Karsten Rist
I could hardly believe my eyes. The e-mail on the computer screen in front of me came from a PhD student in mining at the University of Porto in Portugal. He asked: ”are you the author of a paper Computer Simulation for Solution of a Mine Transportation Problem? I am trying to locate him”. I was. The paper was published in the December issue of Mining World in 1961. As I looked for a copy to send to Portugal memories of days long ago flooded my mind:
The summer Sunday in 1959 when my friend Bob Johnson, an IBM operator, took me to Colorado Interstate Gas in Colorado Springs, where he worked, and ushered me into a big room, humming with air conditioning and filled with huge grey cabinets. Bob told me that I was looking at an IBM650 computer and explained what this machine was capable of doing. I was fascinated. That fall, when I returned to the University of Michigan to finish work on my MBA degree, I also went to the engineering school to see what I might learn about computers. I was in luck. The Industrial Engineering department offered a course in Data Processing and, thanks to my completed mining engineering degree, I was considered qualified to enroll. I never saw a computer or wrote a single instruction but, at the end of the semester, I knew all the University of Michigan had to teach its students about computer usage. After receiving my MBA degree, the Climax Molybdenum Company offered me a job in its engineering department. The assignment: Justify the installation of a computer at the Climax mine.
IBM was on the verge of releasing its first “small” computers based on transistor technology, the IBM 1620 and the IBM 1401. Big Blue was eager to break into the mining field and the Climax Molybdenum Company, highly profitable, with the biggest underground operation in the world, was a logical target. IBM supported my efforts with highly qualified people and its vast network of computer know-how.
We started out with a multidimensional statistical analysis of mill recovery data. IBM furnished the programs. The results were interesting, even enlightening, but did not come close to justifying the investment in a computer.
I thoroughly enjoyed the work on “a mine transportation problem”. My knowledge of mine operations and the course work at the University of Michigan merged to spark the idea that a model might give a meaningful answer to the question whether or not the investment in a second access tunnel for the Storke level was justified. The understanding and patience of my supervisors and the support offered by IBM were vital to the project.
The round trip of a train on the Storke level had four phases:
- Entering the mine through the portal and proceeding to #1 Switch where all dispatching decisions were made. This trip took 3.4 minutes for the empty train.
- The loading loop inside the mine, represented by 400 actually observed loading times ranging from 28 minutes to 190 minutes with a mean of 63 minutes.
- The trip of the loaded train out of the mine, lasting 5.2 minutes.
- The last phase, the unloading of the train at the crusher. Crusher times ranged from 3 to 15 minutes with an average of 5.5 minutes. They were also represented by a table of actually observed times.
The program kept track of the status of every train in the system minute by minute. A logic section dealt with the decisions which train should advance if more than one train was ready to use the mine access tunnel. The decisions were programmed to minimize total train waiting time.
IBM made time available on a brand new IBM1620 computer, installed at Kaman Aircraft in Colorado Springs. While primitive by today’s standards, it represented the latest available technology in 1961. The machine was the first to use transistors rather than vacuum tubes like all previous computers. “Our” computer had 20,000 characters of core storage (also a new technology) but no additional disk or drum storage. Each character represented one decimal digit. Alphabetic characters required two digits of storage. Addition and multiplication were handled by the use of tables. Each instruction was 12 characters long, two characters for the operation code and five each for two addresses. Input and output to the machine was either by typewriter or by paper tape. The programming system was SPS (Symbolic Programming System) that means each machine instruction was represented by one line of symbolic code. My program consisted of about 400 instructions and the SPS processor took about an hour and a half to generate the object code.
Together with the systems engineer from IBM I spent many hours in Colorado Springs compiling, debugging, testing and retesting. I remember the engineers from Kaman Aircraft walking by, kidding us and asking us how the trains were running. When we finally generated results they appeared to be consistent with the experience at the mine. The fact that the model did not require higher level mathematics or statistics made it easy to explain and the actual data embedded in the program made the model more convincing to the men who were responsible for mine operations. The model showed that there was no need for an additional access tunnel, which meant a saving in the vicinity of a million dollars for the Climax Molybdenum Company.
After the engineering report had been distributed at the mine, the general manager called for a meeting of his senior staff in his office. As the discussion progressed the Storke level manager, a capable, practical man, old enough to be my father, leaned back in his chair and said:” This model is very interesting and engineering has done a fine job, but the reality of mining is, that there are always unexpected events and problems that theory can not account for.” I had expected this comment and responded by explaining the specific, unusual circumstances represented by the six loading times above 150 minutes. These times represented the unexpected problems. They had an even chance with all the other times to be selected for any particular train run. The model really did account for the uncertainties of mine operations. The second tunnel was never constructed.
While the model showed off the capability of the new computer equipment and generated truly valuable information, it was unrealistic to argue that it justified the need for a computer at the mine.
IBM introduced PERT (Program Evaluation and Revue Technique) to the mine. This was a planning system which focused on all the tasks which needed to be completed to achieve a certain goal and how these tasks were dependent on and related to each other. The mine was operating 24 hours, seven days a week. All major maintenance work in the mill had to be accomplished during one of two annual, three day shutdowns at Christmas and at the Fourth of July.
We decided to organize the next shutdown under PERT. In preparation numerous meetings took place and we discussed in detail each task to be accomplished. The PERT program produced a lovely network image. The shutdown was a great success and described as “the best shutdown we ever had”. It was clear, however, that the success was the result of the meticulous planning and preparation, not of the printed reports the computer furnished.
I had the opportunity to present a paper describing my work at the first APCOM conference held at the University of Arizona in Tucson in 1961 (proceedings pp L2-1 to L2-15).
I had received a thorough education in computer usage and stood at the leading edge of this new technology. At the same time it became clearer day by day that Climax was not likely to install a computer in the near future. So I made the decision in 1962 to accept a job offer from Northern Natural Gas in Omaha, Nebraska, to work as an industrial engineer. While my own career advanced and demanded my full attention I lost track of what was happening in modeling and mining. That is – until I got an e-mail from Portugal.
I was eager to learn how the graduate student in Porto had found me. The answer came by e-mail: “I googled you.” What else did the Internet tell the world about me that I was not aware of? I entered my name in the search engine. There were various references to my civic involvements in South Florida. Then there was a call for papers by the journal of The Society for Modeling and Simulation International. The leading sentence caught my attention: “One of the first computer simulations in any field was the first mine simulation model written in 1961 by Karsten Rist (1).” Interesting. I continued my research into modeling and simulation by entering the name of the society. The result was 833,000 hits, of which 855 were displayed on 86 pages of references. A little more than I was prepared to read right then and there. I found it difficult to learn about the early days of the Society prior to the availability of the internet and received some most valuable help from Professor John Sturgul from the University of Idaho, whose name is intimately connected with mine simulation. To make a long story short, I was awed by the variety of fields which are now using modeling and simulation, the international scope of the ongoing work and its ever growing sophistication.
The problems and the principles of engineering work have not changed much. The tools and the sophistication of the solutions offered surely have. I hope you enjoyed stepping back in time and getting just a glimpse of the early days. I can still hear the question of the Kaman engineers: “Are the trains running yet?”.