Pub Date : 2024-11-01DOI: 10.1016/j.cad.2024.103821
Xinming Li, Lujie Ma, Bowen Ji, Kuan Fan, Zhengdong Huang
This paper presents a novel approach for extracting fiber paths from the optimized lamination parameters (LPs) of variable stiffness laminated shells, utilizing the framework of physics-informed neural network (PINN). In this methodology, each fiber layer is associated with a specific stream function, which is approximated by an independent neural network. The stream function is governed by a partial differential equation (PDE) derived from the fiber orientation field in the parameter space. Moreover, the isocontours of the stream function are transformed into the actual fiber paths in the physical space. To account for manufacturing constraints, Riemannian geometry serves as a computational tool to determine the intrinsic distance between adjacent fiber paths and the geodesic curvature of the isocontours. By incorporating regularization terms into the loss function based on the physical relationships, the constrained optimization problem is converted into an unconstrained one, making it more suitable for neural network training. Meanwhile, a fiber path extraction (FPE) algorithm is used to minimize the loss function at randomly sampled points through gradient descent. The numerical results suggest that the extraction of fiber paths using PINN can achieve satisfactory levels of accuracy while effectively satisfying the imposed constraints.
{"title":"Extracting fiber paths from the optimized lamination parameters of variable-stiffness laminated shells based on physic-informed neural network","authors":"Xinming Li, Lujie Ma, Bowen Ji, Kuan Fan, Zhengdong Huang","doi":"10.1016/j.cad.2024.103821","DOIUrl":"10.1016/j.cad.2024.103821","url":null,"abstract":"<div><div>This paper presents a novel approach for extracting fiber paths from the optimized lamination parameters (LPs) of variable stiffness laminated shells, utilizing the framework of physics-informed neural network (PINN). In this methodology, each fiber layer is associated with a specific stream function, which is approximated by an independent neural network. The stream function is governed by a partial differential equation (PDE) derived from the fiber orientation field in the parameter space. Moreover, the isocontours of the stream function are transformed into the actual fiber paths in the physical space. To account for manufacturing constraints, Riemannian geometry serves as a computational tool to determine the intrinsic distance between adjacent fiber paths and the geodesic curvature of the isocontours. By incorporating regularization terms into the loss function based on the physical relationships, the constrained optimization problem is converted into an unconstrained one, making it more suitable for neural network training. Meanwhile, a fiber path extraction (FPE) algorithm is used to minimize the loss function at randomly sampled points through gradient descent. The numerical results suggest that the extraction of fiber paths using PINN can achieve satisfactory levels of accuracy while effectively satisfying the imposed constraints.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"179 ","pages":"Article 103821"},"PeriodicalIF":3.0,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142657976","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 : 2024-11-01DOI: 10.1016/j.cad.2024.103813
Jianping Yang , Qiaoyun Wu , Yuan Zhang , Jiajia Dai , Jun Wang
For the highly interacting machining features, Layered Projection Decomposition Method presents inferior recognition efficiency and accuracy, due to its high-cost 3D projection and failures in determining projection faces for internal occluded faces. To address these issues, we propose a potential hybrid recognition framework. We first introduce a straightforward adjacent projection wire (APW) over UV wires, automatically restoring the full projection wires from highly interacting features. Building on APWs, an efficient hybrid boundary representation and its corresponding unambiguous primitive definitions are proposed by combining with graph-based boundary representations. Subsequently, we design an efficient primitive decomposition method by introducing primitive boundary matching to decide the initial projection faces, and introducing iterative projection boundary expansion to complete the full primitives from occluded faces. Moreover, we establish an efficient Graph Neural Network to learn the distinguishable distributions over the decomposed primitives. Specifically, an Adjacency Attention Unit is proposed to automatically perceive the influence weight of adjacent nodes, leading to more discriminative self-adaptive shape embedding for efficient primitive recognition. Finally, we summarize convenient reconstruction rules to correct the wrong predictions of feature faces with indistinguishable adjacent relationships. To evaluate the effectiveness of the proposed recognition framework, CAD models of complex aircraft structural parts are collected to present a challenging machining feature dataset. Extensive numerical experiments demonstrate that the proposed hybrid recognition framework enables significant improvements over the state-of-the-art machining feature recognition techniques.
