes a knowledge-informed PGD ...optimization framework for sustainable buildings, aimed to mitigate the time issue by integrating a knowledge graph (KG) into PGD. The first component of the framework is a PGD-KG schema that represents the topological relations within PGD. A generation method is then proposed for automatically developing PGD-KG models from parametric design models that are enhanced with semantic information. Furthermore, cross-domain reasoning algorithms are developed to enable automated compliance checking and performance evaluation based on regulatory requirements and sustainable design strategies, respectively. The proposed framework is applied to a design project focused on optimizing module layout to minimize cooling energy and maximize daylighting. The results demonstrate that the proposed framework can generate a satisfactory number of Pareto-optimal solutions while reducing computational time by 73.25% compared with the general optimization framework.
•A knowledge graph for performance-based generative design (PGD-KG) is proposed.•Automated PGD-KG generation algorithm from parametric design models is developed.•Reasoning for automated compliance checking & performance evaluation is developed.•The PGD-KG optimization framework narrows the solution space of optimization.•The PGD-KG optimization framework saves computational time by 73.25%.
Display omitted
•State-of-the-art for biomimicry in additive manufacturing is summarized•Various forms of biomimicry in additive manufacturing are discussed•Categorization of biomimetic design ...approaches is presented•Present limits and challenges are outlined•Exciting capabilities and opportunities are discussed
This review article summarizes the current state-of-the-art for biomimicry in additive manufacturing. Biomimicry is the practice of learning from and emulating nature - which can be increasingly realized in engineering applications due to progress in additive manufacturing (AM). AM has grown tremendously in recent years, with improvements in technology and resulting material properties sometimes exceeding those of equivalent parts produced by traditional production processes. This has led to the industrial use of AM parts even in highly critical applications, most notably in aerospace, automotive and medical applications. The ability to create parts with complex geometries is one of the most important advantages of this technology, allowing the production of complex functional objects from various materials including plastics and metals that cannot be easily produced by any other means. Utilizing the full complexity allowed by AM is the key to unlocking the huge potential of this technology for real world applications – and biomimicry might be pivotal in this regard. Biomimicry may take different forms in AM, including customization of parts for individuals (e.g. medical prosthesis, implants or custom sports equipment), or optimization for specific properties such as stiffness and light-weighting (e.g. lightweight parts in aerospace or automotive applications). The optimization process often uses an iterative simulation-driven process analogous to biological evolution – with an improvement in every iteration. Other forms of biomimicry in AM include the incorporation of real biological inputs into designs (i.e. emulating nature for its unique properties); the use of cellular or lattice structures – for various applications and customized to the application; incorporating multi-functionality into designs; the consolidation of numerous parts into one and the reduction of waste, amongst others. Numerous biomimetic design approaches may be used – broadly categorized into customized/freeform, simulation-driven and lattice designs. All these approaches may be used in combination with one another, and in all cases with or without direct input from nature. The aim of this review is to unravel the different forms of biomimetic engineering that are now possible – focusing mainly on functional mechanical engineering for end-use parts, i.e. not for prototyping. The current limits of each design approach are discussed and the most exciting future opportunities for biomimetic AM applications are highlighted.
Artificial intelligence, and in particular machine learning, is a fast-emerging field. Research on artificial intelligence focuses mainly on image-, text- and voice-based applications, leading to ...breakthrough developments in self-driving cars, voice recognition algorithms and recommendation systems. In this article, we present the research of an alternative graph-based machine learning system that deals with three-dimensional space, which is more structured and combinatorial than images, text or voice. Specifically, we present a function-driven deep learning approach to generate conceptual design. We trained and used deep neural networks to evaluate existing designs encoded as graphs, extract significant building blocks as subgraphs and merge them into new compositions. Finally, we explored the application of generative adversarial networks to generate entirely new and unique designs.
Building layout generation has entered a new era in recent years, leveraging state-of-the-art deep generative methods to learn morphological properties of exiting urban structures and synthesize ...building alternatives responsive to local context. However, most existing research generally follows an image-to-image translation idea, while overlooking the impact of site/design attributes on building configuration, making their results less performative. Besides, most synthesized layouts are commonly displayed in 2D pixelized images, limiting further performance evaluation and informed decision-making. This study, therefore, proposes a novel GAN-based model, namely site-embedded generative adversarial networks (ESGAN) for automated building layout generation. Both qualitative and quantitative results in New York City indicate ESGAN is capable of synthesizing visually realistic and semantically reasonable layouts. This end-to-end generative system can not only encode a conditional vector to improve performance in different design scenarios but also display synthesized layouts at different levels of detail for human-system interaction.
•A novel deep learning-based generative model is proposed for automated building layout generation.•The model incorporates a conditional vector into the generation process to improve performance in different design scenarios.•The model can synthesize visually realistic and semantically reasonable building layouts.•The synthesized layouts are displayed in different levels of detail for human-system interaction.•The model is an end-to-end generative system without much expert knowledge and user intervention.