SOLIDIFICATION MODELING AND RAPID PROTOTYPING

The use of solidification simulation is widely practiced in American foundries of all sizes. A recent study indicated that approximately 30% of U.S. foundries use solidification software, and all of the automotive foundries use it (Jensen, Beckermann, and Fisher 1996). Perhaps half of the castings poured in the United States today are poured in foundries that make use of solidification simulation programs. There are over a dozen commercially available simulation programs in the United States today. Two Japanese universities (Tohoku and Osaka) are among the world leaders in the research into these models so it is logical to expect that Japanese foundries would also be among the leaders in implementing this technology. However, Japanese foundries appear to be behind U.S. practice in the application of solidification simulation programs to casting production. Of the 1500 foundries in Japan, only about 150 use solidification software to design casting gating systems.

As in other countries, the Japanese have chosen to develop their own models rather than purchase models commercially available in the West (although some large foundries have purchased European and U.S. models, it appears that they are used primarily for benchmarking). There are three Japanese models in use in that country; none are marketed in the United States.

Komatsu "SOLDIA" Model

This finite difference model has been licensed to about 100 foundries in Japan. It has the capability to make fluid flow, heat flow, and cooling rate calculations and Niyama criterion maps. Those who use the model are quite satisfied with it, but its actual extent of use is not known. (Personnel at one Komatsu foundry reported that they did not use it to perform fluid flow or cooling rate calculations while a foundry at another company reported using it successfully for those applications.) One automobile company transfers data files electronically to its foundries, which are expected to use them in this program; at least one supplier foundry does so although another does not.

There is no English language version of the program, and it has not been discussed in international conferences on solidification simulation. (There have been seven such conferences in the last 16 years.)

Hitachi "S-Cast" Program

This was one of the earliest programs developed in Japan. It is FDM-based and has been used for both metal and sand molds. It uses the same mesh to do metal flow, solidification, and structural analysis in terms of stress, deformation, and estimated life. However, there has been no report on its use or improvements for a decade, and its current utility is unknown. There is no English language version available.

"JS-CAST"

One of the reasons that few Japanese foundries use solidification simulation programs is their expense. The average cost of these programs in Japan is about ¥100 million, two to three times their cost in the United States. A recent survey of Japanese foundrymen indicated that they would be willing to pay ¥30 million for a solidification simulation program. Komatsu and Osaka University are developing such a model, called "JS-CAST," which is a modification of SOLDIA. This work is supported by MITI's Small and Medium Enterprise Agency. (Of the 1500 foundries in Japan, 98% are classified as "small" or "medium" enterprises, meaning that they employ fewer than 300 or have a capitalization of less than ¥1 billion.)

JS-CAST is finite-difference based, has an open architecture, will accept CAD files, and has been designed with ease of use in the small foundry in mind. It can run on work stations or PCs using a keyboard or mouse. It is available to Japanese foundries for the target cost of ¥30 million. One limitation is that it does not yet calculate radiation view factors so it is not suitable for use in investment foundries.

Tohoku University's "STEFAN" Model

This model is under development by Prof. Niyama, co-developer of S-Cast when he was at Hitachi. The program does fluid flow and heat flow quite impressively. No micromodeling or distortion code is included, and no use is made of expert systems to shorten development time. The screen prompts are in English. The program has not been benchmarked against foreign commercial codes. Prof. Niyama has demonstrated this model at a number of conferences.

The model is under development at Tohoku University, with funding from a consortium of 30 foundries, half of them die casters. The graduate students who work on the model are Japanese, Chinese, Korean and Thai, which is why the prompts are in English. (Asian regional technical meetings often are carried on in English as a way to avoid the problem of deciding which Asian language should dominate the proceedings.) Only consortium members had access to the model at the time of the WTEC visit, and there were no plans to commercialize it.

Other Uses of Modeling in Japan

Honda has used the Western code "Flow 3D" to model filling in die casting dies with success. Researchers at Nagoya University, who are active in the development of methods of using electromagnetic fields to remove inclusions from molten metal, have also developed models to help them in this research. Toyota also has investigated the use of models for its foundry.

Morikawa representatives stated that they do not use models today for the Morikawa lost foam process because no current model adequately models this process. WTEC's hosts at Daido Precision Parts indicated that they do not use modeling because the fluid flow capabilities of current models are not sufficient to permit them to predict flow in the thin sections of the turbocharger wheels they make. One small foundry reported that it does not use models because of the expense of hiring an engineer to run the models, as well as the expense of purchasing the software and hardware required.

Although Japanese researchers are in the forefront of development of solidification models, application in their foundry industry lags behind that in the United States. Many Japanese foundrymen consider models to be too expensive and believe that it would be difficult to hire engineers to run the models. Models are in use in large foundries but in few of the smaller ones. There is little interest in models that can predict microstructure or casting distortion, and we saw no evidence of the use of expert systems to aid the foundry engineer or of models of core injection or curing (which are under development in the United States).

Observations on Modeling in Europe

Solidification simulation is highly developed in Europe. Models have been developed in England, France, Switzerland, Germany, and the Scandinavian countries; one German model (MagmasoftTM) is commercially available worldwide and is highly regarded by many foundries. A second European model "SIMULOR," developed by Pechiney, is also in use in Europe, and some copies have been sold in the United States. SIMULOR is noted for its ease of use. Both Magmasoft and SIMULOR predict mold filling and solidification patterns for castings.

European university researchers are actively involved in the field. Their progress in modeling is similar to that of the United States; however, there appears to be more interest and work in the area of residual stress prediction in castings in Europe at the present time.


Published: March 1997; WTEC Hyper-Librarian