1999 casting simulation software survey.

An updated survey of available solidification modeling packages illustrates how technology is advancing to ensure sound, cost-effective castings.

The purpose of simulating any industrial process is to model the underlying physics so that important process variables can be identified and controlled, resulting in significant benefits.

If the process of filling and solidifying a mold cavity is accurately modeled, shrinkage cavities and other casting defects can be predicted; the effects on metal fluid flow and solidification of changing the gating/risering method or many other process variable can be simulated; and molten metal trial iterations can be significantly reduced and used primarily for concept validation.

There are three basic types of computer simulation tools: empirical programs based on experimental results and experience; semi-empirical programs based on experimental results in addition to basic physics; and physics-based first principles programs that require complex mathematics and accurate material thermophysical data.

Empirical and semi-empirical programs use tables of experimental results, rules and guidelines and physical and algebraic equations to model a physical process. Typically, the problem is broken down into many small calculations via a finite differencing method.

In first principles programs, complex physical relationships and equations are used along with detailed material physical data. The problem must be broken into small calculations via either a finite differencing method or a finite element method. This allows for calculations profile process changes as a function of time.

Filling and Solidification Modeling

The aim of modeling the filling and solidification of a casting is to:

* predict the influence of the gating/methoding/casting design on turbulence;

* predict oxide entrapment and other flow-related defects;

* provide temperature profiles during and at the end of filling for more accurate solidification analysis;

* predict the pattern of solidification, indicating where shrinkage cavities and associated defects may arise;

* predict other solidification defects (such as hot tears);

* predict solidification times;

* predict microstructure of casting sections (such as segregation);

* predict stress/strain.

Mold filling can be predicted using a first principles program. An approach is to employ Navier-Stokes flow equations with a two-equation turbulence model. This approach accounts for all relevant flow characteristics (such as viscosity, friction, Reynold's Number, etc.) in all spatial degrees of freedom. Continuity and the conservation of momentum and energy equations form the basis of these algorithms. This approach is accurate and has been time-tested in the aerospace industry for decades.

Casting solidification can be predicted using both semi-empirical and first principles programs. The differences, advantages and disadvantages of each of these approaches are as follows.

Simple empirical programs only can model simple, repeatable processes where the process variables only change slightly. While feeding distance (low carbon steel in green sand) could be accurately predicted, defect prediction from a simple empirical analysis would likely be in error. Semi-empirical programs can effectively model relatively simple casting processes. Process variables must have limited, known ranges, and the physics must be simple and straightforward. An example of such a situation is steel solidification characteristics for simple shapes, and in this case, semi-empirical programs can provide accurate defect predictions.

First principle programs can model simple or complex processes. The physics can be complex, the process variables can change greatly and new processes can be analyzed. This approach is well-suited for solidification predictions for all metal types.

First principle programs generally employ Fourier heat transform equations, which model transient, 3-D heat transfer. Convective heat flows also are accounted for, which is important for large casting sections. Detailed material thermophysical data allows conduction, convection and radiation heat transfer equations to be applied to the entire process. The approach is a stepwise, time-marching analysis and allows temperature predictions at any location at any time in the process. Time to solidify, porosity, riser piping, alloy segregation and stress/strain profiles also can be predicted.

Use of Simulation

Foundries increasingly are using computer simulation tools daily. A foundry may employ various levels of computer simulation at one time or another to optimize casting processes. For example, an empirical program is used to estimate methoding requirements. A semi-empirical program is used to quickly check the method and to further optimize the process. Finally, a first principles program could then be used to evaluate filling characteristics and provide a more detailed analysis of the method and the casting process. In addition, a first principles program also is used to evaluate a new process or very complex design.

Ideally, the use of computer simulation allows the foundry to reduce turbulence and fluid flow defects. In addition, yield, riser size/placement, gating changes and casting orientation are optimized, and the cost of improving the casting process is reduced significantly.

