Design Space Research Articles

Exploring unseen 3D scenarios of physics variables using machine learning-based synthetic data: An application to wave energy converters

This work uses machine learning to produce synthetic data of wave energy converters from time-expensive 3D simulations based on computational fluid dynamics models. The simulations to analyse the response of these systems to incoming waves are lengthy and computationally expensive to obtain. Here, we explore the use of a beta-VAE and a Principal Components-based adversarial autoencoder for generating new synthetic data. The compression plus the generation of synthetic data introduces an exceptionally fast surrogate model of the original simulation and delivers more samples of either dynamic viscosity and velocity fields, enlarging the design space. The new generated synthetic samples can have a speed up from 5 to 6 orders of magnitude. The new design space can be used to improve the prediction of dynamic viscosity given the velocity fields. The generative model has the potential to capture the transition and the new physical phenomena under extreme initial conditions.

Experimental characterisation of rotor noise in tandem configuration

This paper presents a comprehensive investigation into the noise radiation characteristics of two rotors in a tandem configuration. Through an extensive experimental campaign, the study explores the effect of the separation distance between two rotors on the noise directivity patterns, spectral characteristics and temporal features of the radiated noise under both hovering and edge-wise flight conditions. The directivity patterns of the overall sound pressure level obtained from the polar and side arrays demonstrate a dependence on the separation distance and advance ratio. Spectral characteristics elucidate the influence of separation distance on tonal and broadband energy content. While the high-frequency broadband noise increases for both hovering and edge-wise flight similarly, the low-frequency broadband energy strongly depends on the advance ratio and the rotor separation distance. The results reveal a distinct envelope in the design space, where the radiated broadband noise for tandem configurations can be minimized, offering potential insights for noise reduction strategies. Additionally, the wavelet analysis for tandem configurations reveals that the emitted noise at the first and second blade passing frequency is strongly intermittent with amplitude modulation, which shall be considered for annoyance perception.

DRLAR: A deep reinforcement learning-based adaptive routing framework for network-on-chips

Adaptive routing plays a pivotal role in the overall performance of Network-on-Chips (NoCs). However, with many-core architectures supporting complex and constantly changing traffic patterns for emerging applications, this aspect presents significant challenges. Our meticulous examination and analysis of existing heuristic adaptive routing algorithms revealed three key limiting factors: single network status metric, lack of system feedback awareness, and lack of ability to customize. Reinforcement Learning (RL) methods have demonstrated promising opportunities for exploring adaptive routing design. Deep reinforcement learning (DRL) techniques, in particular, enable efficient exploration in adaptive routing design spaces where heuristic strategies may be inadequate. This paper proposes a novel deep reinforcement learning framework for adaptive routing called DRLAR that is suitable for diversified traffic patterns and resolves multi-objective optimization simultaneously. DRLAR formulates routing as an agent, which makes routing decisions based on autonomous learning using multiple network state features and system-level feedback information. We conduct extensive experiments against state-of-the-art routing algorithms to evaluate our design. The results show that DRLAR reduces packet latency by an average of 33.3%, achieving a reduction of 41.6% and 10.5% on average under heavy synthetic traffic and the PARSEC 2.1 benchmark, respectively. We also perform a cost analysis to validate the potential implementation of DRLAR on NoCs with low computational, storage, and power.

Design Space Exploration of Low-Bit Quantized Neural Networks for Visual Place Recognition

Design Space Exploration of Low-Bit Quantized Neural Networks for Visual Place Recognition

A machine-learning based framework for design and characterization of honeycombs with partial self-similar hierarchical architectures

Honeycomb structures with hierarchical topology are widely used in engineering fields for their excellent mechanical properties. Multiphase topology is proved to possess a larger design space than the traditional hierarchical topology. In this study, a partial self-similar hierarchical honeycomb (PSSHH), comprising of zero-order cell and first-order cell, was designed. Numerical simulations, validating against the experimental results, were carried out to investigate the in-plane crushing performance of PSSHHs with different configurations. An artificial neural network model was trained using the results of simulations. The ability and effectiveness of artificial neural networks in predicting and characterizing the crushing responses of PSSHH were evaluated through comparison with simulations. The model was primarily employed for optimizing the design of PSSHH to enhance the crushing performance. The results indicate that the optimized PSSHH outperforms traditional hexagonal honeycombs and fully self-similar hierarchical honeycombs in terms of initial peak force, mean crushing force, and energy absorption. Furthermore, the effect of number and organized patterns of the first-order cells along the loading direction on the mechanical properties of PSSHH is examined. The findings reveal that the formation and development of localized collapse bands are significantly affected by the number and distribution of hierarchical first-order cells. The deformation, which involves a sequential crushing process with the initial collapse of specific zero-order cells followed by the subsequent collapse of first-order cells, contributes to enhancing the crashworthiness performance of PSSHH.

