Pub Date : 2025-06-18DOI: 10.1016/j.cad.2025.103897
Wenqiang Tang , Zhouwang Yang
B-spline curve parameterization is a complex nonlinear and non-convex optimization problem. Traditional optimization methods often struggle with local minima and are computationally expensive, especially in high-dimensional spaces. We proposes a deep neural network (DNN)-based method to efficiently solve the parameterization problem in B-spline curve approximation. The designed parameterization network (PNet) maps the initial parameterization to an optimized one, transforming the problem into a search for suitable network parameters in a high-dimensional feature space. Due to the over-parameterization nature of DNNs, PNet is robust to initial conditions and less prone to local minima. Furthermore, the smooth regularization and top- loss function are introduced to further enhance optimization performance. Experimental results show that PNet achieves high-precision approximation with remarkable efficiency, even for large-scale point clouds.
{"title":"DNN-based Parameterization for B-Spline Curve Approximation","authors":"Wenqiang Tang , Zhouwang Yang","doi":"10.1016/j.cad.2025.103897","DOIUrl":"10.1016/j.cad.2025.103897","url":null,"abstract":"<div><div>B-spline curve parameterization is a complex nonlinear and non-convex optimization problem. Traditional optimization methods often struggle with local minima and are computationally expensive, especially in high-dimensional spaces. We proposes a deep neural network (DNN)-based method to efficiently solve the parameterization problem in B-spline curve approximation. The designed parameterization network (PNet) maps the initial parameterization to an optimized one, transforming the problem into a search for suitable network parameters in a high-dimensional feature space. Due to the over-parameterization nature of DNNs, PNet is robust to initial conditions and less prone to local minima. Furthermore, the smooth regularization and top-<span><math><mi>K</mi></math></span> loss function are introduced to further enhance optimization performance. Experimental results show that PNet achieves high-precision approximation with remarkable efficiency, even for large-scale point clouds.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"186 ","pages":"Article 103897"},"PeriodicalIF":3.0,"publicationDate":"2025-06-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144330099","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-11DOI: 10.1016/j.cad.2025.103899
Soroush Masoudi , Barun K. Das , Muhammad Aamir , Majid Tolouei-Rad
The cooling phase of a moulded part plays a crucial role in the injection moulding (IM) process, accounting for 50 to 80 % of total cycle time, and significantly impacting the quality of moulded parts. During the last decade, the advancement of different additive manufacturing (AM) processes, especially metal 3D printing, has facilitated the production of mould parts, such as cores and cavities, with complex-shaped internal conformal cooling channels (CCCs). These innovative cooling systems exhibit significant potential to replace traditional straight-drilled cooling design, as they offer more efficient and uniform cooling performance by facilitating more effective heat transfer, considerably enhancing production quality and efficiency. Despite the growing attention being given to the design and manufacturing of CCC systems, there is still a lack of systematic and comprehensive classification, comparison and evaluation methodologies. This paper aims to review various types of conformal cooling channels, such as spiral, zigzag and nature-based designs, among others, and to provide an overview of advancements in design, process modeling and simulation of this new cooling technology. Previous studies have indicated that conventional straight-drilled cooling channels are likely to be replaced by CCCs, especially for complex and sensitive parts, due to their superior performance in reducing cycle times and enhancing product quality. The present study explores various challenges that arise when developing conformal cooling channels. These issues range from optimizing complex geometry to improving thermal performance through modelling. However, these challenges also present opportunities for innovation and advancement in mould design and manufacturing.
