From: Charles Cutler Subject: Cutler's Matrix Control Date: December 16, 2010 at 9:50:03 PM GMT+1 To: Cc: 'Matthew Hetzel' , 'June Cutler' Resent-From: Reply-To: Content-Type: multipart/mixed; boundary="----=_NextPart_000_0061_01CB9D30.862788C0" Hello Man fred It was good to make contact again with you. I have attached a file that contains communications I had when I was working for Shell Oil Company. It provides some historical perspective on the evolution of the dynamic matrix control algorithm. I was working for Dr. Huang on an experimental thesis at the University of Houston, when Shell offered me a group leader position in a newly created process control group in their New Orleans refinery. I had this idea of using an LP to control a process at multiple constraints. While taking an electrical engineering course on Z transforms, I realized the process dynamics could be represented numerically by truncating the infinite series and evaluating the terms of the series at each time interval. I ask Huang to let me develop this idea remotely. My proposal to him in 1969 is included in the attachment. He agreed to let me return to Houston once a month to review my progress and he would let me switch my dissertation topic. In 1973, the refinery was struck and staff personal took over the operation of the refinery. My control group in 1970 and 1971 had build a RTO and computer control system for the catalytic cracking unit. Since at the time of the strike I was the process manager of the catalytic cracking department, I did not have to ask anyone for permission to test my matrix control concept. I was locked in the refinery for 3 or 4 months so I wrote the program. Ramaker's contribution was a program that he had help develop at LSU while working on this PHD that fit a first order model to data. The inlet temperature to the furnace that went through a surge drum before entering the furnace. Ramaker's program permitted us to get the disturbance model for the effect of inlet temperature to the surge drum to the outlet temperature of the furnace. The effect of the fuel gas and draft damper were taken from plots of the data versus time. As you can see from my proposal to C.J. Huang, the concepts for algorithm were develop years before Ramaker's involment. I kept the file on the communications in Shell in the event someone challenged me as to the ownership of the algorithm. I was never challenged, but the word on the street was the algorithm belonged to Shell. It did prevent me from getting project work for my new control company. Stone and Webster Engineering Company gave me my first opportunity to use the algorithm outside of Shell, but only after their lawyers kept the attached file for two months. I still work 50 plus hours per week for my company. Our new controller is a significantly more powerful than the version Aspen still sells. We build a controller this last year that has 41 manipulated variables and 160 control variables on the largest single train ammonia plant in the world in Saudi Arabia. The controller runs at a 10 second frequency, has 420 coefficients, and executes in less than a half a second. If you get to San Antonio for some reason, I would love to have lunch or dinner with you. Have a great day !!! Charlie -----Original Message----- From: June A Cutler [mailto:june.cutler@cutler-tech.com] Sent: Tuesday, December 14, 2010 10:38 AM To: charles.cutler@cutler-tech.com Subject: FW: Hello and question -----Original Message----- From: Manfred Morari [mailto:morari@ethz.ch] Sent: Sunday, December 12, 2010 9:46 AM To: charles.cutler@cutler-tech.com Subject: Hello and question Hi Charlie, I hope that you are doing well. I will be giving a plenary this week in Atlanta at the CDC, where I will comment on the history of DMC. The story that I remember is that you (and Brian Ramaker?) developed the concept while on strike duty somewhere in Texas. Could you confirm this and also tell me the year? Clearly it is not critical, but for posterity it would be great to nail down the story and the date. Thank you very much for your help. If you happen to be in Europe at any time we would be most pleased to invite you for a visit - even on short notice. Just let me know. If nothing else I can introduce you to my favorite Italian restaurant here in Zurich. Best wishes, Manfred ____________________________________________________________ Prof. Manfred Morari phone +41 44-632-7626 Automatic Control Lab mobile phone +41 79-642-2394 ETHZ / ETL I 29 private fax +41 44-912-3843 Physikstrasse 3 email Morari@ethz.ch CH-8092 Zurich / Switzerland http://control.ee.ethz.ch