{"title":"A Hybrid Recognition Framework for Highly Interacting Machining Features Based on Primitive Decomposition, Learning and Reconstruction","authors":"Jianping Yang , Qiaoyun Wu , Yuan Zhang , Jiajia Dai , Jun Wang","doi":"10.1016/j.cad.2024.103813","DOIUrl":"10.1016/j.cad.2024.103813","url":null,"abstract":"<div><div>For the highly interacting machining features, Layered Projection Decomposition Method presents inferior recognition efficiency and accuracy, due to its high-cost 3D projection and failures in determining projection faces for internal occluded faces. To address these issues, we propose a potential hybrid recognition framework. We first introduce a straightforward adjacent projection wire (APW) over UV wires, automatically restoring the full projection wires from highly interacting features. Building on APWs, an efficient hybrid boundary representation and its corresponding unambiguous primitive definitions are proposed by combining with graph-based boundary representations. Subsequently, we design an efficient primitive decomposition method by introducing primitive boundary matching to decide the initial projection faces, and introducing iterative projection boundary expansion to complete the full primitives from occluded faces. Moreover, we establish an efficient Graph Neural Network to learn the distinguishable distributions over the decomposed primitives. Specifically, an Adjacency Attention Unit is proposed to automatically perceive the influence weight of adjacent nodes, leading to more discriminative self-adaptive shape embedding for efficient primitive recognition. Finally, we summarize convenient reconstruction rules to correct the wrong predictions of feature faces with indistinguishable adjacent relationships. To evaluate the effectiveness of the proposed recognition framework, CAD models of complex aircraft structural parts are collected to present a challenging machining feature dataset. Extensive numerical experiments demonstrate that the proposed hybrid recognition framework enables significant improvements over the state-of-the-art machining feature recognition techniques.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"179 ","pages":"Article 103813"},"PeriodicalIF":3.0,"publicationDate":"2024-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142657977","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 : 2024-10-18DOI: 10.1016/j.cad.2024.103812
Zhiyong Su , Changchang Wang , Kun Jiang , Kai Jiang , Weiqing Li
Despite existing supervised point cloud denoising methods having made great progress, they require paired ideal noisy-clean datasets for training which is expensive and impractical in real-world applications. Moreover, they may perform the denoising process multiple times with fixed network parameters for better denoising results at test time. To address above issues, this paper proposes a self-supervised iterative training framework (SITF) for point cloud denoising, which only requires single noisy point clouds and a noise model. Given an off-the-shelf denoising network and original noisy point clouds, firstly, an intermediate noisier-noisy dataset is created by adding additional noises from the known noise model to noisy point clouds (i.e. learning targets). Secondly, after training on the noisier-noisy dataset, the denoising network is employed to denoise the original noisy point clouds to obtain the learning targets for the next iteration. The above two steps are iteratively and alternatively performed to get a better and better trained denoising network. Furthermore, to get better learning targets for the next round, this paper also proposes a novel iterative denoising network (IDN) architecture of stacked source attention denoising modules. The IDN explicitly models the iterative denoising process internally within a single network via reforming the given denoising network. Experimental results show that existing supervised networks trained through the SITF can achieve competitive denoising results and even outperform supervised networks under high noise conditions. The source code can be found at: https://github.com/VCG-NJUST/SITF.
