CumInCAD is a Cumulative Index about publications in Computer Aided Architectural Design
supported by the sibling associations ACADIA, CAADRIA, eCAADe, SIGraDi, ASCAAD and CAAD futures

PDF papers
References

Hits 121 to 140 of 144

_id c5ec
authors Smith Shaw, Doris
year 1988
title The Conceptual Approach to CAD Education
source Computing in Design Education [ACADIA Conference Proceedings] Ann Arbor (Michigan / USA) 28-30 October 1988, pp. 35-45
summary Recent research at the Corps of Engineers Construction Engineering Research Laboratory (CERL) investigated embedded computer-based instruction for AutoCAD. The results of this study, which are the focus of this paper, indicated that the only factor which correlated with success in completing the final test was previous experience with another CAD system. Those who knew another CAD system had higher scores and required less than half the time to complete the lessons. Presumably their conceptual knowledge about CAD transferred to the new software environment, even though the Corps' study showed that they were initially biased against learning the new system. Such biased attitudes have been observed when users are asked to learn a second similar software of any kind.

Architects who are deeply involved in computer-aided design have stated that one must learn to program the computer to build the conceptual framework for the creative process. We at CERL agree that an understanding of underlying graphics concepts is essential to the designer. Our research shows that giving students the freedom to explore an existing software program can result in the development of conceptual knowledge. Interviews also reveal that students can invent ways to meet individual objectives when "guided discovery" learning is encouraged.

series ACADIA
last changed 1999/01/01 18:23

_id acadia16_98
id acadia16_98
authors Smith, Shane Ida; Lasch, Chris
year 2016
title Machine Learning Integration for Adaptive Building Envelopes: An Experimental Framework for Intelligent Adaptive Control
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 98-105
summary This paper describes the development of an Intelligent Adaptive Control (IAC) framework that uses machine learning to integrate responsive passive conditioning at the envelope into a building’s comprehensive conventional environmental control system. Initial results show that by leveraging adaptive computational control to orchestrate the building’s mechanical and passive systems together, there exists a demonstrably greater potential to maximize energy efficiency than can be gained by focusing on either system individually, while the addition of more passive conditioning strategies significantly increase human comfort, health and wellness building-wide. Implicitly, this project suggests that, given the development and ever increasing adoption of building automation systems, a significant new site for computational design in architecture is expanding within the post-occupancy operation of a building, in contrast to architects’ traditional focus on the building’s initial design. Through the development of an experimental framework that includes physical material testing linked to computational simulation, this project begins to describe a set of tools and procedures by which architects might better conceptualize, visualize, and experiment with the design of adaptive building envelopes. This process allows designers to ultimately engage in the opportunities presented by active systems that govern the daily interactions between a building, its inhabitants, and their environment long after construction is completed. Adaptive material assemblies at the envelope are given special attention since it is here that a building’s performance and urban expression are most closely intertwined.
keywords model predictive control, reinforcement learning, energy performance, adaptive envelope, sensate systems
series ACADIA
type paper
email shaneida@u.arizona.edu
last changed 2016/10/24 11:12

_id acadia16_470
id acadia16_470
authors Sollazzo, Aldo; Baseta, Efilena; Chronis, Angelos
year 2016
title Symbiotic Associations
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 470-477
summary Soil contamination poses a series of important health issues, following years of neglect, constant industrialization, and unsustainable agriculture. It is estimated that 30% of the total cultivated soil in the world will convert to degraded land by 2020 (Rashid et al. 2016). Finding suitable treatment technologies to clean up contaminated water and soil is not trivial, and although technological solutions are sought, many are both resource-expensive and potentially equally unsustainable in long term. Bacteria and fungi have proved efficient in contributing to the bioavailability of nutrients and in aggregating formation in degraded soils (Rashid et al. 2016). Our research aims to explore the possible implementation of physical computing, computational analysis, and digital fabrication techniques in the design and optimization of an efficient soil remediation strategy using mycelium. The study presented here is a first step towards an overarching methodology for the development of an automated soil decontamination process, using an optimized bio-cell fungus seed that can be remotely populated using aerial transportation. The presented study focuses on the development of a methodology for capturing and modeling the growth of the mycelium fungus using photogrammetry-based 3D scanning and computational analysis techniques.
keywords computational design, photogrammetry, simulation, mycelium, 3d scanning, growth strategies
series ACADIA
type paper
email info@noumena.io
last changed 2016/10/24 11:12

