Simulation is more than Software

0 0
The early years of CADFEM GmbH

Guenter Mueller, Founder of CADFEM: the finite element method has been with me all my life

The early years of CADFEM GmbH

In 1941, Konrad Zuse invented his computer (the Z3) in Germany and Eckert and Killby invented their computer in the USA. Professor Stiefel of ETHZ in Zurich was able to solve differential equations using a follow-up model, the Z4. I was born in 1944.

Education at University of Stuttgart | 1963 - 1970

After graduating from high school, I started a civil engineering course at the University of Stuttgart. Children from a humble background were not able to choose where they studied. Assuming you got the chance, you could study a subject that to some degree suited your inclinations, at some nearby institution. I had always enjoyed playing with my metal construction set – I was simply predestined for civil engineering. 

On my course I attended lectures on the principles of mechanics (J. Bernoulli, …), differential equations, (Navier, …), and variational methods (Leibnitz, …). In 1968, Erwin Stein of the Institute for Structural Design (Institut für Baustatik) gave a fresh and challenging lecture on tensor algebra at which I was one of three participants. This lecture provided me with the grounding I needed for my diploma thesis, in which – building on the work of Ray Clough, Professor at the University of California, Berkeley, who had coined the term “finite element method” in 1960 – I created an FE program for analyzing planar structures.

Scientific Assistant, Institute for Structural Design, Stuttgart University – Visiting Scholar, University of California, Berkely | 1971-1977

In 1970, Erwin Stein offered me a position as a scientific assistant so that, during my tenure as an assistant from 1970 to 1976, I would be able to focus on the work of the 20th century mathematicians who had developed the mathematical background to the finite element method – the calculus of variations. Ritz, Galerkin, Trefftz, and Courant had developed numerical methods for solving variational problems, however without the use of computers, they were only able to solve problems containing a small number of unknowns. My dissertation was about discretizing the Trefftz method and expanding it to cover multiple field regions. 

I was able to spend the period between the middle of 1976 and the end of 1977 as Visiting Scholar at the University of California, Berkely, funded by a grant from the German Research Foundation (DFG). During lectures, research activities, and faculty meetings, I had contact with Ray Clough, Edward Wilson, and other world-renowned scientists from the field of FEM. My supervisor was Professor Alex Scordelis, who was conducting researching into FEM methods in relation to shell structures.  

Working at IKOSS and Control Data | 1978-1982

In 1978, after returning from the USA, I was employed at IKOSS GmbH as a computational engineer, however after a few months I went over to Control Data GmbH. Control Data Corporation (CDC) was the leading provider of mainframe computers designed for technical and scientific applications, renting out computing power on a time-sharing basis. My job was to provide support to customers who were using a variety of FEM programs such as TPS-10, STARDYNE, ASKA, MARC, ADINA, ABAQUS, and ANSYS.

In 1979, I was offered a position as “District Manager, Professional Services”, which meant I had to move with my family from Stuttgart to Munich. The work was challenging and well-paid, but the company had been spoiled by success, and the development of microcomputers passed it by. Fearing that CDC had no future, it struck me that I might go self-employed. It was Christmas 1981 when I phoned John Swanson and asked him whether he might need someone in Europe to deal with marketing, distribution, and support for his ANSYS program. John gave the brief reply “Yes”. When I asked what I would need to do, he replied: “You have to leave your company”.

The office of Guenter Mueller’s Company for Computer-aided Analysis in Civil and Mechanical Engineering | 1982 - 1985

So that is what I did. In May 1982, I started the Company for Computer-aided Analysis in Civil and Mechanical Engineering, based at my home office in Grafing, near Munich. I was aware of the fact that by resigning from my well-paid job at CDC I was taking a risk. The responsibility for feeding my family with two small children lay with me, as my wife was only permitted to teach in the Baden-Wuerttemberg region and so was not allowed to teach in the high schools in Bavaria.

Being an ANSYS Support Representative (ASR), the main activity of my new engineering company was providing Swanson Analysis Systems, Inc. (SASI), with representation in terms of regional marketing, distribution, and support. Looking across the globe, there were only six ASRs in North America and three in Europe. Stewart Morrison was responsible for Great Britain, and Hervè DeBaecker was responsible for France, while I was concerned with Germany, Switzerland, and Austria, and other countries in the rest of Europe.