Both computer hardware and software technology is advancing at an increasingly rapid pace, and new, calculated processes are being performed that were not even thought possible a couple of years ago. The following survey was compiled to provide detail on the latest features of the process modeling software displayed at this year's CastExpo in St. Louis. Each description includes a success story of how a particular program has optimized a foundry's casting operation.

AFS SOLIDIFICATION SYSTEM (3-D)

American Foundrymen's Society, Inc. 505 State Street Des Plaines, Illinois 60016 847/824-0181, fax [phone], email: [e-mail], website: www.afsinc.org

Contact: Laura Moreno.

General Methodology: Finite difference.

Software Features: Combined thermal and volumetric calculations provide accurate predictions of shrinkage porosity. Cast scan and slide shows provide X-ray and animated results that can be viewed without extra viewing software. The complete package is sold as a site license, and the program can be installed on multiple machines at one location at no extra cost.

Process Applications: Sand, shell, investment and permanent mold.

Computer Platforms: Pentium PCs.

Number of Installations: More than 330, including Prospect Foundry, Minneapolis; Bimac Corp., Dayton, Ohio; and Stahl Specialty Co., Kingsville, Missouri.

Training, Support and Update Options: Training is available for two people per site at no charge as a bimonthly training course. Telephone, fax and email support is offered at no charge. There is no charge for updates for the first year. The fee for continuing support, training and updates after the first year is currently $1000/yr/site.

Case History: A steel grinding blank for the food processing industry was produced via investment casting by Bimac Corp. The foundry had produced three wax patterns via rapid prototyping. The first two had been rigged using conventional methods but had not produced a sound casting. Using modeling to help design a better rigging system, 12 simulations were run to test various options but did not show significant improvement.

The design team then did some brainstorming in the hope of producing a successful design, and a top risering scheme was discussed as a possible alternative. This approach was modeled, and the results showed that a casting could be made this way. The final pattern was rigged according to the simulation results, and a sound casting was produced.

This part was new business that the foundry was trying to acquire. The simulation helped it bid successfully for the job, which provided $500.000/yr in business for the foundry.

CASTCAE

CT-Castech, Inc. Oy (CASTech, Inc.) P.O. Box 524, Tekniikantie 21 B FIN-O2151 Espoo, Finland 773/755-0441, [phone], fax [phone], email: [e-mail], website: www.castech.fi

Contact: Anna-Leena Yli-Karhu.

General Methodology: Calculates filling and solidification (misrun, defect and hardness prediction), finite difference, control volume and full Navier-Stokes.

Software Features: QuickTime and Windows movie files (simulation results can be viewed without simulation software), advanced iron model, exothermic model, X-ray visualization and industry-verified filling calculations (video experiment results available).

Process Applications: Sand, shell, investment, diecasting, permanent mold and semi-solid.

Computer Platforms: Windows95/98/NT, MacOS, Linux and UNIX.

Number of Installations: Approximately 30, including Giddings & Lewis Casting Technology, Menominee, Michigan; Sparta Foundry (Kurdziel Industries, Inc.), Sparta, Michigan; and Ellwood Engineered Castings Co., Hubbard, Ohio.

Training, Support and Update Options: Training (1-5 days) is available on site or at CASTech. Support is available via phone (toll-free), fax, email and Internet. Updates are made several times a year and are included in the maintenance option. Maintenance and training are optional and manuals are included in the software.

Case History: Sparta Foundry, Inc. uses the vertically parted green sand process to cast pulley rings in gray iron. After experiencing some problems with the parts, the foundry had the situation simulated with CastCAE.

The simulation revealed that the ingates were not working properly, and the pressure distribution was causing the parts at the bottom row to fill too quickly. In addition, parts on the top row were filling slowly, and the melt was rippling in through the ingates, causing rapid melt cooling. This resulted in the mixing of liquid and solid metal, leading to surface problems. At the end of the filling, before the last parts were filled, there seemed to be a risk of cold laps forming at the top of the castings on the top row. Also, some shrinkage was predicted inside the castings.

