Printer Friendly

Controlling die configuration through software.

Inside This Story:

* Die configuration is arguably one of the most important aspects of diecasting. However, it also is one of the hardest parts of the process to automate.

* Detailed within is a dexel-based software program that has helped automate the process and produce more accurate models.

Decisions made early in die casting design have an immense impact on the final product. Of everything that goes into the design of a diecast component, die configuration is arguably the most crucial. These decisions (parting direction, parting line location, core/slides, etc.) have a pro found impact on quality, manufacturing cost, operational costs and final component design.

In component and early die design, it is desirable to have a computer tool to explore, visualize and evaluate different configurations quickly and easily. When it comes to computer aided design (CAD), accuracy is of the utmost importance. A 3-D representation of the component and die can play a pivotal role in simulation and answering the questions needed for the final design. However, calculation of some factors is complicated (if not impossible) to automate. This article details some guidelines of die configuration and how an automated program was developed to implement them and produce more accurate models.

Ground Rules

Many criteria have been proposed to achieve successful die configuration. Projected area, parting line, die block depth, undercuts, tolerance, heat flow, etc., must all be taken into consideration. The ultimate goal in creating a CAD tool for die configuration is to construct a program that incorporates these guidelines to accomplish the configuration task automatically. However, it is difficult to follow these rules in a computer-based tool because calculation and reasoning for factors such as flow and heat are extremely complicated to automate.

In previous attempts to solve the problem, methods to find feasible parting direction, undercut and parting line location were investigated with little success. Most of the work focused on component castability or moldability analysis. In these cases, only the component model was considered and analyzed, which is not enough for die configuration purposes. Also, most methods proposed can only be used with a polyhedral model, which assumes the component has only planar surfaces and limits their applicability to standard accurate curved surface models in commercial CAD systems. Finally, many proposed algorithms are complicated and thus difficult to implement robustly.

Approach & Overall Framework

The exception to the previous attempts is a simple dexel-based tool to assist in die configuration. The dexel model (line segment representation) enables efficient geometric reasoning and can quickly provide complete con figuration results, including not only a component model but also models of the cover die, ejector die, undercut, slide and die bases. The resulting models provide 3-D graphics with automatically calculated configuration factors and properties.

The evaluation process is easier and more direct than simply looking at a component model and attempting to imagine the corresponding die. Also, the underlying dexel-based method can be used on both polyhedral models and advanced curved surface models (though only polyhedral implementations have been developed at this time).

A prototype implementation using the method adopts the STL file of the component as the input because of its simplified triangular representation and independence of specific CAD systems. Models of the component cover and ejector dies and undercuts are automatically generated based upon the component model and user-specified candidate parting direction and parting surface information. The undercuts (if any) are then used to guide the construction of die slides. The resulting cover and ejector die models also can be treated as die cavity inserts, and die bases can be constructed by user input, important die configuration factors (undercut, parting line, projected area, etc.) are automatically calculated and can be used to evaluate different die configuration plans. The dexel models of cover die and ejector die are converted back to STL models and can be subsequently simplified and exported for visualization and other application purposes.

Dexel Models

The dexel model was first introduced by silicon Graphics Inc., Mountain View, Calif., for the purpose of Numerical Control Simulation. It represents a solid as a collection of simple line segments. The model is modified to include not only the component model, but also die components and features.

Once the component model is input, the user picks the die parting direction. Then, the component model is rotated so that the die parting direction is parallel to the Z coordinate axis. The bounding box of the component model is calculated and expanded to represent the die box, which is defined according to user-specified resolution of the dexel model and number of dexels in the die.

Subsequently, the part model is translated to the proper position so that the die box is located in the first octant of the coordinate system and the low left corner of the die box is al the origin of the coordinate system (Fig. 1).

[FIGURE 1 OMITTED]

Rays are then cast from the grid points on the XOY plane along the Z direction. These rays intersect with the part model as well as the die box. The intersection points break each ray into several line segments, which represent a dexel. The dexels are recorded in a 2D array that corresponds to the grid mesh on the XOY plane. Each cell in the 2 D array corresponds to a grid point on the mesh. Every element of the 2-D array contains a linked list, which corresponds to a ray. Each node in the list defines the start of a line segment or dexel.

Die Base Construction

The die models generated in the dexelization process can be treated as cavity inserts. Then, the bases can be used to combine with the inserts to form the die (Fig. 2). If a design has some unavoidable undercuts, the die has to be made with one or more slides to obtain the desired result without affecting ejection of the casting from the cavity. A slide consists of two parts, the slide tip and the slide carrier (Fig. 3). The tip forms a part of the cavity that involves one of more undercuts. It takes the shape of the corresponding undercut feature of the component.

[FIGURES 2-3 OMITTED]

A carrier is represented by four cross section rectangles (front rectangle, mid-1 rectangle, mid-2 rectangle and back rectangle). The front and the mid-1 rectangle have the same dimensions and the back and the mid-2 rectangle have the same dimensions. The difference in the cross-section rectangles between the front portion of the carrier and the back portion of the carrier creates the lock face. One simple method for slide construction is to treat each undercut as the tip of a slide and attach a slide carrier to the tip. However, this method is not going to capture the desired slide in some scenarios.

