I have MS degree in control eng. My MS thesis was about distillation column control using "feedback linearization". i studied alot of papers on the application of advanced control theories in chemical processes. Now i'm working in research institute of petroleum industry in iran. the important fact is that even in a research institute, i'm confused about how to apply those theories in a real industrial plant, and how i can prepare these information in a well known engineering procedure.

 

I have a tale to tell in this respect. A fellow I once worked with did his masters thesis on fuzzy logic. His work included writing a PLC program (in RLL no less) that used fuzzy logic to control the motor speed on a textile machine. His thesis regarding this project was over two hundred pages long, and the code he wrote to impliment his fuzzy logic algorithm was several hundred rungs long. I understood from talking with him he spent many hundred hours just writing and testing the code. It worked OK, but it was nothing spectacular.

I spent a couple hours one afternoon in a factory in Virginia solving the same problem he solved. It took about 15 rungs of RLL and worked far better then his did. I pointed this out to him, and after looking at it he declared that the logic I had created was "cascade" logic, and his logic was "fuzzy" logic. I am not all that sure what the difference is, but I did get a big kick out of it.

Point being is that many of the techniques used in academia are uneconomic in any production environment. Now and then they turn out to be useful. Thats kind of what research is about, determing what is and what is not usable in the real world.