Thin-walled plastic injection parts are common components in many consumer products. They areusually characterized by thin shell on the main body and with many functional geometric features inside.One inherent problem of such a part in injection molding is that the outer surface tends to be shrunk locallyowing to uneven thickness at the junction of the features and the main body. Mold flow analysis has beenemployed in injection molding for the product design, the mold design and manufacturing parameterssetting, so as to improve the quality of the injection part. In mold flow analysis, it is well known that ahybrid combination of hexahedral and prism meshes are preferable to tetrahedral meshes because they arecharacterized by higher accuracy, convergence, and application specificity. However, they are inherentlymore difficult to generate because it requires the decomposition of the CAD model carefully, which isusually complex and difficult even for well-trained CAE engineers. Feature recognition has been studiedfor decades and several feature recognition methods are available on commercial CAD systems. Substantialefforts have been made on automatic decomposition and meshing, but, none of them is currently reliableenough for industrial applications because real CAD models are more complex and variable.To provide a better quality of meshes for thin-walled injection parts, main features on the part shouldbe recognized and decomposed first. Since the inner face of a thin-walled part is usually composed of manytypes of features, it is necessary to develop specific recognition and decomposition algorithms, and thenapply appropriate meshing method for each of them. The most critical issue in mesh generation is to plannodes on all edges of a face. When a feature or a region is extracted from a model, the nodes at thetransition must be consistent so that all meshes generated can be connected correctly. A gradual change ofthe density of nodes near the transition region might be necessary to maintain the connection of all meshes.This means that the planning of feature recognition and decomposition should be consideredsimultaneously, and the data used for meshing should be analyzed and computed.The purpose of this project is to develop algorithms for the recognition and decomposition of commonfeatures on thin-walled plastic parts, including rib, tube, column, boss, extrusion, and model base, and studythe geometric and topological data that should be recorded for each of the features. A hybrid meshescomposed of prism and hexahedron are used for all features extracted. When a feature is extracted, thetransition data between this feature and the model base is also recorded, which is needed when the modelbase is meshed. The proposed method can be divided into four parts. First, development of featurerecognition algorithms: we focus on the recognition of rib, tube, boss, column, extrusion and model base.The data required for each type of features are planned and computed. Second, preprocess of featuredecomposition: we focus on the planning and computation of all slicing faces for each of the features. Thedata for each slicing face are computed. Third, development of feature decomposition algorithms: we focuson the decomposition of each feature form the model. The transition data between each feature and themodel base is also recorded. Finally, solid mesh generation: All feature regions obtained are saved as B-repmodels and exported to Moldex3D for generating solid meshes. It is used to verify the feasibility of theproposed feature recognition and decomposition method to deal with real CAD models.
|Effective start/end date||1/08/18 → 31/07/19|
UN Sustainable Development Goals
In 2015, UN member states agreed to 17 global Sustainable Development Goals (SDGs) to end poverty, protect the planet and ensure prosperity for all. This project contributes towards the following SDG(s):
- Feature recognition
- Feature decomposition
- Solid mesh generation
- Thin-walled parts
Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.