{"title":"SITF: A Self-Supervised Iterative Training Framework for Point Cloud Denoising","authors":"Zhiyong Su , Changchang Wang , Kun Jiang , Kai Jiang , Weiqing Li","doi":"10.1016/j.cad.2024.103812","DOIUrl":"10.1016/j.cad.2024.103812","url":null,"abstract":"<div><div>Despite existing supervised point cloud denoising methods having made great progress, they require paired ideal noisy-clean datasets for training which is expensive and impractical in real-world applications. Moreover, they may perform the denoising process multiple times with fixed network parameters for better denoising results at test time. To address above issues, this paper proposes a self-supervised iterative training framework (SITF) for point cloud denoising, which only requires single noisy point clouds and a noise model. Given an off-the-shelf denoising network and original noisy point clouds, firstly, an intermediate noisier-noisy dataset is created by adding additional noises from the known noise model to noisy point clouds (i.e. learning targets). Secondly, after training on the noisier-noisy dataset, the denoising network is employed to denoise the original noisy point clouds to obtain the learning targets for the next iteration. The above two steps are iteratively and alternatively performed to get a better and better trained denoising network. Furthermore, to get better learning targets for the next round, this paper also proposes a novel iterative denoising network (IDN) architecture of stacked source attention denoising modules. The IDN explicitly models the iterative denoising process internally within a single network via reforming the given denoising network. Experimental results show that existing supervised networks trained through the SITF can achieve competitive denoising results and even outperform supervised networks under high noise conditions. The source code can be found at: <span><span>https://github.com/VCG-NJUST/SITF</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"179 ","pages":"Article 103812"},"PeriodicalIF":3.0,"publicationDate":"2024-10-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142528272","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 : 2024-10-17DOI: 10.1016/j.cad.2024.103811
Rafael Merli , Antolín Martínez-Martínez , Juan José Ródenas , Marc Bosch-Galera , Enrique Nadal
In today’s industry, the rapid evolution in the design and development of optimized mechanical components to meet customer requirements represents a significant challenge for companies. These companies seek efficient solutions to enhance their products in terms of stiffness and strength. One powerful approach is Topology Optimization, which aims to determine the optimal material distribution within a predefined domain to maximize the overall component’s stiffness. Achieving high-resolution solutions is also crucial for accurately defining the final material distribution. While standard Topology Optimization tools can propose optimal solutions for entire components, they struggle with small-scale details (such as trabecular structures) due to prohibitive computational costs. To address this issue, our proposed approach introduces a two-level topology optimization methodology considering density-based techniques. The proposed methodology includes three steps: The first one subdivides the whole component in cells and generates a coarse optimized low-definition material distribution, assigning a different density to each cell. Since the output stresses from the coarse problem are not equilibrated into each cell, they must not be directly used in the fine level. Thus, the second step uses the equilibrating traction recovery approach to convert the cell nodal forces into equilibrated lateral tractions over the cell boundary. Finally, taking as input data the densities from the coarse optimization and imposing these lateral tractions as Neumann boundary conditions, each cell is optimized at fine level. The main goal of this work is to efficiently solve high-resolution topology optimization problems using a two-level mechanically-continuous method, which would be unaffordable with standard computing facilities and the current techniques.
{"title":"Two-Level High-Resolution Structural Topology Optimization with Equilibrated Cells","authors":"Rafael Merli , Antolín Martínez-Martínez , Juan José Ródenas , Marc Bosch-Galera , Enrique Nadal","doi":"10.1016/j.cad.2024.103811","DOIUrl":"10.1016/j.cad.2024.103811","url":null,"abstract":"<div><div>In today’s industry, the rapid evolution in the design and development of optimized mechanical components to meet customer requirements represents a significant challenge for companies. These companies seek efficient solutions to enhance their products in terms of stiffness and strength. One powerful approach is Topology Optimization, which aims to determine the optimal material distribution within a predefined domain to maximize the overall component’s stiffness. Achieving high-resolution solutions is also crucial for accurately defining the final material distribution. While standard Topology Optimization tools can propose optimal solutions for entire components, they struggle with small-scale details (such as trabecular structures) due to prohibitive computational costs. To address this issue, our proposed approach introduces a two-level topology optimization methodology considering density-based techniques. The proposed methodology includes three