• Geometric_and_Feature_Modelling
• Current_Projects

Current Projects

Validity maintenance for freeform feature modelling

In the semantic feature modelling approach, recently developed in our group, the semantics of regular-shaped features can be defined and maintained in a model. For relatively simple features, having prismatic or cylindrical shapes, it is possible to define all kinds of properties; an example is that the entrance of a hole should be open. The approach uses constraint specification and solving techniques, and guarantees that only valid feature models can be created.

In practice, many products have freeform faces because of aerodynamic, aestetic or other requirements. Modelling systems exist with which objects can be designed that contain freeform surfaces. However, several issues are yet to be solved in order to create full-fledged freeform feature modelling systems.

The goal of this project is to develop a feature validation approach for modelling objects with freeform features. A useful set of freeform features, with their characteristic parameters and important properties, will be determined. Methods for defining such freeform features and their properties with constraints, and for maintaining the latter during the modelling process, will be developed. A prototype system with an appropriate model representation and the required constraint solving algorithms will be implemented.

This project is supported by the Netherlands Organization for Scientific Research (NWO).

Semantics of families of objects

Parametric modelling, and in particular feature modelling, is now the most important way to model the shape of 3D objects. It characterises a class of similar objects by a number of parameters; an instance of such a class can be created by specifying values for its parameters. Advantages of parametric modelling are that the shape of an object can easily be fine-tuned, and that objects that differ only in details can easily be modelled. Parametric modelling is profitably used in, for example, computer-aided design and manufacturing.

A class of objects is often called a family of objects, and an instance a member of the family. A fundamental problem in parametric modelling, which hinders general acceptance, is that there is not yet a good way to define the semantics of a family of objects, and therefore objects can wrongly be considered or excluded as a member of a family. A sound definition should exactly determine which objects belong to a family.

The goal of this project is to develop methods to precisely define the semantics of a family of objects. These methods will be based on semantic feature modelling, in which shapes with well-defined semantics and constraints are used to model objects. Methods to determine members of a family will also be developed, in particular a family membership test to check whether given parameter values specify a member of a family, and techniques to compute the ranges of the parameters that correspond to members of a family.

This project is supported by the Netherlands Organization for Scientific Research (NWO).

Freeform shape techniques for improved product design

Computer-aided creation, handling and management of three-dimensional (3D) shapes is becoming an essential technology. Key processes in product design, engineering and manufacturing, but also in the graphical and entertainment industry, medical healing and urban planning, are heavily dependent on techniques for effective and efficient control of 3D shapes. In particular the requirements on the quality of engineered products are rapidly becoming more severe, putting strict conditions on aesthetic, ergonomic and cost aspects, which are all implicitly or explicitly related to shape. From literature, but also from professional users of the relevant computer-based tools, it is known that the control of 3D shapes suffers from serious limitations.

The project addresses the following issues:

1) translation of engineering requirements into shape constraints, 2) finding appropriate procedures (algorithms) that enable computer-based variation of shape in conformity with the constraints, 3) presenting appropriate shape handles (or parameters) to the user, and 4) providing relevant feedback to the user when he/she changes a parameter.

Addressing these issues is crucial to solve a very general problem: the lack of flexibility of shape handles. Existing methods cannot solve this, because they are too much based on low-level geometric editing.

In the project we will introduce an innovative method, composed of:

1) selection of regions in 3D objects by the user, 2) shape matching and shape fitting, 3) shape recognition, introduction of parameters, and constraint maintenance, 4) techniques for merging freeform shapes into new models, and 5) shape deformation techniques, with or without topological change.

As a result, the project will deliver a methodology and a set of computer tools supporting the required shape modifications.

This is a joint project with the Faculty of Design, Engineering and Production.

This project is supported by the Technology Foundation STW.

Integration of Design and Analysis Models

Computer-aided design involves several phases, including design and analysis. For the most important type of analysis, finite element analysis, the design model has to be converted into an analysis model. The analysis model consists of a mesh with a large number of simple three-, two- or one-dimensional (3D, 2D or 1D) elements, e.g. tetrahedrons, triangles or line segments. After an analysis step, the design model may have to be modified, after which a new mesh has to be generated, for another analysis.

Because several iterations may be required before the optimal model is found, it is important to use efficient meshing methods. The goal of this project is to develop such efficient meshing methods. For cases in which a full 3D model has to be analysed, an efficient method will be developed to only locally update a 3D mesh where the design model has actually been modified, instead of regenerating the complete mesh.

For cases in which it is unfeasible to analyse a full 3D model, an idealised model, which may be 2D or 1D, can be used. Efficient methods will be developed to generate meshes for idealised models, and to associate idealised models to the design model for automatic propagation of changes in an analysis model to the design model. All methods will make use of an intermediate model, a non-manifold cellular model. The result will be an integration of design and analysis models, which will lead to a much more efficient design and analysis cycle.

This project is supported by the Netherlands Organisation for Scientific Research (NWO).

An integrated collaborative design framework for application in simulation-based ship design

The design of a naval ship is a very complex process. Many requirements from different life cycle stages must be met, and these are often dependent and possibly conflicting. Examples of aspects that must be considered are operational performance and behaviour, which are related to the capabilities of the ship to fulfill missions and tasks. These aspects must already be considered in the early design phase, which is often done with simulations.

This approach of design requires collaboration of several design disciplines, in an integrated way. The goal of this research project is therefore to develop a framework for an integrated collaborative design process for application in simulation-based ship design. In particular, attention will be paid to the representation of different design requirements, a product model that incorporates all relevant information for the different design disciplines, the relation of simulation with the product model, and mechanisms for resolving design conflicts. Such a framework can lead to much shorter design times, and much better designs, for future ships.

This project is supported by TNO Defence, Security and Safety.