{"title":"Recent advancement in conformal cooling channels: A review on design, simulation and future trends","authors":"Soroush Masoudi , Barun K. Das , Muhammad Aamir , Majid Tolouei-Rad","doi":"10.1016/j.cad.2025.103899","DOIUrl":"10.1016/j.cad.2025.103899","url":null,"abstract":"<div><div>The cooling phase of a moulded part plays a crucial role in the injection moulding (IM) process, accounting for 50 to 80 % of total cycle time, and significantly impacting the quality of moulded parts. During the last decade, the advancement of different additive manufacturing (AM) processes, especially metal 3D printing, has facilitated the production of mould parts, such as cores and cavities, with complex-shaped internal conformal cooling channels (CCCs). These innovative cooling systems exhibit significant potential to replace traditional straight-drilled cooling design, as they offer more efficient and uniform cooling performance by facilitating more effective heat transfer, considerably enhancing production quality and efficiency. Despite the growing attention being given to the design and manufacturing of CCC systems, there is still a lack of systematic and comprehensive classification, comparison and evaluation methodologies. This paper aims to review various types of conformal cooling channels, such as spiral, zigzag and nature-based designs, among others, and to provide an overview of advancements in design, process modeling and simulation of this new cooling technology. Previous studies have indicated that conventional straight-drilled cooling channels are likely to be replaced by CCCs, especially for complex and sensitive parts, due to their superior performance in reducing cycle times and enhancing product quality. The present study explores various challenges that arise when developing conformal cooling channels. These issues range from optimizing complex geometry to improving thermal performance through modelling. However, these challenges also present opportunities for innovation and advancement in mould design and manufacturing.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"186 ","pages":"Article 103899"},"PeriodicalIF":3.0,"publicationDate":"2025-06-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144321227","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-06-04DOI: 10.1016/j.cad.2025.103898
Jin Wang, Mingliang Zhu, Zhiwei Miao
This paper proposes a two-stage optimization deep neural network method for form-finding of tensegrity structures, based on physical information. The total loss function of the neural network is constructed by comprehensively considering the physical information, including nodal residual forces, element length constraints, and minimum node distance. To enhance the learning ability of the neural network, a two-stage optimization model is adopted. In the first stage, the AdamW optimizer is employed for preliminary training of the network's hyperparameters, quickly reducing the loss values. Following the preliminary training, the l-BFGS optimizer is utilized in the second stage to refine the optimization and converge toward the optimal solution, resulting in the nodal coordinates that satisfy the structural equilibrium. The paper includes case studies on five different tensegrity models. The results show that the proposed two-stage physics-informed deep neural network (PIDNN) approach, utilizing dual optimizers, can efficiently and accurately perform form-finding for various tensegrity structures, including both single- and multi-stable models. The method provides reliable results, avoids complex finite element computations, and offers high computational efficiency.
{"title":"Two-stage physics-informed deep neural networks framework for form-finding of tensegrity structures","authors":"Jin Wang, Mingliang Zhu, Zhiwei Miao","doi":"10.1016/j.cad.2025.103898","DOIUrl":"10.1016/j.cad.2025.103898","url":null,"abstract":"<div><div>This paper proposes a two-stage optimization deep neural network method for form-finding of tensegrity structures, based on physical information. The total loss function of the neural network is constructed by comprehensively considering the physical information, including nodal residual forces, element length constraints, and minimum node distance. To enhance the learning ability of the neural network, a two-stage optimization model is adopted. In the first stage, the AdamW optimizer is employed for preliminary training of the network's hyperparameters, quickly reducing the loss values. Following the preliminary training, the l-BFGS optimizer is utilized in the second stage to refine the optimization and converge toward the optimal solution, resulting in the nodal coordinates that satisfy the structural equilibrium. The paper includes case studies on five different tensegrity models. The results show that the proposed two-stage physics-informed deep neural network (PIDNN) approach, utilizing dual optimizers, can efficiently and accurately perform form-finding for various tensegrity structures, including both single- and multi-stable models. The method provides reliable results, avoids complex finite element computations, and offers high computational efficiency.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"186 ","pages":"Article 103898"},"PeriodicalIF":3.0,"publicationDate":"2025-06-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144243171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-05-24DOI: 10.1016/j.cad.2025.103896
Zhentong Xu , Long Zeng , Junli Zhao , Baodong Wang , Zhenkuan Pan , Yong-Jin Liu
With the development of generative techniques, sketch-driven 3D reconstruction has gained substantial attention as an efficient 3D modeling technique. However, challenges remain in extracting detailed features from sketches, representing local geometric structures, and ensuring generated fidelity and stability. To address these issues, in this paper we propose a multi-spectral channel cross-attention model for sketch reconstruction, which leverages the complementary strengths of frequency and spatial domains to capture multi-level sketch features. Our method employs a two-stage diffusion generation mechanism, additionally, a Sparse Feature Enhancement Module (SFE) replaces traditional down-sampling, reducing feature loss and enhancing detail preservation and noise suppression through a Laplace voxel smoothing operator. The Wasserstein distance introduced and integrated as part of the loss function, stabilizes the generative process using optimal transport theory to support high-quality 3D model reconstruction. Extensive experiments verify that our model surpasses state-of-the-art methods in terms of generation accuracy, local control, and generalization ability, providing an efficient, precise solution for transforming sketches into 3D models.