_id ijac20053301
id ijac20053301
authors Srinivasan, Ravi S.; Malkawi, Ali M.
year 2005
title Real-time Simulations Using Learning Algorithms for Immersive Data Visualization in Buildings
source International Journal of Architectural Computing vol. 3 - no. 3, 265-280
summary Computational Fluid Dynamic (CFD) simulations are used to predict indoor thermal environments and assess their response to specific internal/external conditions. Although computing power has increased exponentially in the past decade, CFD simulations are still time-consuming and their prediction results cannot be used for real-time immersive visualization in buildings. A method that can bypass the timeconsuming simulations and generate "acceptable" results will allow such visualization to be constructed.This paper discusses a project that utilizes a supervised Artificial Neural Network (ANN) as a learning algorithm to predict post-processed CFD data to ensure rapid data visualization. To develop a generic learning model for a wide range of spatial configurations, this paper presents a pilot project that utilizes an unsupervised Reinforcement Learning (RL) algorithm. The ANN technique was integrated with an interactive, immersive Augmented Reality (AR) system to interact with and visualize CFD results in buildings. ANN was also evaluated against a linear regression model. Both models were tested and validated with datasets to determine their degree of accuracy. Initial tests, conducted to evaluate the user's experience of the system, indicated satisfactory results.
series journal
more http://www.ingentaconnect.com/search/expand?pub=infobike://mscp/ijac/2005/00000003/00000003/art00002
last changed 2007/03/04 06:08

_id 08c3
authors Steinfeld, Edward
year 1988
title Using CAD to Teach Architectural Design
source Computing in Design Education [ACADIA Conference Proceedings] Ann Arbor (Michigan / USA) 28-30 October 1988, pp. 75-86
summary A CAD program with solid modeling capabilities was used as the primary instructional media for a graduate level architectural design studio. The objective of the studio was to enhance design education by use of the CAD tool. A series of projects were designed to utilize CAD most effectively without sacrificing complexity and depth of educational content. The experiment identified several pedagogical advantages of CAD for design education and also some problems that have to be resolved to take maximum advantage of these tools.

series ACADIA
last changed 1999/01/01 18:25

_id 0bf9
authors Stevens, Garry and Radford, Antony D.
year 1988
title Educating for the Unknown: Present Computer Education for Future Design Practice
source Computing in Design Education [ACADIA Conference Proceedings] Ann Arbor (Michigan / USA) 28-30 October 1988, pp. 47-57
summary The penetration of computing and computer-aided design into current and future architectural practise is examined and the implications for education discussed. It is argued that computing is part of the working environment in which both education and practise exist and that the content of education needs to be reinterpreted in the context of this new environment.

series ACADIA
email anthony.radford@adelaide.edu.au
last changed 2003/05/16 17:23

_id acadia16_414
id acadia16_414
authors Tabbarah, Faysal
year 2016
title Almost Natural Shelter: Non-Linear Material Misbehavior
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 414-423
summary This paper critiques computational design and digital fabrication’s obsession with both precision and images of natural patterns by describing a messy attitude towards digital and material computation that integrates and blurs between linear and non-linear fabrication, resulting in material formations and spatial affects that are beyond pattern and image and are almost natural. The motivation behind the body of work presented in the paper is to question the production of space and aesthetics in a post-human frontier as we embark on a new geological era that is emerging out of the unprecedented influence of the human race on the planet’s ecological systems. The paper and the body of work posit that the blurring between the natural and the synthetic in the post-human frontier can materialize a conception of space that exhibits qualities that are both natural and synthetic. The paper is organized in three parts. It begins by describing the theoretical framework that drives the body of work. Next, it describes early digital and material casting explorations that began to blur between linear and non-linear fabrication to produce almost natural objects. Finally, it describes the process of designing and making Almost Natural Shelter, a spatial installation that emerges from the integration of messy computational design methodologies and chemically volatile non-linear fabrication. In specific, High Density Foam is persuaded to chemically self-compute in an attempt at uncovering a shelter that has almost natural spatial qualities, such as non-linear textural differentiation and sudden migration between different texture types.
keywords natural, texture, nonlinear fabrication, sensate systems
series ACADIA
type paper
email ftabbarah@aus.edu
last changed 2016/10/24 11:12

_id acadia16_280
id acadia16_280
authors Thomsen, Mette Ramsgaard; Tamke, Martin; Karmon, Ayelet; Underwood, Jenny; Gengnagel, Christoph; Stranghoner, Natalie; Uhlemann, Jorg
year 2016
title Knit as bespoke material practice for architecture
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 280-289
summary This paper presents an inquiry into how to inform material systems that allow for a high degree of variation and gradation of their material composition. Presenting knit as a particular system of material fabrication, we discuss how new practices that integrate material design into the architectural design chain present new opportunities and challenges for how we understand and create cycles of design, analysis, specification and fabrication. By tracing current interdisciplinary efforts to establish simulation methods for knitted textiles, our aim is to question how these efforts can be understood and extended in the context of knitted architectural textiles. The paper draws on a number of projects that prototype methods for using simulation and sensing as grounds for informing the design of complex, heterogeneous and performative materials. It asks how these methods can allow feedback in the design chain and be interfaced with highly craft-based methods of fabrication.
keywords cross disciplinary collaboration, knitting, light weight simulation, idesign integrated fe simulation, interfacing, sensing, bespoke material fabrication
series ACADIA
type paper
email mette.thomsen@kadk.dk
last changed 2016/10/24 11:12

_id ga0016
id ga0016
authors Thum, Robert and Derix, Christian
year 2000
title Artificial Neural Network Spaces
source International Conference on Generative Art
summary In the processes of continues functional differentiation the contemporary society increasingly displays the characteristics of complex, distributed systems. In this condition architecture can no longer rely on top-down reductionist methodologies in ignoring the constituting importance of contextual parameters. The present paper describes how Artificial Neural Networks can be employed to design with-in the underlying logic of our society -the logic of distributed systems. In computer simulations the paper explores the capability of ANN to Self-organization: Neural Networks Architectures absorb and adapt vast amount of urban data in order to adjust their organization to exterior changes through interior structural reconfiguration, thus producing adaptive spatial formations.    
series other
type normal paper
email Christian.Derix@aedas.com
more http://www.generativeart.com/
last changed 2012/09/25 05:42

_id acadia16_424
id acadia16_424
authors Twose, Simon; du Chatenier, Rosa
year 2016
title Experimental Material Research - Digital Chocolate
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 424-431
summary This research investigates the aesthetics of a shared agency between humans, computation and physical material. ‘Chocolate’ is manipulated in physical and virtual space simultaneously to extract aesthetic conditions that are a sum of human and non-human relations. This is an attempt to further the knowledge of designing, giving physical and digital materials force in determining their own aesthetics. The research springs from work in speculative aesthetics, particularly N. Katherine Hayles’s OOI (object-oriented inquiry) and Graham Harman’s OOO (object-oriented ontology) and explores how these ideas impact contemporary computational architectural design. To study this, a simple material has been chosen, chocolate, and used as a vehicle to investigate the dynamics of physical and digital materials and their shared/differing ‘resistances to human manipulation’ (Pickering 1995). Digital chocolate is ‘melted’ through virtual heat, and the results printed and cast in real chocolate, to be further manipulated in real space. The resistances and feedback of physical and digital chocolate to human ‘prodding’ (Hayles 2014) are analyzed in terms of a material’s qualities and tendencies in digital space versus those in physical space. Observations from this process are used to speculate on an aesthetics where humans, computation and physical material are mutually agential. This research is a pilot for a larger study taking on more complex conditions, such as building and cities, with a view to broadening how aesthetics is understood in architectural design. The contribution of this research to the field of architectural computation is thus in areas of aesthetic speculation and human/non-human architectural authorship.
keywords object-oriented inquiry, speculative aesthetics, mutual agency, big data
series ACADIA
type paper
email rosaduchatenier@gmail.com
last changed 2016/10/24 11:12

_id acadia16_184
id acadia16_184
authors Vasey; Lauren; Long Nguyen; Tovi Grossman; Heather Kerrick; Danil Nagy; Evan Atherton; David Thomasson; Nick Cote; David Benjamin; George Fitzmaurice; Achim Menges
year 2016
title Collaborative Construction: Human and Robotic Collaboration Enabling the Fabrication and Assembly of a Filament-Wound Structure
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 184-195
summary In this paper, we describe an interdisciplinary project and live-exhibit that investigated whether untrained humans and robots could work together collaboratively towards the common goal of building a large-scale structure composed out of robotically fabricated modules using a filament winding process. We describe the fabrication system and exhibition setup, including a custom end effector and tension control mechanism, as well as a collaborative fabrication process in which instructions delivered via wearable devices enable the trade-off of production and assembly tasks between human and robot. We describe the necessary robotic developments that facilitated a live fabrication process, including a generic robot inverse kinematic solver engine for non-spherical wrist robots, and wireless network communication connecting hardware and software. In addition, we discuss computational strategies for the fiber syntax generation and robotic motion planning which mitigated constraints such as reachability, axis limitations, and collisions, and ensured predictable and therefore safe motion in a live exhibition setting. We discuss the larger implications of this project as a case study for handling deviations due to non-standardized materials or human error, as well as a means to reconsider the fundamental separation of human and robotic tasks in a production workflow. Most significantly, the project exemplifies a hybrid domain of human and robot collaboration in which coordination and communication between robots, people, and devices can enhance the integration of robotic processes and computational control into the characteristic processes of construction.
keywords machin vision, cyber-physical systems, internet of things, robotic fabrication, human robot collaboration, sensate systems
series ACADIA
type paper
email lauren.vasey@icd.uni-stuttgart.de
last changed 2016/11/17 09:20

_id acadia16_342
id acadia16_342
authors Velikov, Kathy
year 2016
title Posthuman Engagements
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 342-343
series ACADIA
type introduction
email kvelikov@umich.edu
last changed 2016/10/24 11:12

_id acadia16_412
id acadia16_412
authors Velikov, Kathy
year 2016
title Material Frontiers
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 412-413
series ACADIA
type introduction
email kvelikov@umich.edu
last changed 2016/10/24 11:12

_id acadia16_0
id acadia16_0
authors Velikov, Kathy; Ahlquist, Sean; del Campo, Matias; Thün, Geoffrey (eds.)
year 2016
title ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, 514p.
summary The ACADIA 2016 Conference will feature research and design work from practice and academia that is positioned at the intersection of procedural design, digital environments, and autonomous machines. POSTHUMAN FRONTIERS: DATA, DESIGNERS, AND COGNITIVE MACHINES will bring together architects, designers, material scientists, engineers, programmers, and artists to explore the current trend in computational design to develop and apply quasi-cognitive machines, and to advance the integration of software, information, fabrication, and sensing in the generation of mechanisms for interfacing with the physical realm
series ACADIA
email kvelikov@umich.edu
last changed 2016/10/24 11:12

_id acadia16_8
id acadia16_8
authors Velikov, Kathy; Ahlquist, Sean; del Campo, Matias; Thün, Geoffrey
year 2016
title Introduction: Posthuman Frontiers
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 8-9
series ACADIA
type introduction
email kvelikov@umich.edu
last changed 2016/10/24 11:12

