Development of 2x2 Model Predictive Control Model For Crude Distillation Unit

• Main Author: Balakrishan, Vijendran
• Format: Final Year Project
• Language: English
• Published: Universiti Teknologi PETRONAS 2012
• Subjects: TP Chemical technology
• Online Access: http://utpedia.utp.edu.my/9710/1/2012%20-%20Development%20of%202X2%20Model%20Predictive%20Control%20Model%20for%20Crude%20Distillation%20Unit.pdf; http://utpedia.utp.edu.my/9710/

This report is documented mainly discuss about the final year project entitled "Development of 2x2 Model Predictive Model for Crude Distillation Unit". Advancements in the oil and gas industries requires parallel progress both in maximizing production rate and profit. One sector in which those objectives are accessible is in the refinery business. Core business ofthe refinery sector is swarmed around the crude distillation unit (CDU) which separates raw crude into few marketable products. Due to its high nonlinearity profile and sensitivity of profit margin, any advancement in CDU is considered to be essential. Many researches and engineers use CDU as their case study for projects and paper works to contribute on the optimization, control and production problems. This piece of literature narrows it's scope to control issue of the CDU in which system identification and simulation of CDU system will be developed. Main purpose of this study is to investigate whether development of 2 by 2 MIMO model using Model Predictive Controller (MPC) can increase the performance and reproduce actual data ofCDU to the respect to the variables chosen. Contribution of this research channels to error minimization produced by MPC in which evaluated by minimal controller moves and fluctuations of chosen control variables in comparative to its set points. Testing data from virtual plant will be used as base case to develop relevant robust mathematical model to be eligible for representing CDU system and performance analysis on the chosen model were conducted to derive relevant conclusions. Both research work is possible using MATLAB and HYSYS in which needed materials and toolboxes are available.