Volume 3B: 48th Design Automation Conference (DAC)最新文献

{"title":"Generative Adversarial Design Analysis of Non-Convexity in Topology Optimization","authors":"Nathan Hertlein, A. Gillman, P. Buskohl","doi":"10.1115/detc2022-89997","DOIUrl":"https://doi.org/10.1115/detc2022-89997","url":null,"abstract":"\u0000 Material penalization and filtering schemes are key strategies applied to topology optimization (TO) to promote more discrete and manufacturable designs. However, these modifications introduce fluctuations in the design landscape that amplify non-convexity and influence the local minima identified by TO. Harnessing the machine learning approach of generative adversarial networks (GAN), we investigate the role of penalization and filtering by comparing the designs between TO and GAN-based TO surrogates. A total of 17 GANs were constructed to predict 2D minimum compliance topologies across a set of penalization factors and filters, each interpolating a design space of 270,000 boundary condition and loading scenarios. The prevalence of GAN-predicted topologies with better compliance than TO-calculated topologies was estimated via a random sampling of the design space. GAN ‘over-performance’ occurs across material penalization and filtering conditions, where the frequency tends to increase as penalization increases. Analysis of this test set is leveraged to highlight trends regarding the conditions under which this ‘over-performance’ occurs, and the geometric characteristics these designs exhibit. Collectively, this study presents an alternative method to characterize the effects of penalization and filtering on design outcomes and motivates the use of data-driven surrogates to augment traditional approaches.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"34 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121206550","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Ship Deck Object Placement Optimization Using a Many-Objective Bilevel Approach","authors":"Noah J. Bagazinski, Faez Ahmed","doi":"10.1115/detc2022-89797","DOIUrl":"https://doi.org/10.1115/detc2022-89797","url":null,"abstract":"\u0000 The placement of objects on a ship is critical to many facets of the performance of a ship. Most notably, the mass distribution properties of objects in a ship affect the ship’s stability, trim, and structural loading. Information gathered from object placement optimization can allow naval architects to further optimize the design of the whole ship by potentially reducing the structural weight of the vessel, and adjusting the shape of the hull or the general arrangements based on available space in the ship. This paper presents a novel, many-objective bin packing problem for object placement across multiple decks on a ship. This problem is also highly constrained to avoid object intersection and protrusion. The problem was optimized with the NSGA-II algorithm, utilizing a heuristic population initialization and by separating the objectives into a bilevel optimization scheme. The bilevel scheme decouples certain objectives and design variables from the rest of the problem and sequences the evaluation for the objectives in a two-stage process. The hypervolume of the final population measured the performance of the optimization test. The results indicate that sequencing the objectives with a bilevel scheme produces an 80.3% larger hypervolume than an all-in-one optimization for the same problem. The findings from this study provide a systematic way by combining concepts from many-objective optimization, bin packing heuristics, and bilevel optimization to sequence the optimization of many-objective, object placement problems.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"48 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123010734","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Decision-Centric Perspective on Evolving Cyber-Physical-Social Systems: Effectiveness, Group Value, and Opportunities","authors":"Anton van Beek","doi":"10.1115/detc2022-90161","DOIUrl":"https://doi.org/10.1115/detc2022-90161","url":null,"abstract":"\u0000 In this paper, we view evolving cyber-physical-social systems (CPSSs) from a group decision-making perspective, introduce the group value concept as a potential approach to improve their effectiveness, and conclude by identifying a set of research directions for further scientific inquiry. An evolving CPSS is a system in which the digital and physical spaces adapt to changing interests in the social space. In this paper, we introduce the group value concept as an approach to balancing the interests of individuals in the social space and deciding how a CPSS should evolve. The advantages afforded by the group value concept are twofold: (i) it enables CPSSs to evolve along with the interests of the social space, and (ii) it provides transparency in the decision-making process that will improve public support. The group value is a stochastic function that is constructed by modeling the distributions of individual value functions and shares a similarity with utility-theory and normative models for group decision-making. Through analysis of the introduced framework, we show: (i) how the group value concept can be