{"title":"Sketch123: Multi-spectral channel cross attention for sketch-based 3D generation via diffusion models","authors":"Zhentong Xu , Long Zeng , Junli Zhao , Baodong Wang , Zhenkuan Pan , Yong-Jin Liu","doi":"10.1016/j.cad.2025.103896","DOIUrl":"10.1016/j.cad.2025.103896","url":null,"abstract":"<div><div>With the development of generative techniques, sketch-driven 3D reconstruction has gained substantial attention as an efficient 3D modeling technique. However, challenges remain in extracting detailed features from sketches, representing local geometric structures, and ensuring generated fidelity and stability. To address these issues, in this paper we propose a multi-spectral channel cross-attention model for sketch reconstruction, which leverages the complementary strengths of frequency and spatial domains to capture multi-level sketch features. Our method employs a two-stage diffusion generation mechanism, additionally, a Sparse Feature Enhancement Module (SFE) replaces traditional down-sampling, reducing feature loss and enhancing detail preservation and noise suppression through a Laplace voxel smoothing operator. The Wasserstein distance introduced and integrated as part of the loss function, stabilizes the generative process using optimal transport theory to support high-quality 3D model reconstruction. Extensive experiments verify that our model surpasses state-of-the-art methods in terms of generation accuracy, local control, and generalization ability, providing an efficient, precise solution for transforming sketches into 3D models.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"185 ","pages":"Article 103896"},"PeriodicalIF":3.0,"publicationDate":"2025-05-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144139049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-05-09DOI: 10.1016/j.cad.2025.103888
Jie Zhang , Luwei Chen , Hang Ren , Yan Luximon , Ping Li
The realm of personalized human wearable has undergone significant evolution. However, existing methodologies often lack in addressing crucial factors essential for comprehensive facewear customization, including automation, accuracy, convenience, and comfort. To bridge this gap, we propose an automated and user-friendly approach for ergonomic facewear personalized based on 3D face modeling and registration. Our approach involves the development of a user-friendly 3D face scanning technique utilizing RGB-D selfies captured by smartphone depth cameras, simplifying the process of personal data collection. Subsequently, we present a precise 3D symmetry-aware face registration method integrated with automatic facial landmark detection to produce parameterized 3D facial meshes featuring accurate symmetry panels. Additionally, we introduce a comfort-centric eyeglasses customization process to showcase its practical applications. Qualitative and quantitative comparisons with state-of-the-art techniques highlight the efficacy of our symmetry-aware registration method. Experimental findings also showcase the versatility of our approach across various scanning data. Evaluation results underscore the practical utility of our method in creating personalized eyeglasses with significantly enhanced comfort levels. Moreover, our method can be extended to other symmetric facewear designs, such as sunglasses, swimming goggles, and respirators.
{"title":"Face2Wear: An automatic and user-friendly facewear personalization framework with 3D symmetry-aware face registration using RGB-D selfies","authors":"Jie Zhang , Luwei Chen , Hang Ren , Yan Luximon , Ping Li","doi":"10.1016/j.cad.2025.103888","DOIUrl":"10.1016/j.cad.2025.103888","url":null,"abstract":"<div><div>The realm of personalized human wearable has undergone significant evolution. However, existing methodologies often lack in addressing crucial factors essential for comprehensive facewear customization, including automation, accuracy, convenience, and comfort. To bridge this gap, we propose an automated and user-friendly approach for ergonomic facewear personalized based on 3D face modeling and registration. Our approach involves the development of a user-friendly 3D face scanning technique utilizing RGB-D selfies captured by smartphone depth cameras, simplifying the process of personal data collection. Subsequently, we present a precise 3D symmetry-aware face registration method integrated with automatic facial landmark detection to produce parameterized 3D facial meshes featuring accurate symmetry panels. Additionally, we introduce a comfort-centric eyeglasses customization process to showcase its practical applications. Qualitative and quantitative comparisons with state-of-the-art techniques highlight the efficacy of our symmetry-aware registration method. Experimental findings also showcase the versatility of our approach across various scanning data. Evaluation results underscore the practical utility of our method in creating personalized eyeglasses with significantly enhanced comfort levels. Moreover, our method can be extended to other symmetric facewear designs, such as sunglasses, swimming goggles, and respirators.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"185 ","pages":"Article 103888"},"PeriodicalIF":3.0,"publicationDate":"2025-05-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144068931","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-23DOI: 10.1016/j.cad.2025.103887
Michal Bizzarri , Kanika Rajain , Michael Bartoň
Geometric modeling is traditionally a key part of an efficient manufacturing pipeline as one can decide, in virtual realm, what specific manufacturing tools to use and how to move them. Flank milling is the finishing stage of 5-axis Computer Numerically Controlled (CNC) machining, a stage where the machining accuracy is equally important as the smooth surface finish of the to-be-manufactured workpiece. The benchmark machining geometries such as propellers or blisks are doubly-curved surfaces and one typically needs several paths of the tool to get highly accurate surface finish. However, navigating a tool to move tangentially (i.e., in flank fashion) to the surface is very restrictive and in order to get highly accurate approximation, one typically has to compromise the smoothness across the neighboring paths.
To connect neighboring paths in smooth (-continuous) fashion using a conical tool is possible only for reasonably flat target geometries, such as spiral bevel gears, however, for a general free-form surface conical tools do not offer sufficient degrees of freedom. In this work, we consider generally curved, custom-shaped, cutting tools, whose shape is a design parameter computed by the proposed optimization-based framework to adapt their motions globally to the input free-form surface, supporting a feature of connection across the neighboring paths. We demonstrate our algorithm on synthetic free-form surfaces as well as on industrial benchmark datasets, showing that optimizing the shape of the tool offers more flexibility to produce connections between neighboring strips and outperforms conical tools both in terms of the approximation error and the smoothness.
