Minimum entropy control of COT distribution in feeding process of industrial ethylene cracker

Industrial ethylene crackers have many tubes. It is difficult to control each tube COT (Coil Outlet Temperature) accurately. The traditional feeding operation in the industrial ethylene cracker is to form a cascade control structure with PID controllers, where the average COT of multiple tubes is chosen as the master control loop and the fuel gas flow rate is chosen as the slave control loop. The good average COT control result cannot guarantee the small COT deviation of each tube. Uneven COT distribution may cause process limit violation and prolong the feeding process. Therefore, it is important to control the COT distribution evenly in the feeding process. In this work, to address the problem of uneven distribution of COT in the feeding process, an advanced COT distribution control method is proposed. The fuel gas flow rate is regarded as the manipulated variable of the COT distribution, and the probability density function (PDF) model is applied to describe the distribution of COT. The PDF model based minimum entropy controller has been developed and solved with Bayesian Optimization (BO). The simulation and industrial application results have verified the feasibility and effectiveness of the proposed method.