The predictions were found to match well with the real situation. The foundry had problems with bad surface quality on many rings, as well as some shrinkage appearing inside the rings in predicted locations that was revealed in machining.

The simulation showed the cause of the problem and, based on this information, the foundry engineers redesigned the gating system so that all parts would fill simultaneously. After the change, the scrap rate dropped dramatically. The foundry engineers commented that the reason for the problems they had with the part were a mystery until the simulation showed what really was happening.

EKK METAL CASTING SIMULATION SOFTWARE

EKK, Inc. 2065 West Maple, Suite C309 Walled Lake, Michigan 48390 248/624-9957, fax [phone], email: [e-mail], website: http://ic.net/~ekk

Contact: Shawn Mahaney.

General Methodology: Finite element.

Software Features: Accuracy, speed and efficiency.

Process Applications: all.

Computer Platforms: LINUX, IRIX (SGI) and HPUX (HP).

Number of Installations: 18, including Hayes-Lemmerz, Romulus, Michigan; Madison-Kipp, Madison, Wisconsin; and GM Powertrain, Defiance, Ohio.

Training, Support and Update Options: Licenses typically include 2 weeks training at EKK's facility and unlimited phone/email/ftp support. Additional training also is available. Updates are annual, sometimes sooner when important features are introduced.

Case History: A differential carrier bracket was analyzed for Bohn Aluminum Corp. (Citation Co.), Butler, Indiana. The aluminum casting was to be made two-on in a permanent mold gravity-pour carousel. This work was performed at the earliest stages of tool design for maximum benefit. After 2 weeks of initial model construction, a series of simulations was performed on a variety of tool and process designs. Seventeen variants were simulated in about 3 weeks. A process was designed that could make good parts, but some process sensitivities remained. After initial tool tryout and additional simulations, part design changes were requested of the final customer. Seeing the simulation and testing work done, the customer allowed the changes despite the addition of some mass in order to maintain casting quality.

FLOW-3D

Flow Science, Inc. 1257 40th Street Los Alamos, New Mexico 87544 505/662-2636, fax [phone], email: [e-mail], website: www.flow3d.com.

Contact: Michael Barkhudarov.

General Methodology: Control volume and finite difference.

Software Features: Free surface modeling, geometry embedding with the Favor method and free gridding.

Process Applications: Sand, diecasting, permanent mold, lost foam, sand coreblowing, semi-solid metalcasting, tilt pour and centrifugal casting.

Computer Platforms: SGI, Sun, DEC, HP, IBM and Windows95/98/NT.

Number of Installations: 90, including General Motors, Saginaw, Michigan; Alcoa, New Kensington, Pennsylvania; and Kolbenschmidt Aluminium Technologie AG, Neckarsulm, Germany.

Training, Support and Update Options: Two-day training at Flow Science consists of lectures and hands-on sessions with the third day dedicated to individual consulting and examining specific customer problems. Support is provided via fax, phone and email by highly trained engineers. Updates are provided to all current users on an annual basis.

Case History: Werkzeugbau Schaufler, a pressure diecasting toolmaker in Laichingen, Germany, uses FLOW-3D to validate the design of its dies. With the requirements for pressure-tight castings for the automotive industry becoming more stringent, it is important that the entrapment of air during filling of a die at high velocity is minimal. Schaufler engineers validate the runner and gating system, the die cavity itself and the overflows by running filling simulations with simultaneous heat transfer before any part of the die is made.

The runner/gating system for a six-cylinder engine magnesium inlet manifold produced by Kolbenschmidt Pierburg AG for Daimler-Benz, Stuttgart, Germany, was a particular area that needed Schaufler's attention. With the wall thickness below 2 mm, it was critical that the filling pattern be smooth and balanced with minimal fluid breakup and air entrapment. The simulation result highlighted significant flow separation in the gating system and showed that metal arrived at the gates at different times, which is undesirable. Alterations were made to the runners for a more efficient feeding system.