{"title":"Smooth surface finishing for 5-axis flank CNC machining of free-form geometries using custom-shaped tools","authors":"Michal Bizzarri , Kanika Rajain , Michael Bartoň","doi":"10.1016/j.cad.2025.103887","DOIUrl":"10.1016/j.cad.2025.103887","url":null,"abstract":"<div><div>Geometric modeling is traditionally a key part of an efficient manufacturing pipeline as one can decide, in virtual realm, what specific manufacturing tools to use and how to move them. Flank milling is the finishing stage of 5-axis Computer Numerically Controlled (CNC) machining, a stage where the machining accuracy is equally important as the smooth surface finish of the to-be-manufactured workpiece. The benchmark machining geometries such as propellers or blisks are doubly-curved surfaces and one typically needs several paths of the tool to get highly accurate surface finish. However, navigating a tool to move tangentially (i.e., in flank fashion) to the surface is very restrictive and in order to get highly accurate approximation, one typically has to compromise the smoothness across the neighboring paths.</div><div>To connect neighboring paths in smooth (<span><math><msup><mrow><mi>G</mi></mrow><mrow><mn>1</mn></mrow></msup></math></span>-continuous) fashion using a conical tool is possible only for reasonably flat target geometries, such as spiral bevel gears, however, for a general free-form surface conical tools do not offer sufficient degrees of freedom. In this work, we consider generally curved, custom-shaped, cutting tools, whose shape is a design parameter computed by the proposed optimization-based framework to adapt their motions globally to the input free-form surface, supporting a feature of <span><math><msup><mrow><mi>G</mi></mrow><mrow><mn>1</mn></mrow></msup></math></span> connection across the neighboring paths. We demonstrate our algorithm on synthetic free-form surfaces as well as on industrial benchmark datasets, showing that optimizing the shape of the tool offers more flexibility to produce <span><math><msup><mrow><mi>G</mi></mrow><mrow><mn>1</mn></mrow></msup></math></span> connections between neighboring strips and outperforms conical tools both in terms of the approximation error and the smoothness.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"185 ","pages":"Article 103887"},"PeriodicalIF":3.0,"publicationDate":"2025-04-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143888011","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-21DOI: 10.1016/j.cad.2025.103886
Bohumír Bastl
In this paper, we present a fully automatic approach to generate a geometric model of a Francis turbine runner wheel based on NURBS surfaces from given shape parameters and also a fully automatic approach to generate NURBS meshes of the inner parts of the runner wheel based on NURBS volumes. All the steps of the presented approaches are described in detail, including several challenges that need to be overcome, such as e.g. obtaining conformal parameterizations of streamsurfaces or automatic determination of suitable B-spline approximation curves for representing spatial blade profiles. NURBS meshes generated by the presented method are of good quality and can be used directly for numerical simulations of incompressible turbulent fluid flows based on isogeometric analysis, or, after simple conversion to hexahedral meshes, based on finite element method. The presented approaches can also be used in automatic shape optimization algorithms for Francis turbine runners based on gradient or gradient-free approaches.
{"title":"Parametric generators of geometric models and computational meshes of Francis turbines runners","authors":"Bohumír Bastl","doi":"10.1016/j.cad.2025.103886","DOIUrl":"10.1016/j.cad.2025.103886","url":null,"abstract":"<div><div>In this paper, we present a fully automatic approach to generate a geometric model of a Francis turbine runner wheel based on NURBS surfaces from given shape parameters and also a fully automatic approach to generate NURBS meshes of the inner parts of the runner wheel based on NURBS volumes. All the steps of the presented approaches are described in detail, including several challenges that need to be overcome, such as e.g. obtaining conformal parameterizations of streamsurfaces or automatic determination of suitable B-spline approximation curves for representing spatial blade profiles. NURBS meshes generated by the presented method are of good quality and can be used directly for numerical simulations of incompressible turbulent fluid flows based on isogeometric analysis, or, after simple conversion to hexahedral meshes, based on finite element method. The presented approaches can also be used in automatic shape optimization algorithms for Francis turbine runners based on gradient or gradient-free approaches.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"184 ","pages":"Article 103886"},"PeriodicalIF":3.0,"publicationDate":"2025-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143869118","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-04-02DOI: 10.1016/j.cad.2025.103876
Jiajie Guo, Ming Li
We propose an augmented sphere tracing (AST) pipeline that seamlessly integrates editing, rendering, and slicing of mega-scale periodic shell-lattice structures. Traditional STL-based pipelines face challenges such as time-consuming format conversions, high storage requirements, and complex blending issues between discrete lattice and shell components, often resulting in a loss of geometric accuracy. Alternatively, implicit-based pipelines excel at smooth modeling and robust Boolean operations but require inefficient and error-prone conversions of STL shells into implicit forms, complicating the rendering process. To address these issues, AST combines hybrid implicit lattice and mesh shell representations, eliminating the need for explicit 3D model construction and unnecessary geometric format conversions. It overcomes the major challenges of hybrid forms and mega-scale rendering by using an augmented tracing distance query that avoids costly signed distance field (SDF) calculations while preserving geometric details. Additionally, it employs a local tracing distance query within a single cell, leveraging lattice periodicity for efficiency. The pipeline also supports various types of shell-lattices in industrial applications, including blending, warping, field-directed distributions, region-specific cell types, and produces arbitrary directional slicing for manufacturing. As demonstrated by various examples implemented in WebGPU, AST archives high efficiency and accuracy in real-time rendering of shell-lattices with billions of beams on an RTX 3090, outperforming traditional pipelines in storage, frame time, and detail preservation.