steps: The first one subdivides the whole component in cells and generates a coarse optimized low-definition material distribution, assigning a different density to each cell. Since the output stresses from the coarse problem are not equilibrated into each cell, they must not be directly used in the fine level. Thus, the second step uses the equilibrating traction recovery approach to convert the cell nodal forces into equilibrated lateral tractions over the cell boundary. Finally, taking as input data the densities from the coarse optimization and imposing these lateral tractions as Neumann boundary conditions, each cell is optimized at fine level. The main goal of this work is to efficiently solve high-resolution topology optimization problems using a two-level mechanically-continuous method, which would be unaffordable with standard computing facilities and the current techniques.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"179 ","pages":"Article 103811"},"PeriodicalIF":3.0,"publicationDate":"2024-10-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142528347","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 : 2024-10-10DOI: 10.1016/j.cad.2024.103809
Qiang Zou, Lizhen Zhu, Jiayu Wu, Zhijie Yang
This paper presents a learning-based method to solve the traditional parameterization and knot placement problems in B-spline approximation. Different from conventional heuristic methods or recent AI-based methods, the proposed method does not assume ordered or fixed-size data points as input. There is also no need for manually setting the number of knots. Parameters and knots are generated in an associative way to attain better parameter-knot alignment, and therefore a higher approximation accuracy. These features are attained by using a new generative model SplineGen, which casts the parameterization and knot placement problems as a sequence-to-sequence translation problem. It first adopts a shared autoencoder model to learn a 512-D embedding for each input point, which has the local neighborhood information implicitly captured. Then these embeddings are autoregressively decoded into parameters and knots by two associative decoders, a generative process automatically determining the number of knots, their placement, parameter values, and their ordering. The two decoders are made to work in a coordinated manner by a new network module called internal cross-attention. Once trained, SplineGen demonstrates a notable improvement over existing methods, with one to two orders of magnitude increase in approximation accuracy on test data.
{"title":"SplineGen: Approximating unorganized points through generative AI","authors":"Qiang Zou, Lizhen Zhu, Jiayu Wu, Zhijie Yang","doi":"10.1016/j.cad.2024.103809","DOIUrl":"10.1016/j.cad.2024.103809","url":null,"abstract":"<div><div>This paper presents a learning-based method to solve the traditional parameterization and knot placement problems in B-spline approximation. Different from conventional heuristic methods or recent AI-based methods, the proposed method does not assume ordered or fixed-size data points as input. There is also no need for manually setting the number of knots. Parameters and knots are generated in an associative way to attain better parameter-knot alignment, and therefore a higher approximation accuracy. These features are attained by using a new generative model SplineGen, which casts the parameterization and knot placement problems as a sequence-to-sequence translation problem. It first adopts a shared autoencoder model to learn a 512-D embedding for each input point, which has the local neighborhood information implicitly captured. Then these embeddings are autoregressively decoded into parameters and knots by two associative decoders, a generative process automatically determining the number of knots, their placement, parameter values, and their ordering. The two decoders are made to work in a coordinated manner by a new network module called internal cross-attention. Once trained, SplineGen demonstrates a notable improvement over existing methods, with one to two orders of magnitude increase in approximation accuracy on test data.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103809"},"PeriodicalIF":3.0,"publicationDate":"2024-10-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142433629","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 : 2024-10-01DOI: 10.1016/j.cad.2024.103810
Kaixin Yu , Bohan Wang , Xuejuan Chen , Ying He , Jianjun Chen
This paper presents a novel method for generating higher-order meshes for CAD surfaces by leveraging minimal surface theory to improve element shapes. We explore the concept of higher-order mesh distortion through deformation gradients and introduce an energy function designed to minimize the surface area of these meshes, providing a theoretical justification for its effectiveness in untangling. The process of mesh generation starts with segmenting CAD surfaces into linear elements, followed by the insertion of higher-order nodes within these elements. These nodes are then projected onto the CAD surface to form the initial higher-order elements. By optimizing energy functions related to minimal surfaces and the projection distances, we achieve high-quality, geometrically accurate higher-order surface meshes. Our method has been validated on complex geometries, showcasing its potential in creating effective higher-order meshes for industrial CAD models.