One example is shown in Fig. 4. In this situation, two separate undercuts that should be addressed by only one slide are identified during the dexelization process. If the two undercuts are assigned to one slide carrier to form the slide, the middle portion of the tip is missing.

[FIGURE 4 OMITTED]

To overcome this problem, a method called local dexelization is used. The only difference from the previous dexelization process is that rays are shot from the front face of the slide carrier and only one dexel is needed at each position. Once the slide carrier is designed, the front face of the carrier is examined and divided into a 2-D mesh (Fig. 5). The grid size of the mesh can be the same as the dexel size used in the previous process.

[FIGURE 5 OMITTED]

Subsequently, rays are shot from the grid points of the 2 D mesh along opposite directions of the slide draw direction. Each my terminates if it hits the part surface or reaches a pre-defined maximum length to form slide tip dexel. The maximum length is set as the longest length of the bounding box of the part. Then, all dexels are checked against the existing undercut dexels to make sure the generated slide tip marches the corresponding undercuts. Once the slide construction is finished, the slide model is subtracted from the die cavity insert models and die base models to create slide passages.

Implementation Results

Prototype software based upon the proposed methods has been implemented on the Windows platform and was used to test various die casting models. For the test, a housing part was used as the input model. After specifying the die parting direction and parting surfaces, the configuration results (including cover die, ejector die and undercut model) were automatically generated. Die bases and a slide are then constructed. Based upon the test results, the advantages of using the proposed methodology can be summarized as:

* the application of a dexel model simplifies the reasoning process because reasoning algorithms are based on the uniform dexel representation of any component;

* many factors influencing die configuration are volumetric characteristics of the part. The dexel model is a volumetric model, therefore, it explicitly describes these factors naturally;

* after the dexel transformation, a component model and dexel model is established;

* the die dexel model can be readily converted back to STL models and subsequently simplified for rendering and other applications;

* the dexel model also could be used on a curved surface model to be compatible with commercial CAD systems.

For the current implementation, STL files from commercial CAL) systems are used as the input to the tool. The STL models are triangle representations of the exact curved surface models used in the CAD system. Under certain circumstances, the approximation introduces false undercuts for the corresponding STL models. For example, in Fig. 6, the original cylindrical surface is perpendicular to the top face of the base box, but the triangles generated from the surface are not. Some triangles face upward and some face downward.

[FIGURE 6 OMITTED]

If the die parting direction is chosen as shown in the example, the space between the triangles facing downward and the top face of the base box becomes undercuts. Because these undercuts do not exist if the accurate model is considered, they are called false undercuts. In the dexelization process, these concave features introduced by the approximation are picked up as false undercut dexels. Fortunately, in most cases, false undercuts are easy to identify by comparing them against the STL model. They generally show up in a scattered manner on the surfaces parallel to the die parting direction.

To completely eliminate false undercuts, an accurate geometric model instead of an approximate model has to be used as the input to the dexelization process. This can be achieved by either building curved surface handling capability to this system, or building the dexel models on top of a commercial CAD system using its application programming interfaces.

Back In Control

This methodology for building a geometric based die configuration tool was used to construct a prototype tool that can be used to explore, visualize and evaluate different die configuration plans quickly and easily. By using this tool, complete configuration results can be generated and presented in 3-D graphics. A number of die configuration factors, as well as important proper ties of the die and component are automatically identified and/or calculated. The dexel-based software takes much of the guesswork out of the die configuration process and puts more control of an integral aspect of diecasting back into the hands of diecasters.

This article was adapted from a paper presented at the 2002 North American Die Casting Assn. Congress, Rosemont. Ill.

For More Information

"Simulation As a Tool in High-Pressure Diecasting Processes," W. Maus, H. Rockmann, R. Seefeldt, C. Kleeburg, G. Vong, D.C. Schmidt and J.F. Meredith, Metal Casting Technologies, June 2001.

"Computer Model Diecasting Shot Sleeves to Predict Distortion," J. Brevick, K. Narayan, P. Shastri and A. Jain, MODERN CASTING, March 2002, p. 34 37.

Dongtao Wang is a PhD student at the Ohio State Univ., Columbus, Ohio. His research concentration includes mathematical modeling, numerical simulation and geometric reasoning for manufacturing processes. R. Allen Miller is a professor and the chair of the Dept. of Industrial, Welding and Systems Engineering at The Ohio State Univ. Li Cai has done research involving development of tools for design and manufacturing.
COPYRIGHT 2004 American Foundry Society, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2004, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Author:Wang, Dongtao
Publication:Modern Casting
Geographic Code:1USA
Date:May 1, 2004
Words:2035
Previous Article:Discovering zinc alloys for stronger die castings.
Next Article:Exploring 6 production aspects for vigilant design.
Topics:


Related Articles
Vertical squeeze casting of aluminum components.
Why not extrusion simulation?
Plenty of extrusion news at CMM Show.
World's largest sheet line cuts costs for geomembrane maker.
Zero pressure extrusion.
Automatic gauge control comparing the options.
Flow modeling speeds processing, lowers costs.
Crossheads, tubing dies.
Extrusion Dies for Plastics And Rubber: Design and Engineering Computations 3rd Edition.
Feinfocus' WBI-Fox x-ray: dynamic void detection in wafer bumps, without gray level definition.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters