For part of my Senior Design Project, I have to optimize a plug flow reactor (PFR). In this reactor, ethylene goes through partial combustion with oxygen to produce ethylene oxide. For more details about this project, please see here. Here I’ll go through with you in details how my group and I designed the reactor.
In the ethylene oxide production process, a plug flow reactor was used due to its proficiency in handling vapor phase reactions. In modeling with ASPEN Plus, it is assumed that, inside a PFR, the gas flows consistently and there is no radial variation in velocity, concentration, temperature, or reaction rate. The kinetics for the reactions proved difficult to model correctly in ASPEN, however they were eventually correctly implemented. The partial pressure of the reacting gas is the driving force for the reaction. The coefficients for the driving force were converted from the given kinetics and catalyst weight was selected as the rate basis.
Since the exothermic reaction requires constant cooling from a thermal fluid to prevent a runaway reaction, boiler feed water was chosen to control the reaction and generate high pressure steam. Hence, 0.28 is the correct overall heat transfer coefficient. The reactor pressure is chosen to be 26 bar from a range of values between 25 bar and 35 bar because the highest amount of ethylene oxide was achieved at that pressure.
Once the operating pressure was fixed, the reaction temperature was determined. Based on the sensitivity results, the optimal reaction temperature was found to be 240oC due to the fact that the most ethylene oxide is produced at that temperature. The reactor length and diameter dictate residence time, conversion, and capital cost. The dimensions need to be carefully selected so that the reactor has enough volume for the reaction to run its course. Any excess volume is wasteful as it increases both capital and catalyst costs. It would be optimal to maximize reactor surface area while minimizing volume. A conversion of 57.5% and a residence time of 73 seconds was reached based on the specifications I had just said.
There’s obviously more details that go along with this such as number of kinetics. But basically, that’s been the approach in designing this reactor. It’s difficult to optimize four parameter at once, so that’s the most logical approach that I could come up with.