Weijuan Cao

Hangzhou Dianzi University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Trends

Author Order

Document Type

Co-Authors

Wanbin Pan

University of Science and Technology of China

Andrew Colligan

University of York

Hua Yang

Health Economics and Outcomes Research (United Kingdom)

Shuming Gao

Zhejiang University

Trevor Robinson

Queen's University Belfast

Jieyi Zhao

Huazhong University of Science and Technology

Shufang Wang

Yunnan Agricultural University

Yuqin Jiang

University of Electronic Science and Technology of China

Zhenlei Weng

Hangzhou Dianzi University

Declan Nolan

Queen's University Belfast

Cooperative Institutions

Zhejiang University

Hangzhou Dianzi University

Queen's University Belfast

National University of Singapore

University of Electronic Science and Technology of China

Queen's University

Beihang University

Huazhong University of Science and Technology

Sichuan University

West China Hospital of Sichuan University

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

A real-time approach for failure detection in material extrusion process based on artificial neural network

Rapid Prototyping Journal (2023)

Wanbin Pan Hongyi Jiang Shufang Wang Wen Feng Lu Weijuan Cao

Purpose This paper aims to detect the printing failures (such as warpage and collapse) in material extrusion (MEX) process effectively and timely to reduce the waste of printing time, energy and material. Design/methodology/approach The approach is designed based on the frequently observed fact that printing failures are accompanied by abnormal material phenomena occurring close to the nozzle. To effectively and timely capture the phenomena near the nozzle, a camera is delicately installed on a typical MEX printer. Then, aided by the captured phenomena (images), a smart printing failure predictor is built based on the artificial neural network (ANN). Finally, based on the predictor, the printing failures, as well as their types, can be effectively detected from the images captured by the camera in real-time. Findings Experiments show that printing failures can be detected timely with an accuracy of more than 98% on average. Comparisons in methodology demonstrate that this approach has advantages in real-time printing failure detection in MEX. Originality/value A novel real-time approach for failure detection is proposed based on ANN. The following characteristics make the approach have a great potential to be implemented easily and widely: (1) the scheme designed to capture the phenomena near the nozzle is simple, low-cost, and effective; and (2) the predictor can be conveniently extended to detect more types of failures by using more abnormal material phenomena that are occurring close to the nozzle.

10.1108/rpj-03-2022-0072

Cite

Citations (0)

DCCD: Distributed N-Body Rigid Continuous Collision Detection for Large-Scale Virtual Environments

Arabian Journal for Science and Engineering (2017)

Peng Du Jieyi Zhao Weijuan Cao Yigang Wang

10.1007/s13369-016-2411-0

Cite

Citations (3)

Graph Representation of 3D CAD Models for Machining Feature Recognition With Deep Learning

Weijuan Cao Trevor Robinson Hua Yang Flavien Boussuge Andrew Colligan

Abstract In this paper, the application of deep learning methods to the task of machining feature recognition in CAD models is studied. Four contributions are made: 1. An automatic method to generate large datasets of 3D CAD models is proposed, where each model contains multiple machining features with face labels. 2. A concise and informative graph representation for 3D CAD models is presented. This is shown to be applicable to graph neural networks. 3. The graph representation is compared with voxels on their performance of training deep neural networks to segment 3D CAD models. 4. Experiments are also conducted to evaluate the effectiveness of graph-based deep learning for interacting feature recognition. Results show that the proposed graph representation is a more efficient representation of 3D CAD models than voxels for deep learning. It is also shown that graph neural networks can be used to recognize individual features on the model and also identify complex interacting features.