_id cf2011_p047
id cf2011_p047
authors Vermeersch, Peter-Willem; Nijs Greg, Heylighen Ann
year 2011
title Mediating Objects in Architectural Design: a Non-Visual Exploration
source Computer Aided Architectural Design Futures 2011 [Proceedings of the 14th International Conference on Computer Aided Architectural Design Futures / ISBN 9782874561429] Liege (Belgium) 4-8 July 2011, pp. 721-734.
summary Through their daily interactions with the built environment, people with disabilities become able to appreciate spatial qualities or detect obstacles that architects may not be attuned to. This observation motivated us to explore scenarios for involving people with sensory disabilities as experts in the design process. An architecture office participating in a real-world design competition is teamed up with two blind persons. The design process is studied in real time through a team ethnography. The analysis in this paper focuses on the mediating aspects of objects in the actions, perception and cognition in one collaborative design meeting in particular. In general, disability situations can teach us something about fixed ways of doing by making perceivable, or questioning practices that seem self-evident. In this particular situation, the blind person’s involvement in a design meeting that relies heavily on representational artifacts, makes perceivable or questions everyday practices in architects’ design process that are taken for granted. Examples include knowing and indicating to others where design elements are on the site, knowing what design element is being talked about, holding the element and its environment ‘in place’, or spatially exploring the design’s spatial configuration. As such, our study shows that exploring inclusivity ‘upstream’, i.e. in the design process, may contribute not only to inclusive design, but also to a more articulate understanding of the working of mediating objects and their use in architects’ design processes tout court.
keywords architectural design, blindness, disability, participant observation, users
series CAAD Futures
email peterwillem.vermeersch@asro.kuleuven.be
last changed 2012/02/11 18:21

_id bb8d
id bb8d
authors von Buelow, Peter
year 2007
title Genetically Engineered Architecture: design exploration with evolutionary computation
source VDM Verlag Dr. Mueller, Germany, Dec. 2007. ISBN 978-3-8364-4721-8
summary This book explores design tools based on evolutionary computation (EC), oriented primarily toward conceptual design of architectural and civil engineering structures. EC tools are well suited for exploration in a way which promotes creative design. The multiplicity of solutions generated by EC techniques is less likely to cause design fixation, and so promote a more thorough exploration of possible solutions. The use of such tools also allows the designer greater latitude in exploring design criteria, such as aesthetics, by utilizing an interactive human-computer interface. This book begins with a survey of techniques that have been used in early phases of architectural design, and establishes a set of successful attributes, which are then discussed in the context of EC techniques. Finally, a specific implementation developed by the author is described. Several examples are given in the area of architectural engineering, and comparisons are made with results obtained with more conventional optimization tools. This book is especially useful for designers interested in new methods for generating and exploring structural form, and is accessible to non-programmers in either field.

Dr.-Ing. Peter von Buelow has worked as both architect and engineer, and is currently a professor at the University of Michigan in Ann Arbor, USA, where he teaches structures in the School of Architecture, and conducts research in structural form exploration based on evolutionary computation. For more information visit: www.umich.edu/~pvbuelow.

keywords evolutionary computation, genetic algorithm, design, optimization, structures
series book
type normal paper
email pvbuelow@umich.edu
more http://elib.uni-stuttgart.de/opus/volltexte/2007/3160/pdf/PvB_diss.pdf
last changed 2008/05/12 17:05