{"title":"Augmented Sphere Tracing for Real-time Editing Mega-scale Periodic Shell-lattice Structures","authors":"Jiajie Guo, Ming Li","doi":"10.1016/j.cad.2025.103876","DOIUrl":"10.1016/j.cad.2025.103876","url":null,"abstract":"<div><div>We propose an augmented sphere tracing (AST) pipeline that seamlessly integrates editing, rendering, and slicing of mega-scale periodic shell-lattice structures. Traditional STL-based pipelines face challenges such as time-consuming format conversions, high storage requirements, and complex blending issues between discrete lattice and shell components, often resulting in a loss of geometric accuracy. Alternatively, implicit-based pipelines excel at smooth modeling and robust Boolean operations but require inefficient and error-prone conversions of STL shells into implicit forms, complicating the rendering process. To address these issues, AST combines hybrid implicit lattice and mesh shell representations, eliminating the need for explicit 3D model construction and unnecessary geometric format conversions. It overcomes the major challenges of hybrid forms and mega-scale rendering by using an augmented tracing distance query that avoids costly signed distance field (SDF) calculations while preserving geometric details. Additionally, it employs a local tracing distance query within a single cell, leveraging lattice periodicity for efficiency. The pipeline also supports various types of shell-lattices in industrial applications, including blending, warping, field-directed distributions, region-specific cell types, and produces arbitrary directional slicing for manufacturing. As demonstrated by various examples implemented in WebGPU, AST archives high efficiency and accuracy in real-time rendering of shell-lattices with billions of beams on an RTX 3090, outperforming traditional pipelines in storage, frame time, and detail preservation.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"184 ","pages":"Article 103876"},"PeriodicalIF":3.0,"publicationDate":"2025-04-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776638","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-23DOI: 10.1016/j.cad.2025.103877
Don Pubudu Vishwana Joseph Jayakody , Tak Yu Lau , Hyunyoung Kim , Kai Tang , Lauren E.J. Thomas-Seale
Part orientation is a crucial element that governs the impact of several manufacturing constraints in material extrusion-based additive manufacturing (AM). Although part orientation optimisation has been extensively investigated to improve the manufacturability in 2.5-axis AM configuration, its influence on material extrusion-based multi-axis AM remains underdetermined. In this paper, we propose a computational framework to find the optimal part orientation that maximises the compliance of the tool orientation vector field with respect to several constraints required for support-free multi-axis AM. By combining topological significance, mesh saliency and curvedness metrics, we introduce a new salient feature map to formulate the link between the part orientation and the tool orientation vector field compliance. Once the optimal orientation is computed, our method enables a direct computation of a compliant iso-tool orientation vector field for a set of input iso-tool path points. We demonstrate that the part orientation can indeed be changed to minimise tool angle variation whilst adhering to overhang angle constraints for a range of 3D mesh models. The effectiveness of the proposed method is validated by comparing our method with existing tool orientation vector field design methods. Our promising results reveal the potential in part orientation optimisation as a means to address manufacturing constraints in multi-axis tool path design.
