Modeling, optimizing and control analysis of a debutanizer column using Aspen Plus and Aspen Dynamic

This thesis project is focusing on the modeling, optimization and control analysis of a debutanizer column using Aspen PLUS and Aspen Dynamics. A complex mixture of hydrocarbons contained a different range of hydrogen and carbon from C2 until nC8 was fed into the debutanizer column for the separation process. There are two products coming out from this distillation column; the light-end hydrocarbons (C2-C4) and the heavier-end hydrocarbons (C5+). The C2-C4 became the desired product for debutanizer column which required to be separated from the mixed hydrocarbons. This C2-C4 was removed from distillate stream as an overhead product. Meanwhile, the C5+ was removed from the bottoms stream as a bottoms product. The target of this project was to recover 90% of butane (C4) and maximum 5 mol% of pentane (C5) composition in the distillate stream. This target was achieved at the end of the project by obtaining approximately 91.1% of C4 recovery and 4.039 mol% of C5 in the distillate stream. Therefore, it concluded the recovery of C5 in the bottoms stream was 90.3%. The debutanizer model was firstly constructed in the Aspen PLUS for steady-state simulation which relied on several specifications of the column and the criteria of the process. The simulation of this separation process was designed using rigorous distillation column simulator, RadFrac. A comparison of physical property methods between Peng-Robinson and RK-Soave were investigated by considering the same theoretical stages in each configuration. Then, the final type of property model was selected depending on the lowest offset from industrial data. A sensitivity analysis was performed to simulate the column within a range of the parameter, and an optimization problem was formulated to be solved. The steady-state flowsheet generated in Aspen PLUS was exported into Aspen Dynamics to simulate the column in dynamic simulation. The debutanizer system has multiple input variables to control the multiple output variables. Therefore, the relative gain array (RGA) analysis was calculated based on the steady-state gain obtained from open loop transfer functions to find the best pairing of input-output. The conventional Proportional-Integral (PI) and cascade control were implemented into the debutanizer column and both control required to be tuned. Therefore, a relay auto-tuning in Aspen Dynamics was used to determine the ultimate period (Pu) and ultimate gain (KCU) of each process. Then, the controller parameters could be calculated using Ziegler-Nichols method. The control strategy was carried out to observe the process response towards changes of set-point and to analyze the relationships between the process variables (PV) and manipulated variables (MV). The disturbance rejection was performed to determine the success of established control scheme. At the end of the project, multiple comparisons were made between the results obtained from Aspen PLUS and Aspen Dynamics with the literature papers. Overall, all thesis objectives were completed, and the purpose of the debutanizer column to be simulated in Aspen PLUS and Aspen Dynamics were successful