NSGAN: a non-dominant sorting optimisation-based generative adversarial design framework for alloy discovery

The design and discovery of new materials is fundamental to advancing scientific and technological innovation. The recent emergence of the materials genome concept holds great promise in revolutionising materials science by enabling the systematic utilisation of data for efficient prediction and optimisation of ‘superior’ materials. However, the materials genome approach can be stymied by the vast complexity of design spaces, which often demand substantial computational resources and sophisticated data processing capabilities. To address these challenges, this work introduces a generative design framework called the non-dominant sorting optimisation-based generative adversarial networks (NSGAN). Capitalising on the synergies of genetic algorithms (GA) and generative adversarial networks (GANs), NSGAN provides a robust and efficient approach for tackling high-dimensional multi-objective optimisation design problems. To validate the efficacy of the proposed framework, we applied the model to a comprehensive dataset of aluminium alloys. Additionally, an online tool was created as a supplementary resource, offering a brief introduction to this innovative method for the wider scientific community. This study explores the potential of a predictive and data-driven approach in material design, indicating a promising pathway for widespread applications in the field of materials science.

Theoretical modeling of a bottom-raised oscillating surge wave energy converter structural loadings and power performances

This study presents theoretical formulations to evaluate the fundamental parameters and performance characteristics of a bottom-raised oscillating surge wave energy converter (OSWEC) device. Employing a flat plate assumption and potential flow formulation in elliptical coordinates, closed-form equations for the added mass, radiation damping, and excitation forces/torques in the relevant pitch-pitch and surge-pitch directions of motion are developed and used to calculate the system's response amplitude operator and the forces and moments acting on the foundation. The model is benchmarked against numerical simulations using WAMIT and WEC-Sim, showcasing excellent agreement. The sensitivity of plate thickness on the analytical hydrodynamic solutions is investigated over several thickness-to-width ratios ranging from 1:80 to 1:10. The results show that as the thickness of the benchmark OSWEC increases, the deviation of the analytical hydrodynamic coefficients from the numerical solutions grows from 3 % to 25 %. Differences in the excitation forces and torques, however, are contained within 12 %. While the flat plate assumption is a limitation of the proposed analytical model, the error is within a reasonable margin for use in the design space exploration phase before a higher-fidelity (and thus more computationally expensive) model is employed. A parametric study demonstrates the ability of the analytical model to quickly sweep over a domain of OSWEC dimensions, illustrating the analytical model's utility in the early phases of design.

Research on the Application of Visual Form in Space Design in Digital Media Art

Research on the Application of Visual Form in Space Design in Digital Media Art

Clinical Affordances: Towards an Ecological Account of Health Affordance and Design of Clinical Learning Spaces

Clinical Affordances: Towards an Ecological Account of Health Affordance and Design of Clinical Learning Spaces

Special ship design and ocean space multi-use synergies

Special ship design and ocean space multi-use synergies

FrameGraph: A Scalable Performance Evaluation Method for Frame Structure Designs Using Graph Neural Network

Abstract Multilayer Perceptron (MLP) and Convolutional Neural Network (CNN) encounter a critical scalability issue when applied to the performance evaluation task for frame structure designs. Specifically, a model of MLP or CNN is limited to structures of a particular topology type and fails immediately when applied to other topology types. In order to tackle this challenge, we propose a scalable performance evaluation method (called FrameGraph) for frame structure designs using Graph Neural Network (GNN), offering applicability to a wide range of topology types simultaneously. FrameGraph consists of two main parts: (1) Components and their connections in a frame structure are denoted as edges and vertices in a graph, respectively. Subsequently, A graph dataset for frame structure designs with different topologies is constructed. (2) A well-defined GNN design space is established with a general GNN layer, and a controlled random search approach is employed to derive the optimal GNN model for this performance evaluation task. In numerical experiments of car door frames and car body frames, FrameGraph achieved the highest prediction precisions (96.28% and 97.87%) across all structural topologies compared to a series of classical GNN algorithms. Furthermore, the comparison with MLP and FEM highlighted FrameGraph's significant efficiency advantage. This verifies the feasibility and optimality of FrameGraph for the performance evaluation task of frame structures with different topologies.

