This paper proposes an augmented reality (AR) framework and tool on smartphones as an alternative to conventional inspection for AM parts. The framework attempts to introduce the rapid inspection potential of smartphone based AR within manufacturing by leveraging the manufacturing capability of additive manufacturing (AM) to integrate markers onto AM parts. The key step from this framework that is explored in this paper is the design and quality assessment of AM markers for marker registration. As part of the marker design and quality assessment objectives, this research conducts an evaluation on the effects of different AM processes on the quality of augmentation achieved from AM fiducial markers. Furthermore, it evaluates the minimum fiducial pattern size that on integration onto AM parts will be viable for augmentation. The results suggest that the AM process and the size of the fiducial pattern play a significant role in determining the quality of the AM markers. The paper concludes by stating that dual material extrusion AM markers provide the highest number of detectable features and therefore the highest quality of AM markers, and the smallest viable fiducial pattern for Cybercode/QR code marker can be sized at 19 × 19mm2.
{"title":"Quality Assessment of Additively Manufactured Fiducial Markers to Support Augmented Reality-Based Part Inspection","authors":"Jayant Mathur, S. Basu, Jessica Menold, N. Meisel","doi":"10.1115/detc2020-22172","DOIUrl":"https://doi.org/10.1115/detc2020-22172","url":null,"abstract":"\u0000 This paper proposes an augmented reality (AR) framework and tool on smartphones as an alternative to conventional inspection for AM parts. The framework attempts to introduce the rapid inspection potential of smartphone based AR within manufacturing by leveraging the manufacturing capability of additive manufacturing (AM) to integrate markers onto AM parts. The key step from this framework that is explored in this paper is the design and quality assessment of AM markers for marker registration. As part of the marker design and quality assessment objectives, this research conducts an evaluation on the effects of different AM processes on the quality of augmentation achieved from AM fiducial markers. Furthermore, it evaluates the minimum fiducial pattern size that on integration onto AM parts will be viable for augmentation. The results suggest that the AM process and the size of the fiducial pattern play a significant role in determining the quality of the AM markers. The paper concludes by stating that dual material extrusion AM markers provide the highest number of detectable features and therefore the highest quality of AM markers, and the smallest viable fiducial pattern for Cybercode/QR code marker can be sized at 19 × 19mm2.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"142 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122062342","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}
Particle swarm optimization (PSO) method is a well-known optimization algorithm, which shows good performance in solving different optimization problems. However, PSO usually suffers from slow convergence. In this paper, a reinforcement learning method is used to enhance PSO in convergence by replacing the uniformly distributed random number in the updating function by a random number generated from a well-selected normal distribution. The mean and variance of the normal distribution are estimated from the current state of each individual through a policy net. The historic behavior of the swarm group is learned to update the policy net and guide the selection of parameters of the normal distribution. The proposed algorithm is tested with numerical test functions and the results show that the convergence rate of PSO can be improved with the proposed Reinforcement Learning method (RL-PSO).
{"title":"Enhanced Particle Swarm Optimization via Reinforcement Learning","authors":"Di Wu, G. Wang","doi":"10.1115/detc2020-22519","DOIUrl":"https://doi.org/10.1115/detc2020-22519","url":null,"abstract":"\u0000 Particle swarm optimization (PSO) method is a well-known optimization algorithm, which shows good performance in solving different optimization problems. However, PSO usually suffers from slow convergence. In this paper, a reinforcement learning method is used to enhance PSO in convergence by replacing the uniformly distributed random number in the updating function by a random number generated from a well-selected normal distribution. The mean and variance of the normal distribution are estimated from the current state of each individual through a policy net. The historic behavior of the swarm group is learned to update the policy net and guide the selection of parameters of the normal distribution. The proposed algorithm is tested with numerical test functions and the results show that the convergence rate of PSO can be improved with the proposed Reinforcement Learning method (RL-PSO).","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"136 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131322130","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}
Topology optimization (TO) was conducted for three dimensional static fluid mixers. The problem is optimized using the weakly coupled Navier-Stokes equation at low Reynolds number (Re ≤ 1) and a convection-diffusion equation. The domain was discretized with up to 10 million cells. The optimizations were run with 1024 to 2048 CPUs on a national supercomputer. For a mixer in a square cross-section channel, the mixing was improved by 83% for a modest 2.5 times higher pressure drop compared with the open straight channel. For a cylindrical cross-section tee arrangement, the mixing improved by 91% with a 2.5 times higher pressure drop compared to the straight channel.