_id acadia16_290
id acadia16_290
authors Wang, Adam; Ahlquist, Sean
year 2016
title Pneumatic Textile System
source ACADIA // 2016: POSTHUMAN FRONTIERS: Data, Designers, and Cognitive Machines [Proceedings of the 36th Annual Conference of the Association for Computer Aided Design in Architecture (ACADIA) ISBN 978-0-692-77095-5] Ann Arbor 27-29 October, 2016, pp. 290-297
summary This paper attempts to demonstrate a seamless transformable material system through an interdependent designed assembly of two materials with different material properties (anisotropic knit textile and isotropic silicone) but similar behaviors (stretch). The transformable system is achieved by balancing the volumetric expansion through a silicone tube, under inflation, with the controlled resistance to stretch by a custom knit fabric. The use of a CNC knitting machine allows not only an opportunity to program the stretch behavior of a knit fabric, by controlling the amount of yarn material to be deposited, but also an ability to knit multiple layers of fabric simultaneously, in order to create a space capable of accommodating an external element seamlessly. The paper will showcase a series of experiments ranging from the initial search for compatible material combinations to the varied structures of the tube sleeve and its relationship with surrounding region. The final prototype attempts to utilize the various behavioral properties of the material system learned from the experiments to create a transformable three-dimensional structure.
keywords fabric, knit, cnc knitting, pneumatic textile, embedded responsiveness
series ACADIA
type paper
email wcawang@umich.edu
last changed 2016/10/24 11:12

_id caadria2018_290
id caadria2018_290
authors Wang, Zhenyu, Shi, Jia, Yu, Chuanfei and Gao, Guoyuan
year 2018
title Automatic Design of Main Pedestrian Entrance of Building Site Based on Machine Learning - A Case Study of Museums in China's Urban Environment
source T. Fukuda, W. Huang, P. Janssen, K. Crolla, S. Alhadidi (eds.), Learning, Adapting and Prototyping - Proceedings of the 23rd CAADRIA Conference - Volume 2, Tsinghua University, Beijing, China, 17-19 May 2018, pp. 227-235
summary The main pedestrian entrance of the building site has a direct influence on the use of the buildings, so the selection of the main pedestrian entrance is very important in the process of architectural design. The correct selection of the main pedestrian entrance of building site depends on the experience of designers and environment data collected by designers, the process is time consuming and inefficient, especially when the building site located in complex urban environment. In order to improve the efficiency of design process, we used online map to collect museums information in China as training samples, and constructing artificial neural networks to predict the direction of the main pedestrian entrance. After the training, we get the prediction model with 79% prediction accuracy. Although the accuracy still need to be improved, it creates a new approach to analysis the main pedestrian entrance of the site and worth further researching.
keywords Artificial Neural Network (ANN); Main Pedestrian Entrance of Building Site; Automatic Design
series CAADRIA
email 514656668@qq.com
last changed 2018/05/17 07:08

_id ijac201412204
id ijac201412204
authors Wilkinson, Samuel; Sean Hanna
year 2014
title Approximating Computational Fluid Dynamics for Generative Tall Building Design
source International Journal of Architectural Computing vol. 12 - no. 2, 155-178
summary Background literature review, methodology, results, and analysis are presented for a novel approach to approximating wind pressure on tall buildings for the application of generative design exploration and optimisation.The predictions are approximations of time-averaged computational fluid dynamics (CFD) data with the aim of maintaining simulation accuracy but with improved speed.This is achieved through the use of a back-propagation artificial neural network (ANN) with vertex-based shape features as input and pressure as output.The training set consists of 600 procedurally generated tall building models, and the test set of 10 real building models; for all models in both sets, a feature vector is calculated for every vertex. Over the test set, mean absolute errors against the basis CFD are 1.99–4.44% (_:2.10–5.09%) with an on-line process time of 14.72–809.98s (0.028s/sample). Studies are also included on feature sensitivity, training set size, and comparison of CFD against prediction times. Results indicate that prediction time is only dependent on the number of test model vertices, and is therefore invariant to basis CFD time.
series journal
last changed 2019/05/24 07:55

For more results click below:

show page 0show page 1show page 2show page 3show page 4show page 5this is page 6show page 7HOMELOGIN (you are user _anon_343604 from group guest) CUMINCAD Papers Powered by SciX Open Publishing Services 1.002