{"title":"Minimal surface-guided higher-order mesh generation for CAD models","authors":"Kaixin Yu , Bohan Wang , Xuejuan Chen , Ying He , Jianjun Chen","doi":"10.1016/j.cad.2024.103810","DOIUrl":"10.1016/j.cad.2024.103810","url":null,"abstract":"<div><div>This paper presents a novel method for generating higher-order meshes for CAD surfaces by leveraging minimal surface theory to improve element shapes. We explore the concept of higher-order mesh distortion through deformation gradients and introduce an energy function designed to minimize the surface area of these meshes, providing a theoretical justification for its effectiveness in untangling. The process of mesh generation starts with segmenting CAD surfaces into linear elements, followed by the insertion of higher-order nodes within these elements. These nodes are then projected onto the CAD surface to form the initial higher-order elements. By optimizing energy functions related to minimal surfaces and the projection distances, we achieve high-quality, geometrically accurate higher-order surface meshes. Our method has been validated on complex geometries, showcasing its potential in creating effective higher-order meshes for industrial CAD models.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103810"},"PeriodicalIF":3.0,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142427578","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 : 2024-09-27DOI: 10.1016/j.cad.2024.103808
Puhao Lei , Zhen Chen , Runli Tao , Jun Li , Yuchi Hao
A vision-based boundary detector is crucial for intelligent processing of ship planar components due to its automatically identifying workpiece edges. However, traditional methods suffer from many issues such as low accuracy and excessive detection errors for these workpieces with complex shape profiles. This paper proposes a trimmed Delaunay triangulation method (TDT) for recognizing boundary edges of planar workpieces from point clouds. It begins by distinguishing the difference of binary image pixel generated from point cloud to eliminate redundant points far away from plane boundary. Then, a triangulation trimming algorithm is developed to extract the edge points from the simplified points. Finally, complete plane boundary is reconstructed by a clustering-and-fitting method from the extracted edge points. Experimental results from multiple angles show that average absolute errors of straight edges and angles recognition are 1.29 mm and 1.04° respectively, which demonstrate that TDT has a high identification accuracy and robustness of plane boundary edge.
基于视觉的边界检测器可自动识别工件边缘,对船舶平面部件的智能加工至关重要。然而,传统的方法存在许多问题,例如精度低、检测误差过大,无法识别形状复杂的工件。本文提出了一种修剪德劳内三角测量法(TDT),用于从点云中识别平面工件的边界边缘。该方法首先区分由点云生成的二值图像像素的差异,以消除远离平面边界的冗余点。然后,开发一种三角形修剪算法,从简化点中提取边缘点。最后,通过聚类和拟合方法从提取的边缘点重建完整的平面边界。多角度的实验结果表明,直线边缘和角度识别的平均绝对误差分别为 1.29 mm 和 1.04°,这表明 TDT 对平面边界边缘具有较高的识别精度和鲁棒性。
{"title":"Boundary recognition of ship planar components from point clouds based on trimmed delaunay triangulation","authors":"Puhao Lei , Zhen Chen , Runli Tao , Jun Li , Yuchi Hao","doi":"10.1016/j.cad.2024.103808","DOIUrl":"10.1016/j.cad.2024.103808","url":null,"abstract":"<div><div>A vision-based boundary detector is crucial for intelligent processing of ship planar components due to its automatically identifying workpiece edges. However, traditional methods suffer from many issues such as low accuracy and excessive detection errors for these workpieces with complex shape profiles. This paper proposes a trimmed Delaunay triangulation method (TDT) for recognizing boundary edges of planar workpieces from point clouds. It begins by distinguishing the difference of binary image pixel generated from point cloud to eliminate redundant points far away from plane boundary. Then, a triangulation trimming algorithm is developed to extract the edge points from the simplified points. Finally, complete plane boundary is reconstructed by a clustering-and-fitting method from the extracted edge points. Experimental results from multiple angles show that average absolute errors of straight edges and angles recognition are 1.29 mm and 1.04° respectively, which demonstrate that TDT has a high identification accuracy and robustness of plane boundary edge.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103808"},"PeriodicalIF":3.0,"publicationDate":"2024-09-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142427577","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 : 2024-09-24DOI: 10.1016/j.cad.2024.103806
Chao Zhang , Arnaud Polette , Romain Pinquié , Gregorio Carasi , Henri De Charnace , Jean-Philippe Pernot
This paper introduces a novel framework capable of reconstructing editable parametric CAD models from dumb B-Rep models. First, each B-Rep model is represented with a network-friendly formalism based on UV-graph, which is then used as input of eCAD-Net, the new deep neural network-based algorithm that predicts feature-based CAD modeling sequences from the graph. Then, the sequences are scaled and fine-tuned using a feature matching algorithm that retrieves the exact parameter values from the input dumb CAD model. The output sequences are then converted in a series of CAD modeling operations to create an editable parametric CAD model in any CAD modeler. A cleaned dataset is used to learn and validate the proposed approach, and is provided with the article. The experimental results show that our approach outperforms existing methods on such reconstruction tasks, and it outputs editable parametric CAD models compatible with existing CAD modelers and ready for use in downstream engineering applications.