All in all, four dies were produced to manufacture the different components of the complete manifold. Kolbenschmidt Pierburg AG received the 1998 Design Award by The International Magnesium Assn. for the final casting.

MAGMASOFT

Magma Foundry Technologies, Inc. 2340 South Arlington Heights Road Arlington Heights, Illinois 60005 847/427-1001, fax [phone], email: [e-mail], website: www.magmasoft.com

Contact: Tim McMillin.

Software Features: Alloy-specific simulation, including calibration.

Process Applications: Sand, shell, investment, diecasting and permanent mold.

Computer Platforms: WindowsNT.

Number of Installations: More than 350, including Grede Foundries, Milwaukee; Waupaca Foundry, Waupaca, Wisconsin; and Atchison Casting Corp., Atchison, Kansas.

Training, Support and Update Options: Magma provides foundry-experienced engineers, not software developers, for training and implementation. With the largest development staff of anyone in the industry, the company assures that its tools will be further developed and supported well into the 21st century.

Case History: St. Cloud Foundry (Grede Foundries, Inc.), St. Cloud, Minnesota, was experiencing difficulty obtaining good directional solidification in an iron off-road vehicle hub it was casting in green sand. The difficulty was due to the gating design and use of ductile iron. Through the use of Magmasoft and the predictive tools specific to iron, the foundry was able to visualize the porosity problem and determine the best solution.

The simulation tool also helped the foundry predict mechanical properties of the iron for consistency in hardness and elongation strength. This ensured a sound, cost-effective casting was delivered to the OEM.

NOVAFLOW & SOLID

NovaCast AB Softcenter S-Ronneby, 37225 Sweden +46/457-71830, fax [phone], email: [e-mail], website: www.novacast.se

Contact: Henrik Barth.

General Methodology: Finite difference.

Software Features: Advanced algorithms that take gravity and flow into consideration during mold filling and solidification.

Process Applications: Sand, shell, diecasting and permanent mold.

Computer Platforms: Windows95/98/NT.

Number of Installations: 102, including Liberty Foundry (Grede Foundries, Inc.), Wauwatosa, Wisconsin; Urick Foundry Co., Erie, Pennsylvania; and Mansfield Foundry Corp. (Citation Co.), Mansfield, Ohio.

Training, Support and Update Options: Service, support (via phone, fax and email) and free upgrades for 6 months are included in the purchase price. There are normally two new releases per year. Product training runs on a regular basis, and in-house training is available.

Case History: For Nya Kohlswa Gjuteri AB, Kolsva, Sweden, the primary objective was to investigate possible ways of reducing scrap rate caused by cracks and inclusions, to improve yield and, if possible, to reduce the time used for cleaning, finishing and machining. With a pouring weight of 940 kg and a yield of only 51%, there was ample room for improvement.

A low-alloyed carbon steel casting was modeled, and the gating system was dimensioned for a flow velocity of 500 mm/sec to minimize tendencies for turbulence.

Initial studies and calculations indicated that two feeders instead of the original five might be sufficient to feed the entire casting. A simulation was needed to see if this was true or not.

The flow and solidification simulation was run on a Pentium II 450 256 Mb RAM WindowsNT. With a mesh size of 2.5 million cubes, the time needed to complete the flow simulation was 5.5 hr, and the solidification simulation would take another 30 min.

The simulation displayed a nice, easy and uniform filling of the casting. The solidification pattern and the shrinkage prediction indicated that the use of only two feeders most likely would work.

The new gating and risering layout was tested in the foundry, resulting in a new pouring weight of 725 kg and a 67% yield. Cleaning room and machining time was reduced by 2.5 hr/casting, and two exothermic sleeves were used instead of five.

SIMTEC 

Simtec, Inc. 3663 Broadmoor SE Grand Rapids, Michigan 49512 616/975-3882, fax [phone], email: [e-mail], website: www.simtec-inc.com

Contact: Juergen Biermann.

General Methodology: Finite element.