{"title":"A salient vector field-driven part orientation selection for multi-axis 3D printing","authors":"Don Pubudu Vishwana Joseph Jayakody , Tak Yu Lau , Hyunyoung Kim , Kai Tang , Lauren E.J. Thomas-Seale","doi":"10.1016/j.cad.2025.103877","DOIUrl":"10.1016/j.cad.2025.103877","url":null,"abstract":"<div><div>Part orientation is a crucial element that governs the impact of several manufacturing constraints in material extrusion-based additive manufacturing (AM). Although part orientation optimisation has been extensively investigated to improve the manufacturability in 2.5-axis AM configuration, its influence on material extrusion-based multi-axis AM remains underdetermined. In this paper, we propose a computational framework to find the optimal part orientation that maximises the compliance of the tool orientation vector field with respect to several constraints required for support-free multi-axis AM. By combining topological significance, mesh saliency and curvedness metrics, we introduce a new salient feature map to formulate the link between the part orientation and the tool orientation vector field compliance. Once the optimal orientation is computed, our method enables a direct computation of a compliant iso-tool orientation vector field for a set of input iso-tool path points. We demonstrate that the part orientation can indeed be changed to minimise tool angle variation whilst adhering to overhang angle constraints for a range of 3D mesh models. The effectiveness of the proposed method is validated by comparing our method with existing tool orientation vector field design methods. Our promising results reveal the potential in part orientation optimisation as a means to address manufacturing constraints in multi-axis tool path design.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"184 ","pages":"Article 103877"},"PeriodicalIF":3.0,"publicationDate":"2025-03-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143734738","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-03-12DOI: 10.1016/j.cad.2025.103864
Haroon Ijaz, Xuwei Wang, Wei Chen, Hai Lin, Ming Li
We propose a novel approach to structure-aware topology optimization (SATO) to generate physically plausible multi-component structures with diverse stylistic variations. Traditional TO methods often operate within a discrete voxel-defined design space, overlooking the underlying structure-aware, which limits their ability to accommodate stylistic design preferences. Our approach leverages variational autoencoders (VAEs) to encode both geometries and corresponding structures into a unified latent space, capturing part arrangement features. The design target is carefully formulated as a topology optimization problem taking the VAE code as design variables under physical constraints, and solved numerically via analyzing the associated sensitivity with respect to the VAE variables. Our numerical examples demonstrate the ability to generate lightweight structures that balance geometric plausibility and structural performance with much enhanced stiffness that outperforms existing generative techniques. The method also enables the generation of diverse and reliable designs, maintaining structural integrity throughout, via a direct smooth interpolation between the optimized designs. The findings highlight the potential of our approach to bridge the gap between generative design and physics-based optimization by incorporating deep learning techniques.
{"title":"Physically reliable 3D styled shape generation via structure-aware topology optimization in unified latent space","authors":"Haroon Ijaz, Xuwei Wang, Wei Chen, Hai Lin, Ming Li","doi":"10.1016/j.cad.2025.103864","DOIUrl":"10.1016/j.cad.2025.103864","url":null,"abstract":"<div><div>We propose a novel approach to structure-aware topology optimization (SATO) to generate physically plausible multi-component structures with diverse stylistic variations. Traditional TO methods often operate within a discrete voxel-defined design space, overlooking the underlying structure-aware, which limits their ability to accommodate stylistic design preferences. Our approach leverages variational autoencoders (VAEs) to encode both geometries and corresponding structures into a unified latent space, capturing part arrangement features. The design target is carefully formulated as a topology optimization problem taking the VAE code as design variables under physical constraints, and solved numerically via analyzing the associated sensitivity with respect to the VAE variables. Our numerical examples demonstrate the ability to generate lightweight structures that balance geometric plausibility and structural performance with much enhanced stiffness that outperforms existing generative techniques. The method also enables the generation of diverse and reliable designs, maintaining structural integrity throughout, via a direct smooth interpolation between the optimized designs. The findings highlight the potential of our approach to bridge the gap between generative design and physics-based optimization by incorporating deep learning techniques.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"183 ","pages":"Article 103864"},"PeriodicalIF":3.0,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143628570","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号