used to bring about evolving CPSSs, (ii) introduce the difference between utility theory and normative models for group decision-making, (iii) define the conditions under which the introduced evolving CPSSs framework is valid, and (iv) delineate a set of four research areas for further scientific inquiry. The motivation for delineating a set of additional research challenges comes from the observation that group decisions violate the conditions of logical decision-making that can only be satisfied for an individual’s decisions. Consequently, establishing an agent that controls the evolution of a CPSS needs to consider the consequences of violating these conditions on the effectiveness of the decision. Through continued research in the identified decision-centric research areas, evolving CPPSs can be established to address many societal challenges and will be more effective as they enjoy broader public support.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129913597","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Uncertainty Quantification of Physics-Based Label-Free Deep Learning and Probabilistic Prediction of Extreme Events","authors":"Huiru Li, Jianhua Yin, Xiaoping Du","doi":"10.1115/detc2022-88277","DOIUrl":"https://doi.org/10.1115/detc2022-88277","url":null,"abstract":"\u0000 Surrogate models from machine learning regression have been increasingly used in engineering analysis and design. Since surrogate models are usually built using data from solving expensive physical models, label-free machine learning methodologies have been developed to reduce the computational cost. Understanding and quantifying the model (epistemic) uncertainty of surrogate models is critical for their applications with quantified confidence. It is, however, much more computationally expensive, or even impossible to quantifying the model uncertainty for label-free machine learning. In this work we propose an uncertainty quantification method for the epistemic uncertainty of physics-based label-free regression. The method is used after a surrogate model has already been built by deep neural network based on the data of only input variables without labels (data of responses) and a system of physical equations. A surrogate model of the neural network regression model error is built with Gaussian Process regression using the existing training points and the derivatives of the system of physical equations at the training points. The error model is then used to compensate the error of the neural network surrogate model, therefore producing more accurate predictions. With higher accuracy, the proposed method is applied to probabilistic prediction of extreme events where both (aleatory) data uncertainty and model uncertainty coexist, and higher accuracy is required. Its application to time-dependent reliability prediction of a four-bar linkage mechanism demonstrates the high accuracy of the proposed method.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"11 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129139067","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimization of Self-Heated Vacuum Membrane Distillation Using Response Surface Methodology","authors":"Shaneza Fatma Rahmadhanty, S. T. M., W. Hung, P. Lin","doi":"10.1115/detc2022-89491","DOIUrl":"https://doi.org/10.1115/detc2022-89491","url":null,"abstract":"\u0000 Currently, technology is increasingly advanced and has penetrated into several fields such as Membrane Distillation (MD) in water treatment, especially to purify saline water, in the face of the increasing scarcity of clean water needed for consumption, both in the domestic, agriculture, and industrial sectors. One of the configurations in MD systems, self-heated Vacuum Membrane Distillation (VMD) system, which employs a Graphene-PVDF membrane heated by power supply, in this case using DC power with low voltage, to eliminate feed pre-feed heating and temperature polarization, is interesting to learn. In addition, to enhance the performances, such as optimizing permeate flux (JW), Temperature Polarization Factor (TPF), Specific Heating Energy (QSH), and Gain Output Ratio (GOR), several different designs are made and compared. In this study, four parameters are selected: DC power supply’s voltage, feed flow rate, the length, and the width of the cell-body’s slot. Ansys FLUENT software is utilized to simulate the system, followed by Minitab software to analyze the results using Response Surface Method (RSM) which aims to achieve the optimal design parameters. The simulation data was validated by experimental data and determination of the optimum conditions of the self-heated VMD system led to the best performances such as maximizing JW, TPF, and GOR, as well as minimizing QSH.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131059998","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modeling and Optimizing Multi-Stage Design With Gaussian Process Based on Surrogate Model Chain","authors":"Siyu Yang, Liangyue Jia, J. Hao, R. Alizadeh","doi":"10.1115/detc2022-89859","DOIUrl":"https://doi.org/10.1115/detc2022-89859","url":null,"abstract":"\u0000 Continued progress in the surrogate-model-based evaluation for the single-stage has been explored, but multistage has higher dimension and uncertainty. High dimension and low overall data of multi-stage leads to low accuracy of prediction, and cannot characterize the uncertainty of the final prediction performance. We propose a Gaussian Process-based surrogate model chain (GP-SMC) to evaluate the performance of multi-stage. Also, we combine the GP-SMC with the quasi-Newton method (L-BFGS-B), make full use of the gradient information of the GP-SMC to get an optimization solution rapidly. The MAE (Mean Absolute Error) and MRE (Mean Relative Error) and STD (standard deviation) of GP-SMC’s predicted value are 10% of the prediction of a single surrogate model, which achieves a significant improvement in prediction accuracy and a significant reduction in uncertainty. Compared with the original optimization results, the average performance is improved by 21.05%. Based on the optimal solution and GP-SMC, the confidence interval of the final performance under the optimal solution is obtained, and the confidence level is 99%. The truth probability of GP-SMC is 91.25% in the test dataset, which is higher than single GP’s 85% truth probability. The technology is used in the case of Hot Rod Rolling, and can also be applied to complex product design with multi-stage.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"45 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131462787","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Improving Designer Learning in Design Space Exploration by Adapting to the Designer’s Learning Goals","authors":"Antoni Virós-i-Martin, Daniel Selva","doi":"10.1115/detc2022-89207","DOIUrl":"https://doi.org/10.1115/detc2022-89207","url":null,"abstract":"\u0000 Design space exploration is a design method by which the designer tries to learn important information about a design problem (e.g., main design trade-offs, sensitivities, common features among good designs) to help them make better design decisions. This paper presents preliminary results of a study characterizing the effects on a designer’s learning of an AI assistant that adapts to the designer’s goals during design space exploration. Specifically, we compare the designer’s learning when the AI assistant adapts to explicit learning goals shared by the designer versus when it does not adapt. The AI assistant used for the study is Daphne, which helps engineers design Earth observation satellite systems. The designer’s learning process is modeled as an iterative hypothesis generation and testing process. First, the designer shares with Daphne a certain learning goal in the form of a hypothesis (e.g., designs with feature F are more likely to be on the Pareto front). Then, Daphne adapts to this goal by searching for more designs that have the feature being tested and showing the user the extent to which the data supports their hypothesis. The participants in the preliminary study are N = 10 students from Texas A&M University. We ask each participant to design earth observation satellite constellations to meet a set of requirements while trying to learn about the design problem. The results show that participants with the adaptive AI assistant consistently score higher on their learning about the design task compared to the baseline design assistant as measured by a post-task test. A negative effect is observed on task performance with the adaptive AI assistant condition due to a smaller number of design creation actions, which is consistent with findings from previous studies. Recommendations are provided for the design of similar future AI assistants based on the results of this study. Finally, a power study is done to set a goal for statistical validity of the study.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"90 5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129984064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Topology Optimization of Permanent Magnets for Generators Using Level Set Methods","authors":"Jiawei Tian, Ran Zhuang, Juan Cilia, Arvind Rangarajan, F. Luo, J. Longtin, Shikui Chen","doi":"10.1115/detc2022-90601","DOIUrl":"https://doi.org/10.1115/detc2022-90601","url":null,"abstract":"\u0000 Generators are considered as the core application of electromagnetic machines, which require high-cost rare-earth-based permanent magnets. The development of generators is moving toward high efficiency and increased environmental friendliness. Minimizing the use of rare earth materials such as magnetic materials under the premise of machine performance emerges as a challenging task. Topology optimization has been promisingly applied to many application areas as a powerful generative design tool. It can identify the optimal distribution of magnetic material in the defined design space. This paper employs the level-set-based topology optimization method to design the permanent magnet for generators. The machine under study is a simplified 2D outer rotor direct-drive wind power generator. The dynamic and static models of this generator are studied, and the magnetostatic system is adopted to conduct the topology optimization. The optimization goals in this study mainly focused on two aspects, namely the maximization of the system magnetic energy and the generation of a target magnetic field in the region of the air gap. The continuum shape sensitivity analysis is derived by using the material time derivative, the Lagrange multiplier method, and the adjoint variable method. Two numerical examples are investigated, and the effectiveness of the proposed design framework is validated by comparing the performance of the original design against the optimized design.