Colour Perception in Camper Van Interior Design Using Virtual Reality Technology

Virtual reality (VR) technology has been developing progressively and rapidly in the past decades. VR technology is used to create 3D models of rooms and spaces, which can be used to create more accurate and detailed designs. Colours used in interior design have the ability to create specific moods, create harmony, and change the perception of how large or small a space may look or feel. The purpose of this study is to explore the application of virtual reality technology in the interior design of a camper van considering the role of colour perception. We aim to determine the effect of colour on the interior design creating professional-looking 3D camper van layout. The results show that any non-professional in 3D design can design a functional and comfortable living space in the camper van using Vanspace 3D software. It was found that light colours and shades create the feeling of space, light, comfort and cosiness. When choosing colours for the interior design of a small limited space, as is the case with a camper van, it is advisable to combine warm earth colours, white and beige colours.

Train Station Pedestrian Monitoring Pilot Study Using an Artificial Intelligence Approach

Pedestrian monitoring in crowded areas like train stations has an important impact in the overall operation and management of those public spaces. An organized distribution of the different elements located inside a station will contribute not only to the safety of all passengers but will also allow for a more efficient process of the regular activities including entering/leaving the station, boarding/alighting from trains, and waiting. This improved distribution only comes by obtaining sufficiently accurate information on passengers’ positions, and their derivatives like speeds, densities, traffic flow. The work described here addresses this need by using an artificial intelligence approach based on computational vision and convolutional neural networks. From the available videos taken regularly at subways stations, two methods are tested. One is based on tracking each person’s bounding box from which filtered 3D kinematics are derived, including position, velocity and density. Another infers the pose and activity that a person has by analyzing its main body key points. Measurements of these quantities would enable a sensible and efficient design of inner spaces in places like railway and subway stations.

New Conventions: Intentional Implementation of Set-Based Design Leveraging Limited Point-Based Approaches

This article communicates a practical application of Set-Based Design (SBD) during its adoption within the U.S. Navy’s latest surface ship design processes. It highlights how the team harnessed SBD to explore the vast design space and make critical decisions. Intriguingly, the article explores the design team’s use of point-based designs(PBD), a counterintuitive approach within the context of SBD. It delves into the rationale behind their implementation and the diverse purposes they serve in bridging knowledge gaps such as prototyping, integration, and evaluation of set interfaces. Furthermore, the article presents a preliminary ontology outlining various use cases for these PBDs. A key focus of this article is understanding how these PBDs coexist within the larger set-based construct while recognizing their inherent differences. The article examines the utility and boundaries of this approach, shedding light on the pragmatic interplay.

Complete Solutions for the Approximate Synthesis of Spherical Four-Bar Function Generators

Abstract Kinematic synthesis to meet an approximate motion specification naturally forms a constrained optimization problem. Instead of using local methods, we conduct global design searches by direct computation of all critical points. The idea is not new, but performed at scale is only possible through modern polynomial homotopy continuation solvers. Such a complete computation finds all minima, including the global, which allows for a full exploration of the design space, whereas local solvers can only find the minimum nearest to an initial guess. We form equality-constrained objective functions that pertain to the synthesis of spherical four-bar linkages for approximate function generation. We consider the general case where all mechanism dimensions may vary and a more specific case that enables the placement of ground pivots. The former optimization problem is shown to permit 268 sets of critical points, and the latter permits 61 sets. Critical points are classified as saddles or minima through a post-process eigenanalysis of the projected Hessian. The discretization points are contained within the coefficients of the stationarity polynomials, so the algebraic structure of the synthesis equations remains invariant to the number of points. The results of our computational work were used to design a mechanism that coordinates the folding wings. We also use this method to parameterize mechanism dimensions for a family of hyperbolic curves.