In this initial period, I had to do a lot of multitasking, because as well as distribution, support, training, and project work, I also had to do administrative tasks, though my wife and an acquaintance provided support. I was not able to afford a “real computer”; I only had enough money to purchase an Apple computer with 16 kB of memory (at a cost of 11,000 German marks) – which we used for administrative tasks. On some days I was graciously leant Tektronix terminals and Prime computers for use in seminars and presentations. I was able to undertake seminars providing an introduction to the finite element method, seminars on dynamics, ANSYS information days, and seminars on nonlinearities, super-elements, and heat transfer. John Swanson and Peter Kohnke from SASI came to my assistance when it came to giving seminars on new topics such as optimization and magnetostatics. At this point in time, I was doing a lot of traveling to give presentations and seminars, and this included traveling outside Germany – in Switzerland, Austria, Italy (repeatedly), Norway, Finland, Romania, Serbia, and Egypt.      

The time was ripe for a first customer meeting. I chose “ANSYS Theory” as the theme. The speaker was Peter Kohnke from SASI, who, on account of his German roots, was able to speak some German. Nearly 30 participants gathered in October 1982 at the Kreitmair Inn in Keferloh, near Munich. In the beer garden, a decision was taken to meet up annually. The first users’ meeting, involving around 30 participants was held in Herrsching am Ammersee, in 1983. Further meetings followed every year – with growing success and without interruption.  

CAD-FEM Gesellschaft für computerunterstützte Konstruktion und Berechnung mbH) | The first ten years, 1985-1995

In 1985, John broadened the range of tasks expected of distribution partners, and designated them “ANSYS Support Distributors” (ASDs). They now had the exclusive right to supply ANSYS within their sales territory, in their own names. This meant I had to start a company: CAD‑FEM GmbH. The name, CAD-FEM – which, at the time, was still hyphenated – came about because the intention was that both CAD-based and FEM-based services would be offered. My partner, Otto Jandl, took on the CAD-based activities, however after a little while he left the company and returned to his former employer, the Max Planck Institute.

CAD-FEM obtained an office in the industrial park in Ebersberg near by Munichand invested in its own computers – a MicroVAX II with 2MB of main memory costing 100,000 marks and a Tektronix 4014 terminal costing 40,000 marks. Communication was carried out via telephone, TELEX, and a newly acquired fax machine. Quotations, accompanied by brochures, were taken to the post office in batches. Visiting clients and potential customers by car involved navigating with a Shell Atlas on your lap. In 1985, sales revenue reached 1.15 million marks (€ 608,000), and we had a total of 50 customers – all of whom were marked out with pins on a wall map of Europe. 

Over the years, a number of complimentary programs were brought to the market, including FLOTRAN (fluid mechanics), BEMAP, SYSNOISE, COMET (acoustics), C-MOLD (injection molding), ADAMS, SIMPACK (multi-body systems), and MAKE-FATIGUE (structural durability). Since John Hallquist founded the company LSTC in Livermore in 1987, CAD-FEM has also been the distributer of LS-DYNA software designed for highly dynamic, explicit analysis.

True to John Swanson’s words, “We don’t sell software, we license technology,” a great deal of emphasis was placed on training users and on project activities, which was also a means by which we were ourselves able to gain experience of using the software. Over the years, the range of seminars was constantly extended, so that from fall 1995 through summer 1996, we covered about 30 subjects and held over 70 seminars, each lasting several days. Seminars were also conducted for events run by hardware firms, high schools, and institutes of further education such as Esslingen Technical Academy and Essen House of Technology.

In 1992, the first volume of “FEM for practitioners” was published by expert-verlag (Expert Publisher), which – due to the participation of my colleagues, Clemens Groth and Uli Stelzmann – was followed by a further three volumes. This showed people how to deal with applying the finite element method to structural mechanics, dynamics, and electromagnetic fields.

In the realm of marketing, the first “infoplaner” was issued, providing information every six months concerning news (the info) and the dates of events and seminars (the “scheduler”). 

In 1993, the company had 20 permanent employees, eight freelancers, and eight temporary workers, and generated a turnover of around 11 million marks (5.5 million euros). The following offices were opened: 1990 in Stuttgart, 1991 in Hannover, 1993 in Berlin and 1995 in Chemnitz.