Feature Learning

Representation

Feature (linguistics)

Solid modeling

Feature recognition

10.1115/detc2020-22355

Cite

Citations (36)

Hierarchical CADNet: Learning from B-Reps for Machining Feature Recognition

Computer-Aided Design (2022)

Andrew Colligan Trevor Robinson Declan Nolan Hua Yang Weijuan Cao

Deep learning approaches have been shown to be capable of recognizing shape features (e.g. machining features) in Computer-Aided Design (CAD) models in certain circumstances, yet still have issues when the features intersect, and in exploiting the geometric and topological information which comprises the boundary representation (B-Rep) of the typical CAD model. This paper presents a novel hierarchical B-Rep graph shape representation which encodes information about the surface geometry and face topology of the B-Rep. To learn from this new shape representation, a novel hierarchical graph convolutional network called Hierarchical CADNet has been created, which has been shown to outperform other state-of-the-art neural architectures on feature identification, including machining features that intersect, with improvements in accuracy for some more complex CAD models.

Boundary representation

Feature recognition

Representation

Feature (linguistics)

Feature Learning

10.1016/j.cad.2022.103226

Cite

Citations (44)

Optimal Rotational Symmetry Cell Mesh Construction for FE Analysis by Symmetry-constrained Local Delaunay Refinement

Computer-Aided Design and Applications (2013)

Weijuan Cao Ming Li Shuming Gao

With the development of Rapid manufacturing techniques, a heterogeneous object can be manufactured by processing geometric and material information in each layer while slicing the object.The functionality and surface finish of heterogeneous object depends upon precise slicing and accurate information processing i.e. geometric and material during fabrication.The current work focuses on both the issues.Layer thickness is varied using adaptive thickness method.Horizontal and vertical contours are approximated to reduce the error and minimize the effect of geometry and material stair-step effects.Reference based boundary mesh generation approach is used for defining material composition at each point in the slice domain.An algorithm is proposed for scanning layers to retrieve geometric and material information at each point in the object domain.A database system is developed for faster processing of information and to avoid the data redundancy during operation.Reference based approach, proposed algorithm and adaptive thickness method are implemented and some cases are discussed to show the validity of the work.

Local symmetry

10.1080/16864360.2014.863505

Cite

Citations (3)

Representation and Automated Generation of Analysis Feature Model for Finite Element Analysis

Weijuan Cao Haipang Wu Yuqin Jiang Yusheng Liu Shuming Gao

Integration of finite element analysis (FEA) into design is important for complex product development. However, how to automatically and robustly generate the analysis model from the design model, and how to organize information needed by CAD and FEA for the efficient integration is still a big problem. In this study, analysis feature model (AFM) is proposed and developed as a central model for design and analysis. Five kinds of information are contained in AFM which are analysis-related information, geometric-related information, element-related information, boundary condition and coupling information between features. A systematic approach is elaborated to automatically generate the AFM from the design model which mainly contains four steps: automated recognition of analysis features, decomposition of the design model, reduction and combination into AFM. The analysis model can be easily obtained from the AFM. At last the proposed method is implemented and some examples are given.

Representation

Feature (linguistics)

Boundary representation

Information model

10.1115/detc2009-86322

Cite

Citations (1)

An Approach to Automated Conversion From Design Feature Model to Analysis Feature Model

Weijuan Cao Xiaosheng Chen Shuming Gao

An approach to automatically converting a design feature model to an analysis feature model for downstream finite element analysis is proposed. The analysis feature model is a mixed-dimensional model with shell features representing thin regions and solid features representing thick regions. In the approach, the design feature model is first decomposed into a set of remnants of additive features, each of which represents part of an additive feature’s volume that remain in the final volume of the design model. The remnant of each additive feature is then decomposed into swept bodies and non-swept bodies. After that, the thin regions of each swept body are effectively recognized based on its sketch information. These detected thin regions can be wrongly recognized, which, together with potentially missing ones are thus detected and corrected by a synthesization process. Finally, the analysis features and their relative interfaces are generated, which ultimately gives the corresponding analysis feature model to the input design feature model. Experimental results are also shown to demonstrate the proposed method’s effectiveness.

Feature (linguistics)

Sketch

Feature model

Feature recognition

10.1115/detc2011-47555

Cite

Citations (5)