Chistoff Heisser (r) receiving his speaker award from Meeting Chair Kathy Hayrynen

BIO

Christoff received his equivalent of a Masters Degree in Foundry Technology at the Technical University of Clausthal in Clausthal, Germany.   After his first employment as Leader of Research and Development at Thyssen Feinguss, an aluminum investment casting foundry in Soest, Germany, he joined Magma GmbH in Aachen, Germany in a marketing and support position.  Christoff moved to Magma Foundry Technologies, Inc's Chicago office in 1995 as Foundry Application Engineer.  He now is the President of Magma Foundry Technologies, Inc.  

The development of computer processing speed over the last decade is truly amazing. In 1995, when we started measuring simulation speed to give customers hardware recommendations, an average casting process simulation (filling and solidification) took more than 2 days. This was on high performance workstations at a price point of US$150,000.00 and above. Today, a basic PC for less than US$3,000.00 can perform the very same simulation in less than one hour. Spend twice that much, investing in a “quad-core” PC, and it will be done in less than half an hour. How long would it take if someone would invest the same US$150,000.00 into a 32 core Cluster? Less than 3 minutes - more than 1000 times faster than in 1995!

What implications does this have on the use of Casting Process Simulation? First and foremost, it is much more usable as it provides answers much faster. Secondly, it is affordable even for small foundry operations as the return of investment is achieved much sooner. But it also leads to a paradigm shift. Traditional Casting Process Simulation is based on “What-If” scenarios, demanding a user to make a decision, implement changes, and start another “run.” The user used to wait for the completion of a simulation or ran it over night to check on it the next morning. Today, if the operator starts a simulation just before he leaves for home, the computer might finish the simulation in less than one hour and sits idle for the next 14 hours! One way around this would be to have the engineer set up ten different designs (i.e. ten different risers, gate configurations, pouring temperatures, etc.) and queue them up. Lot ’s of work in setup, and at the end he has to review all simulations and pick the best one – the best one of the ten, not the optimal one! The solution to this dilemma is to combine the Casting Process Simulation tool with an Autonomous Optimization Tool. The later are used for many years in the design community to optimize designs. MAGMAfrontier is an add-on module to the MAGMASOFT suite of application modules. It can be added to the existing software and thereby allows the user to utilize autonomous optimization.  

How does it work? Instead of setting up i.e. ten separate simulations of ten riser designs, the user defines a parametric riser geometry and defines, i.e., upper and lower limits of its radius or several radii over its height, upper and lower limits of it’s height and where to locate it. One could as well define riser neck dimension limits and location. After defining the objectives of producing a defect free casting and minimize riser volume the optimization can be started. A genetic optimization code assures the meaningful combination of parameters and target oriented modifications leading to the optimum result – sort of like children having the best of each parent combined in them. At the end, the optimization will depict the optimal riser, leading to a defect free casting with the smallest riser volume – the only result the user has to review. The result evaluation will provide the user with information on how the design developed and why the depicted best design is the best design.

One setup – multiple autonomous simulations – one optimum result.

Autonomous Casting Process Optimization minimizes the time needed for the operator to spending time on input and evaluation and, at the same time, maximizes the utilization of the hardware. The key to this technology is the easy setup, the multi-objective optimization (just what a foundry engineer has to do every day), and straight forward result evaluation pinpointing the optimal result. MAGMAfrontier can be used to optimize almost any parameter that is considered in state-of-the-art casting process simulation: Casting-, riser-, gating-geometries, process parameters, mechanical properties, residual stresses and hot tears, etc.

In a challenging environment of ever higher demands on casting quality, delivery time, costs, and scars qualified human resources Autonomous Casting Process Optimization is an enviable tool to conquer these challenges and best use the available engineering capacity, retain qualified people and attract the next generation of foundry engineers.