Seeing a Sea of Ships - Exploring the Ship Design Space in the Digital Domain

We tackle in this work aspects of the ship design space in the digital domain, with an overview of the current status and opportunities to shift from fixed arrangements towards open technologies, proposing a mix of open and proprietary databases. The discussion is focused on the visual domain and digital thread in ship design. Literature examples from the Brazilian case and the visualization of the ocean space are presented (Numerical Offshore Tank - TPN), followed by the Japanese services to design and optimize hull for specific missions (NMRI), and lastly the current open ship design library developed in Norway (Vessel.js, NTNU). We present the argument that seeing a sea of ships, that is, visualizing the behavior of many options is already a reality, accessible from a portable device, without the need a large cluster as in the past, exemplified by web-based cases. Our conclusion is that computer graphics approaches to Ship Design should be considered open and exchangeable. Naval architects should focus on what they do best: creating, analyzing, refining, storing and populating the database of the know-how from the institution (e.g. university, research institute or company).

Designing semiconductor materials and devices in the post-Moore era by tackling computational challenges with data-driven strategies.

In the post-Moore's law era, the progress of electronics relies on discovering superior semiconductor materials and optimizing device fabrication. Computational methods, augmented by emerging data-driven strategies, offer a promising alternative to the traditional trial-and-error approach. In this Perspective, we highlight data-driven computational frameworks for enhancing semiconductor discovery and device development by elaborating on their advances in exploring the materials design space, predicting semiconductor properties and optimizing device fabrication, with a concluding discussion on the challenges and opportunities in these areas.

Leveraging the concept of information entropy to improve a multi-fidelity design framework for early-stage design exploration of complex vessels

Early-stage design exploration is crucial since most of the major design decision are locked-in and only small design modifications are possible at later stages. To assess the performance of the various design candidates while performing design exploration, there are available methods and tools of various fidelities. These methods can be combined to form a multi-fidelity (MF) framework that guarantees accuracy through the high-fidelity model and achieves faster computational speeds through low-fidelity models. The present study proposes the adoption of information-theoretic entropy to improve a MF design framework based on Gaussian Processes (GPs). Entropy quantifies the uncertainty associated with the prediction of the design space. We propose using this uncertainty metric both as a criterion to determine whether further designs should be sampled to construct a reliable approximation of the design space and as a criterion to establish in which optimization step the optimization of the covariance matrix for the MF-GPs should be performed. The approach was tested to benchmark analytical functions and to a ship design problem of an AXEfrigate. The approach holds potential in practical applications, as it aids in the determination of whether additional resources should be allocated for high-fidelity analysis to support early-stage exploration.

Design strategies for aging adaptation design of public spaces under the community elderly care model

As the aging population intensifies, the community elderly care model has gradually become an important model for elderly care. The research takes the community elderly care model as the starting point, the public spaces is used as the key research objective. Firstly, the relevant theoretical research and conceptual definition of aging adaptation design are elaborated. Secondly, the excellent practical cases of the construction and renovation of elderly care communities both domestically and internationally are analyzed. The problems in the aging adaptation design of public spaces in China are analyzed. Finally, based on the psychological, physiological, and social needs of the elderly, a design strategy for public spaces that are suitable for aging under the community elderly care model is proposed, namely, “layout consultation spaces at different levels, form leisure spaces that are conducive to communication, and construct kitchen spaces with composite infiltration”. Through reasonable design and planning, the aging adaptation design strategy of public spaces under the community elderly care model can create a livable and enjoyable leisure environment. Elderly people feel the care and support from the community, improve their quality of life, and enhance social well-bei

Comprehensive review of the directed design of optomechanical crystal cavities using intelligent algorithms

Optomechanical crystal cavities are devices based on optomechanical interactions to manipulate photons and phonons on periodic subwavelength structures, enabling precise measurement of the force and displacement. The performance of the target structures varies when applied to different applications. Optomechanical crystal cavities now rely on an empirical forward design, which is inefficient. Therefore, a desired shift is toward directed design with a “problem-oriented” strategy. The directed optimization problem’s nonconvex nature and extensive parameter space necessitate substantial computational resources, driving the need for intelligent algorithms in a sub-wavelength structure design. Intelligent algorithms can surpass the constraints of traditional methods and discover novel structures that are effective in different materials, topologies, modes, and wavelengths. This paper provides an extensive overview of intelligent algorithms for guiding the directed design of optomechanical crystal cavities. It presents a systematic classification of 15 algorithmics, including, but not limited to, topology algorithms, particle swarm optimization algorithms, convolutional neural networks, and generative adversarial networks. The article provides a comprehensive review and thorough analysis of the principle and current application state, as well as the advantages and disadvantages of each intelligent algorithm. By using these intelligent algorithms, researchers can enhance the efficiency and accuracy of optimizing optomechanical crystal cavities in a broader design space.