Software Features: Fluid flow, solidification, stress, microstructure and sand/core shooting.

Process Applications: Sand, shell, investment, diecasting, permanent mold, tilt pour, lost foam, continuous casting, centrifugal casting and squeeze/semi-solid casting.

Computer Platforms: UNIX, Windows 95/98/NT and Linux.

Number of Installations: 150, including Newport News Shipbuilding, Newport News, Virginia; Thyssen, Essen, Germany; and Progress Casting, Plymouth, Minnesota.

Training, Support and Update Options: Training is done, preferably, at the customer's site, on their platform and with their application. Recommended are 5 days of training, after which the first real project will have a solution. The maintenance and support contract allows unlimited telephone support by actual users (project engineers) and free software updates (usually 3/yr).

Case History: Caterpillar, Inc., Peoria, Illinois, experienced problems with a counterweight for its Excavator, cast by Kurdziel Industries, Muskegon, Michigan. The gray iron casting, which weighed 17,000 lb and was 9 ft in length, was becoming distorted after removal from the mold and, as a result, was not meeting Caterpillar's print specifications of [+ or -]6 mills. This created assembly problems when matching the steel door to the casting.

Primary root causes, identified by Simtec using finite element analysis, indicated that the cope pull time was premature. The casting had not yet reached annealing temperature, and the presence of tie bars and large diameter vents with backdraft contributed to the problem when the cope was removed. Even without the cope catching on something, high residual stress has a tendency to cause distortion. This is a direct result of top-directed chilling confronting the still-hot sand from the bottom of the casting.

By simulating cope removal at 24, 48, 72 and 96 hr, it was found that early pulling of the cope was a direct contributing factor to the experienced distortion. Simtec was able to show the optimal reshaping of the risers (vents), demonstrate that the tie bars were unnecessary additions to the casting, and verify the optimum time for opening of the cope and for pulling of the casting.

Simtec is now employed for education and identification of critical process parameters, and providing the results of neglecting these parameters. Casting warpage was reduced by 73% and the majority of the castings have been brought within allowable tolerance specifications.

PROCAST

UES Software, Inc. 4401 Dayton-Xenia Road Dayton, Ohio, 45432 937/426-6922, fax [phone], email: [e-mail], website: ues-software.com

Contact: Stephen Barry.

General Methodology: Finite element.

Software Features: Radiation for investment, ThermalCalc database, micromodeling, stress and distortion, and automatic meshing.

Process Applications: Sand, shell, investment, diecasting, permanent mold and lost foam.

Computer Platforms: WindowsNT and UNIX.

Number of Installations: 150, including Tool Products, Minneapolis; Gould Pumps, Ashland, Pennsylvania; and Howmet Corp., Whitehall, Michigan.

Training, Support and Update Options: Training consists of a 3-day class, which is offered monthly. Technical support is available 8 a.m.-5 p.m. EST on business days, via phone or email. Updates are available to all customers who are leasing the software or who have opted to purchase an annual maintenance agreement.

Case History: A recent success story involves Eagle Alloy, Inc., Muskegon, Michigan, and a railway casting that had to be turned around quickly in order to win the job. The critical issue was casting accuracy and pattern durability.

"Because we wanted to minimize processing on the part, we elected to have 100% CNC-machined patterns built, which more than doubled the cost of the tooling," said Jim Smith, Eagle Alloy's engineering manager. Therefore, it was essential that the tooling worked right the first time. "We did not want to invest time and money into patterns only to find that we needed some costly and time - consuming changes after the first samples were made," Smith continued.

The foundry employed simulation software to avoid such a scenario. Using ProCAST, the tooling engineer identified large shrink porosities in the casting. Working with the customer's design engineers, a solution was quickly proposed and again simulated with ProCAST. The result displayed that the new configuration would produce sound castings. The first article samples were X-rayed and showed no defects. This quick turnaround from initial design to on-time and in-spec part delivery netted over $250,000 in annual sales.