{"title":"Large Scale Topology Optimization of 3D Static Mixers","authors":"Si-ying Sun, J. Ghandhi, Xiaoping Qian","doi":"10.1115/detc2020-22132","DOIUrl":"https://doi.org/10.1115/detc2020-22132","url":null,"abstract":"\u0000 Topology optimization (TO) was conducted for three dimensional static fluid mixers. The problem is optimized using the weakly coupled Navier-Stokes equation at low Reynolds number (Re ≤ 1) and a convection-diffusion equation. The domain was discretized with up to 10 million cells. The optimizations were run with 1024 to 2048 CPUs on a national supercomputer. For a mixer in a square cross-section channel, the mixing was improved by 83% for a modest 2.5 times higher pressure drop compared with the open straight channel. For a cylindrical cross-section tee arrangement, the mixing improved by 91% with a 2.5 times higher pressure drop compared to the straight channel.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131383026","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}
Binyang Song, Emmett Meinzer, Akash Agrawal, Christopher McComb
The elicitation of customer pain points is a crucial early step in the design or redesign of successful products and services. Online, user-generated data contains rich, real-time information about customer experience, requirements, and preferences. However, it is a nontrivial task to retrieve useful information from these sources because of the sheer amount of data, often unstructured. In this work, we build on previous efforts that used natural language processing techniques to extract meaning from online data and facilitate experiential redesign and extend them by integrating a sentiment analysis. As a use case, we explore the airline industry. A considerable portion of potential passengers opt out of traveling by airplane due to aviophobia, a fear of flying. This causes a market loss to the industry and inconvenience for those who experience aviophobia. The potential contributors to aviophobia are complex and diverse, involving physical, psychological and emotional reactions to the air travel experience. A methodology that is capable of accommodating the complexity and diversity of the commercial airline industry user-generated data is necessary to effectively mine customer pain points. To address the demand, we propose a novel methodology in this study. Using passenger commentary data posted on Reddit, the method implements topic modeling to extract common themes from the commentaries and employs sentiment analysis to elicit and interpret the salient information contained in the extracted themes. This paper ends by providing specific recommendations that are germane to the use case as well as suggesting future research directions.
{"title":"Topic Modeling and Sentiment Analysis of Social Media Data to Drive Experiential Redesign","authors":"Binyang Song, Emmett Meinzer, Akash Agrawal, Christopher McComb","doi":"10.1115/detc2020-22567","DOIUrl":"https://doi.org/10.1115/detc2020-22567","url":null,"abstract":"\u0000 The elicitation of customer pain points is a crucial early step in the design or redesign of successful products and services. Online, user-generated data contains rich, real-time information about customer experience, requirements, and preferences. However, it is a nontrivial task to retrieve useful information from these sources because of the sheer amount of data, often unstructured. In this work, we build on previous efforts that used natural language processing techniques to extract meaning from online data and facilitate experiential redesign and extend them by integrating a sentiment analysis. As a use case, we explore the airline industry. A considerable portion of potential passengers opt out of traveling by airplane due to aviophobia, a fear of flying. This causes a market loss to the industry and inconvenience for those who experience aviophobia. The potential contributors to aviophobia are complex and diverse, involving physical, psychological and emotional reactions to the air travel experience. A methodology that is capable of accommodating the complexity and diversity of the commercial airline industry user-generated data is necessary to effectively mine customer pain points. To address the demand, we propose a novel methodology in this study. Using passenger commentary data posted on Reddit, the method implements topic modeling to extract common themes from the commentaries and employs sentiment analysis to elicit and interpret the salient information contained in the extracted themes. This paper ends by providing specific recommendations that are germane to the use case as well as suggesting future research directions.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"523 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115937565","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}
Jean-François Gamache, A. Vadean, Nicolas Dodane, S. Achiche
Reducing the weight of structures remains a major challenge in the aviation industry in order to reduce fuel consumption. The stiffened panel is the main assembly method for primary structures in aircraft, e.g. fuselage or wing. Density-based topology optimization has been used in research and in industry as a tool to help create new stiffening patterns for aircraft components, such as ribs, spars, bulkheads or even floor design. One critical aspect of stiffened panel design for wing structures is the buckling resistance. However, most work found in the literature does not include buckling analysis during optimization which leads to sub-optimal results when the stiffening layout is validated for buckling. Including buckling as a constraint for the density-based topology optimization has proven to be a complex task, mainly caused by the fact that the buckling of the stiffeners is not captured accurately. As such, this work presents an optimization method for stiffened panels based on the ground structure approach usually used for truss topology optimization. The main novelty of the method is the use of a stiffener activation variable (SAV) to activate only one variable at a time, either the height or density variable associated with each stiffeners of the ground structure. This work shows that while ground structure topology optimization requires more setup time and limiting the degrees of freedom of the optimization, it finds the best stiffening layout efficiently when compared to the density method.