{"title":"eCAD-Net: Editable Parametric CAD Models Reconstruction from Dumb B-Rep Models Using Deep Neural Networks","authors":"Chao Zhang , Arnaud Polette , Romain Pinquié , Gregorio Carasi , Henri De Charnace , Jean-Philippe Pernot","doi":"10.1016/j.cad.2024.103806","DOIUrl":"10.1016/j.cad.2024.103806","url":null,"abstract":"<div><div>This paper introduces a novel framework capable of reconstructing editable parametric CAD models from dumb B-Rep models. First, each B-Rep model is represented with a network-friendly formalism based on UV-graph, which is then used as input of eCAD-Net, the new deep neural network-based algorithm that predicts feature-based CAD modeling sequences from the graph. Then, the sequences are scaled and fine-tuned using a feature matching algorithm that retrieves the exact parameter values from the input dumb CAD model. The output sequences are then converted in a series of CAD modeling operations to create an editable parametric CAD model in any CAD modeler. A cleaned dataset is used to learn and validate the proposed approach, and is provided with the article. The experimental results show that our approach outperforms existing methods on such reconstruction tasks, and it outputs editable parametric CAD models compatible with existing CAD modelers and ready for use in downstream engineering applications.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103806"},"PeriodicalIF":3.0,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142322775","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 : 2024-09-24DOI: 10.1016/j.cad.2024.103807
Nafiseh Niknejadi, Bijan Boroomand
This paper introduces an efficient quadrature rule for domains with curved boundaries in 2D/3D. Building upon our previous work focused on polytopes (Comput. Methods Appl. Mech. Engrg. 403 (2023) 115,726), we extend this method to handle volume/boundary integration on domains with general configurations and boundaries. In this method, we approximate a generic function using a finite number of orthogonal polynomials, and we obtain the coefficients of these polynomials through the integration points. The physical domain is enclosed by a fictitious rectangular/cuboidal domain, where a tensor-product of Gauss quadrature points is primarily considered. To locate the integration points that are strictly within the domain under consideration (e.g., the physical 3D domain itself or its mapped boundaries), we form a system of algebraic equations whose dimensions depend solely on the number of polynomials, not the number of quadrature points which may be significantly larger. This allows us to construct a full-rank square coefficient matrix, leading to the uniqueness of the solution, and the system of equations is then solved through a straightforward inverse process. To evaluate the integral of the polynomials, we transform the integration over the domain under consideration into an equivalent integration along the domain's boundaries using the divergence theorem. For 2D cases, we perform the boundary integration using Gauss points along the curved lines. In 3D cases, we provide an efficient algorithm for computing the boundary integrals over curved surfaces. We present several integration problems involving two and three-dimensional curved regions to demonstrate the accuracy and efficiency of the proposed method.