CAD-FEM was represented abroad

  • in Malters (Switzerland) from the SEFAG company by Mr. Wuest (1990),
  • in Moscow by Juri Ryasnov (1990),
  • in Brno for the Czech Republic and Slovakia by SVS FEM (1991), and
  • in Poland from MESco (1994).

CAD.FEM AG, founded in 1995 in Aadorf, Switzerland, succeeded as company representatives in Malters. Markus Dutly – who, in his capacity as a freelance worker for CAD-FEM in Grafing (and subsequently in Vienna) had been assisting with the distribution of ANSYS – became managing director and shareholder. 

John Swanson sells his company, SASI, Inc. | 1993

SASI had come on in leaps and bounds. New versions were being released onto the market every year, with Version 5 coming out in 1993. It had been completely re-written, incorporating a new data structure and numerous new developments, such as a common database for preprocessors and postprocessors, Boolean operations, support for inputting via “control panels”, video output and animation, automatic adaptive meshing, error norms, automatic load increments, large strains and rotations, contact surfaces, hyper-elastic material, fluid mechanics, and interfaces. In 1992, SASI’s sales figures increased to approximately 30 million dollars and the number of employees grew to around 150. The number of ANSYS support distributors was 35, while commercial installations of the software stood at 2,500, plus an additional 1,600 contracts with institutes of higher education.  

It came as a complete surprise when – in typical fashion – I received a brief message in October from John, via our fax machine: “I sold my company. John.” That really hit me! I had just signed a ten‑year lease agreement for an office in Grafing. And a meeting involving Ralph Donahue from SASI was scheduled to take place the very next day at Siemens KWU for the purpose of extending an ANSYS licensing agreement. On the way to Siemens, we talked about whether we ought to mention the news. To our great surprise, the manager from Siemens reacted to the news positively: “The investor, TA Associates, is a well-renowned company!”

The takeover of SASI by TA Associates was concluded in 1994, and in 1996 the company which had now been re-named ANSYS, Inc., went public. Being one of the most important ASDs, I was visited in Munich by Jacqueline Morby of TA-Associates 1994, so that she could bring me up to speed regarding the future direction of the company. It meant a big change in culture. New management came on board, with Peter Smith as CEO and Klaus Schlemper as European Manager. As ANSYS support distributors, we were required to sign new contracts involving quarterly sales targets, and we were now designated “ANSYS channel partners.” 

The years since SASI

We ASDs were concerned about our future. Close friendships had developed over the years, and the meeting in Houston at SASI had been an annual highlight. It had been a real joy working for John’s company, Swanson Analysis Systems, Inc. For several years, we organized meetings to talk about how we could go on working together. The first meeting, involving ASDs from CADFEM, ITALCAE, STRUCOM, and A&Z was held in Rome, in 1994.

At the meeting in Lugano in 1998, we decided to set up a company that would operate either as a general provider of simulation-based expertise and software or simply as an institution comprising an organized network of CAE experts from industry and universities. This company, Technology Network Alliance AG, is still around today, and since the year 2000 it has mainly been concerned with organizing the biannual meetings of the TechNet Alliance – a global network of experts in numerical simulation hailing from industry and science.  

It took a few years for the ASDs to get used to the new situation. It was only in the year 2000, with Jim Cashman as the CEO of ANSYS, that the atmosphere took a turn for the better. Intensive program development, acquisitions, and strengthened sales and marketing activities had led to strong growth in turnover for ANSYS, Inc. The acquisitions had the effect of establishing ANSYS in several countries; in Germany this was achieved on the back of FLUENT and CFX.

Once ANSYS Germany GmbH had been set up, CADFEM GmbH was no longer the exclusive distributor of ANSYS. The year 2006 saw the market being roughly divided into two realms: structural mechanics and fluid mechanics, with CADFEM being confirmed as the center of expertise in the realm of FEM and ANSYS Germany as the center of expertise in the realm of CFD.  

In 2018, CADFEM signed a new long-term partnership agreement with ANSYS, regulating their places in the market. CADFEM (Suisse) AG and CADFEM (Austria) are included in the agreement.

Dr.-Ing. Guenter Mueller, Founder of CADFEM GmbH and Managing Partner at CADFEM International GmbH

© Images: private | Translation: Kieran Scarffe

Read more related content