{"title":"Topology Optimization for Stiffened Panels: A Ground Structure Method","authors":"Jean-François Gamache, A. Vadean, Nicolas Dodane, S. Achiche","doi":"10.1115/detc2020-22103","DOIUrl":"https://doi.org/10.1115/detc2020-22103","url":null,"abstract":"\u0000 Reducing the weight of structures remains a major challenge in the aviation industry in order to reduce fuel consumption. The stiffened panel is the main assembly method for primary structures in aircraft, e.g. fuselage or wing.\u0000 Density-based topology optimization has been used in research and in industry as a tool to help create new stiffening patterns for aircraft components, such as ribs, spars, bulkheads or even floor design.\u0000 One critical aspect of stiffened panel design for wing structures is the buckling resistance. However, most work found in the literature does not include buckling analysis during optimization which leads to sub-optimal results when the stiffening layout is validated for buckling.\u0000 Including buckling as a constraint for the density-based topology optimization has proven to be a complex task, mainly caused by the fact that the buckling of the stiffeners is not captured accurately. As such, this work presents an optimization method for stiffened panels based on the ground structure approach usually used for truss topology optimization. The main novelty of the method is the use of a stiffener activation variable (SAV) to activate only one variable at a time, either the height or density variable associated with each stiffeners of the ground structure.\u0000 This work shows that while ground structure topology optimization requires more setup time and limiting the degrees of freedom of the optimization, it finds the best stiffening layout efficiently when compared to the density method.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122761293","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}
Zhuoyuan Zheng, Parth Bansal, Pingfeng Wang, Yumeng Li
Lithium cobalt oxide (LCO) cathode is one of the most commonly used positive active materials for lithium ion batteries (LIBs). However, one reliability issue that limits its applications is the presence of moisture adsorbing on the LCO cathode surface, which may react with the electrolyte and lead to detrimental effects. In this study, a novel LCO thin film cathode is proposed, by adding a layer of high diffusion resistant (003) phase of LCO as the protective coating onto the diffusion favorable (110) phase, in order to simultaneously achieve good electrochemical performance and chemical stability of the LIB. A multi-physics-based finite element model is built to investigate the performance of the cathode and the influences of the design and operation variables, including the layout the two crystal phases, the fraction of each phase and the lithiation C rate. In addition, a Gaussian Process based surrogate model is developed, using the simulated results from the FE model as training data, to efficiently explore the design space of the cathode. It is found that, the 110//003 layout cathode could provide high capacity and good rate performance; meanwhile, the 003//110 design may lead to a largely reduced capacity, especially at high lithiation C rates.
{"title":"Simulation Assisted Design of LCO Cathode Materials With High Performance Stability","authors":"Zhuoyuan Zheng, Parth Bansal, Pingfeng Wang, Yumeng Li","doi":"10.1115/detc2020-22357","DOIUrl":"https://doi.org/10.1115/detc2020-22357","url":null,"abstract":"\u0000 Lithium cobalt oxide (LCO) cathode is one of the most commonly used positive active materials for lithium ion batteries (LIBs). However, one reliability issue that limits its applications is the presence of moisture adsorbing on the LCO cathode surface, which may react with the electrolyte and lead to detrimental effects. In this study, a novel LCO thin film cathode is proposed, by adding a layer of high diffusion resistant (003) phase of LCO as the protective coating onto the diffusion favorable (110) phase, in order to simultaneously achieve good electrochemical performance and chemical stability of the LIB. A multi-physics-based finite element model is built to investigate the performance of the cathode and the influences of the design and operation variables, including the layout the two crystal phases, the fraction of each phase and the lithiation C rate. In addition, a Gaussian Process based surrogate model is developed, using the simulated results from the FE model as training data, to efficiently explore the design space of the cathode. It is found that, the 110//003 layout cathode could provide high capacity and good rate performance; meanwhile, the 003//110 design may lead to a largely reduced capacity, especially at high lithiation C rates.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"139 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127470692","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}
The data-driven approach is emerging as a promising method for the topological design of the multiscale structure with greater efficiency. However, existing data-driven methods mostly focus on a single class of unit cells without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of inherent ordering or “distance” measure between different classes of unit cells in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) to creating multi-response LVGP (MRLVGP) for the unit cell libraries of metamaterials, taking both qualitative unit cell concepts and quantitative unit cell design variables as mixed-variable inputs. The MRLVGP embeds the mixed variables into a continuous design space based on their collective effect on the responses, providing substantial insights into the interplay between different geometrical classes and unit cell materials. With this model, we can easily obtain a continuous and differentiable transition between different unit cell concepts that can render gradient information for multiscale topology optimization. While the proposed approach has a broader impact on the concurrent topological and material design of engineered systems, we demonstrate its benefits through multiscale topology optimization with aperiodic unit cells. Design examples reveal that considering multiple unit cell types can lead to improved performance due to the consistent load-transferred paths for micro- and macrostructures.
{"title":"Data-Driven Multiscale Topology Optimization Using Multi-Response Latent Variable Gaussian Process","authors":"Liwei Wang, Siyu Tao, P. Zhu, Wei Chen","doi":"10.1115/detc2020-22595","DOIUrl":"https://doi.org/10.1115/detc2020-22595","url":null,"abstract":"\u0000 The data-driven approach is emerging as a promising method for the topological design of the multiscale structure with greater efficiency. However, existing data-driven methods mostly focus on a single class of unit cells without considering multiple classes to accommodate spatially varying desired properties. The key challenge is the lack of inherent ordering or “distance” measure between different classes of unit cells in meeting a range of properties. To overcome this hurdle, we extend the newly developed latent-variable Gaussian process (LVGP) to creating multi-response LVGP (MRLVGP) for the unit cell libraries of metamaterials, taking both qualitative unit cell concepts and quantitative unit cell design variables as mixed-variable inputs. The MRLVGP embeds the mixed variables into a continuous design space based on their collective effect on the responses, providing substantial insights into the interplay between different geometrical classes and unit cell materials. With this model, we can easily obtain a continuous and differentiable transition between different unit cell concepts that can render gradient information for multiscale topology optimization. While the proposed approach has a broader impact on the concurrent topological and material design of engineered systems, we demonstrate its benefits through multiscale topology optimization with aperiodic unit cells. Design examples reveal that considering multiple unit cell types can lead to improved performance due to the consistent load-transferred paths for micro- and macrostructures.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133743960","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}
Accurate fault defection of bearing is critical in condition-based maintenance to improve system reliability and reduce operational cost. This paper introduces a deep transfer learning-based approach for bearing fault diagnosis by fusing heterogeneous information from multiple sources. Convolutional neural networks (CNN) are first designed to extract critical features by mapping extremely high-dimensional signals such as vibration and images to a much lower dimensional latent space. By partially retaining the resultant CNN architectures and parameters, it becomes possible to transfer and fuse the knowledge gained from multiple heterogeneous sources to improve the robustness and accuracy of fault diagnosis of bearings. With the prior knowledge, a deep transfer learning (DTL) architecture is designed to incorporate the heterogeneous data and trained to detect bearing faults. To future improve the performance of bearing fault diagnosis, a performance-driven optimization approach is developed to optimize the validation accuracy of bearing diagnosis by successively designing the architectures of the deep transfer networks. The CWRU experimental data is utilized to demonstrate the performance of the proposed approach.
{"title":"Designing Deep Transfer Networks for Bearing Fault Diagnosis With Heterogeneous Data Fusion","authors":"Yunsheng Su, Zequn Wang","doi":"10.1115/detc2020-22203","DOIUrl":"https://doi.org/10.1115/detc2020-22203","url":null,"abstract":"\u0000 Accurate fault defection of bearing is critical in condition-based maintenance to improve system reliability and reduce operational cost. This paper introduces a deep transfer learning-based approach for bearing fault diagnosis by fusing heterogeneous information from multiple sources. Convolutional neural networks (CNN) are first designed to extract critical features by mapping extremely high-dimensional signals such as vibration and images to a much lower dimensional latent space. By partially retaining the resultant CNN architectures and parameters, it becomes possible to transfer and fuse the knowledge gained from multiple heterogeneous sources to improve the robustness and accuracy of fault diagnosis of bearings. With the prior knowledge, a deep transfer learning (DTL) architecture is designed to incorporate the heterogeneous data and trained to detect bearing faults. To future improve the performance of bearing fault diagnosis, a performance-driven optimization approach is developed to optimize the validation accuracy of bearing diagnosis by successively designing the architectures of the deep transfer networks. The CWRU experimental data is utilized to demonstrate the performance of the proposed approach.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"30 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130312543","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}
Despite the numerous advances, reinforcement learning remains away from widespread acceptance for autonomous controller design as compared to classical methods due to lack of ability to effectively tackle uncertainty. The reliance on absolute or deterministic reward as a metric for optimization process renders reinforcement learning highly susceptible to changes in problem dynamics. We introduce a novel framework that effectively quantify the uncertainty in the design space and induces robustness in controllers by switching to a reliability-based optimization routine. A model-based approach is used to improve the data efficiency of the method while predicting the system dynamics. We prove the stability of learned neuro-controllers in both static and dynamic environments on classical reinforcement learning tasks such as Cart Pole balancing and Inverted Pendulum.
{"title":"Reliability-Based Reinforcement Learning Under Uncertainty","authors":"Zequn Wang, Narendra Patwardhan","doi":"10.1115/detc2020-22019","DOIUrl":"https://doi.org/10.1115/detc2020-22019","url":null,"abstract":"\u0000 Despite the numerous advances, reinforcement learning remains away from widespread acceptance for autonomous controller design as compared to classical methods due to lack of ability to effectively tackle uncertainty. The reliance on absolute or deterministic reward as a metric for optimization process renders reinforcement learning highly susceptible to changes in problem dynamics. We introduce a novel framework that effectively quantify the uncertainty in the design space and induces robustness in controllers by switching to a reliability-based optimization routine. A model-based approach is used to improve the data efficiency of the method while predicting the system dynamics. We prove the stability of learned neuro-controllers in both static and dynamic environments on classical reinforcement learning tasks such as Cart Pole balancing and Inverted Pendulum.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"115 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116882765","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}
Fluid-flow topology optimization (FTO) allows the generation of innovative flow-channel layouts with minimal pressure drop (power dissipation) between inlet and outlet ports in a given design domain. FTO was first explored using Stokes flow with the material in the design domain modeled as a porous medium governed by Darcy’s law. More recently, Navier-Stokes flow has been implemented to consider higher Reynolds numbers. The objective of this work is to demonstrate the effect of the Reynolds number on the FTO results and generate a set of design rules. To this end, a density-based FTO algorithm and an in-house finite element analysis code for incompressible Navier-Stokes flow are developed. The optimization process is updated using the method of moving asymptotes so that the flow’s potential power is maximized. The nonlinear Navier-Stokes equations are solved using a trust region Newton’s method. Sensitivity analysis is carried out using the adjoint method. A parametric study of the underlying parameters of the Reynolds number in two numerical examples shows the effect of the fluid’s dynamic viscosity and velocity on the optimized flow channels. The results show that fluids with the same Reynolds number, but with different dynamic viscosity or velocity values, can generate significantly different flow channels.
{"title":"A Parametric Study on the Effects of Reynolds Number on the Topology Optimization of Navier-Stokes Flows","authors":"Joel C. Najmon, Tong Wu, A. Tovar","doi":"10.1115/detc2020-22690","DOIUrl":"https://doi.org/10.1115/detc2020-22690","url":null,"abstract":"\u0000 Fluid-flow topology optimization (FTO) allows the generation of innovative flow-channel layouts with minimal pressure drop (power dissipation) between inlet and outlet ports in a given design domain. FTO was first explored using Stokes flow with the material in the design domain modeled as a porous medium governed by Darcy’s law. More recently, Navier-Stokes flow has been implemented to consider higher Reynolds numbers. The objective of this work is to demonstrate the effect of the Reynolds number on the FTO results and generate a set of design rules. To this end, a density-based FTO algorithm and an in-house finite element analysis code for incompressible Navier-Stokes flow are developed. The optimization process is updated using the method of moving asymptotes so that the flow’s potential power is maximized. The nonlinear Navier-Stokes equations are solved using a trust region Newton’s method. Sensitivity analysis is carried out using the adjoint method. A parametric study of the underlying parameters of the Reynolds number in two numerical examples shows the effect of the fluid’s dynamic viscosity and velocity on the optimized flow channels. The results show that fluids with the same Reynolds number, but with different dynamic viscosity or velocity values, can generate significantly different flow channels.","PeriodicalId":415040,"journal":{"name":"Volume 11A: 46th Design Automation Conference (DAC)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2020-08-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"129824152","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}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：