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121291688","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Are You Feeling Happy? the Effect of Emotions on People’s Interaction Experience Toward Empathetic Chatbots","authors":"Ting Liao, Bei Yan","doi":"10.1115/detc2022-91059","DOIUrl":"https://doi.org/10.1115/detc2022-91059","url":null,"abstract":"\u0000 People may experience emotions before interacting with automated agents to seek information and support. However, existing literature has not well examined how human emotional states affect their interaction experience with agents or how automated agents should react to emotions. This study proposes to test how participants perceive an empathetic agent (chatbot) vs. a non-empathetic one under various emotional states (i.e., positive, neutral, negative) when the chatbot mediates the initial screening process for student advising. Participants are prompted to recall a previous emotional experience and have text-based conversations with the chatbot. The study confirms the importance of presenting empathetic cues in the design of automated agents to support human-agent collaboration. Participants who recall a positive experience are more sensitive to the chatbot’s empathetic behavior. The empathetic behavior of the chatbot improves participants’ satisfaction and makes those who recall a neutral experience feel more positive during the interaction. The results reveal that participants’ emotional states are likely to influence their tendency to self-disclose, interaction experience, and perception of the chatbot’s empathetic behavior. The study also highlights the increasing need for emotional acknowledgment of people who experience positive emotions so that design efforts need to be designated according to people’s dynamic emotional states.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"91 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126753843","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Surrogate Models and Time Series for Flow Prediction on the Red River Dam Network","authors":"R. Alizadeh, J. Allen, F. Mistree","doi":"10.1115/detc2022-88163","DOIUrl":"https://doi.org/10.1115/detc2022-88163","url":null,"abstract":"\u0000 Surrogate models have been used to replace computationally expensive analysis models in engineering design problems. However, time-dependent variables and historical data are usually ignored in the surrogate modeling process. For instance, in a dam network design, using hydraulic simulations to estimate the water flow is computationally expensive, and the data is in the form of time series. So, we need time-dependent surrogate models to replace these simulations and manage this computational complexity. In this paper, we describe surrogate models to predict the amount of water flow into a reservoir. The challenge is that the flow is a time-dependent variable, and we need to incorporate time series into surrogate models. Thus, there are three contributions: (1) using surrogate modeling to predict flow for dam network design, (2) incorporating time series analysis in surrogate models for water network design, (3) using an ensemble of surrogates to increase the accuracy of prediction. We also demonstrate how to integrate surrogate models and machine learning with time series analysis for more accurate and faster prediction. Due to the availability of data, we use the Buffalo Reservoir in the Red River Basin as an example. Based on the time series data for flow, evaporation, precipitation, and maximum and minimum temperature, three surrogate models are used to examine the impact of integrating time series into surrogate models. These are multivariate autoregressive integrated moving average (MARIMA), a classic time series analysis method; artificial neural network (ANN), and random forest (RF) methods, two machine learning surrogate models. We use seven different time lags as features within an RF model, as an ensemble of surrogate models, and predict the flow for seven-time steps ahead. We successfully incorporate the time series data and particularly the concept of the time lag within surrogate models. We show that RF as the ensemble of surrogates provides more accurate predictions than the other two surrogate models. Although this method has been demonstrated for the Red River Basin, it could also be applied to designing anything in which time-dependent flow is an issue, for example, in biomedical applications, the management of manufacturing processes and product sales as well as any products in which fluid flow is an issue.","PeriodicalId":394503,"journal":{"name":"Volume 3B: 48th Design Automation Conference (DAC)","volume":"34 12","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"120921255","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}