{"title":"Numerical integration on 2D/3D arbitrary domains: Adaptive quadrature/cubature rule for domains with curved boundaries","authors":"Nafiseh Niknejadi, Bijan Boroomand","doi":"10.1016/j.cad.2024.103807","DOIUrl":"10.1016/j.cad.2024.103807","url":null,"abstract":"<div><div>This paper introduces an efficient quadrature rule for domains with curved boundaries in 2D/3D. Building upon our previous work focused on polytopes (Comput. Methods Appl. Mech. Engrg. 403 (2023) 115,726), we extend this method to handle volume/boundary integration on domains with general configurations and boundaries. In this method, we approximate a generic function using a finite number of orthogonal polynomials, and we obtain the coefficients of these polynomials through the integration points. The physical domain is enclosed by a fictitious rectangular/cuboidal domain, where a tensor-product of Gauss quadrature points is primarily considered. To locate the integration points that are strictly within the domain under consideration (e.g., the physical 3D domain itself or its mapped boundaries), we form a system of algebraic equations whose dimensions depend solely on the number of polynomials, not the number of quadrature points which may be significantly larger. This allows us to construct a full-rank square coefficient matrix, leading to the uniqueness of the solution, and the system of equations is then solved through a straightforward inverse process. To evaluate the integral of the polynomials, we transform the integration over the domain under consideration into an equivalent integration along the domain's boundaries using the divergence theorem. For 2D cases, we perform the boundary integration using Gauss points along the curved lines. In 3D cases, we provide an efficient algorithm for computing the boundary integrals over curved surfaces. We present several integration problems involving two and three-dimensional curved regions to demonstrate the accuracy and efficiency of the proposed method.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103807"},"PeriodicalIF":3.0,"publicationDate":"2024-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142432808","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 : 2024-09-21DOI: 10.1016/j.cad.2024.103805
Lucas Vergez, Arnaud Polette, Jean-Philippe Pernot
This paper presents a machine learning-based approach to predict kinematic constraints between CAD models that have potentially never been assembled together before. During the learning phase, the algorithm is trained to predict the next-possible-constraints between a set of parts candidate to the assembly. Assemblies are represented in a new graph-based formalism that is capable of capturing features associated with parts, interfaces between parts and constraints between them. Using such a multi-level feature extraction strategy coupled to a state-by-state graph decomposition, the approach does not need to be trained on a large database. This formalism is used to model both the network input and output where the next-possible-constraints appear after evaluation. The core of the approach relies on a series of networks based on a link-prediction encoder–decoder architecture, integrating the capabilities of several convolutional networks trained in an end-to-end manner. A decision-making algorithm is added to post-process the output and drive the prediction process in finding one among the set of next-possible-constraints. This process is repeated until no more constraints can be added. The experimental results show that the proposed approach outperforms state-of-the-art methods on such assembly tasks. Although the state-by-state assembly algorithm is iterative, it still takes into account the whole set of parts as well as the whole set of constraints already predicted, and this makes it possible to handle constraint cycles, which is generally not possible when not considering multiple parts as input.
{"title":"Multi-part kinematic constraint prediction for automatic generation of CAD model assemblies using graph convolutional networks","authors":"Lucas Vergez, Arnaud Polette, Jean-Philippe Pernot","doi":"10.1016/j.cad.2024.103805","DOIUrl":"10.1016/j.cad.2024.103805","url":null,"abstract":"<div><div>This paper presents a machine learning-based approach to predict kinematic constraints between CAD models that have potentially never been assembled together before. During the learning phase, the algorithm is trained to predict the next-possible-constraints between a set of parts candidate to the assembly. Assemblies are represented in a new graph-based formalism that is capable of capturing features associated with parts, interfaces between parts and constraints between them. Using such a multi-level feature extraction strategy coupled to a state-by-state graph decomposition, the approach does not need to be trained on a large database. This formalism is used to model both the network input and output where the next-possible-constraints appear after evaluation. The core of the approach relies on a series of networks based on a link-prediction encoder–decoder architecture, integrating the capabilities of several convolutional networks trained in an end-to-end manner. A decision-making algorithm is added to post-process the output and drive the prediction process in finding one among the set of next-possible-constraints. This process is repeated until no more constraints can be added. The experimental results show that the proposed approach outperforms state-of-the-art methods on such assembly tasks. Although the state-by-state assembly algorithm is iterative, it still takes into account the whole set of parts as well as the whole set of constraints already predicted, and this makes it possible to handle constraint cycles, which is generally not possible when not considering multiple parts as input.</div></div>","PeriodicalId":50632,"journal":{"name":"Computer-Aided Design","volume":"178 ","pages":"Article 103805"},"PeriodicalIF":3.0,"publicationDate":"2024-09-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142318590","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}