Perancangan Sistem Kontrol Proses Kolom Distilasi Minyak Mentah Berbasis Anfis

Abstract

In the oil and gas industry, the process of separating a liquid material to increase the value of the product is usually performed by using distillation method. To maintain the stability of distillation process, control algorithm is required to keep the process variable remains at the desired value. The objective of this study is to carried out the design of neuro-fuzzy controller and to observe a neuro-fuzzy controller towards simulation. Neuro-fuzzy system which is used in the study is the Adaptive Neuro Fuzzy Inference System (ANFIS). ANFIS controller is designed using the inverse dynamics method. The study was conducted in a crude distillation unit column in national oil and gas company. To maintain the purity of the product, industry use instrument analyzer but the operational and maintanance cost of the instrument is expensive. Therefore, this study using tray temperature control or product temperature which indicate the purity of the product indirectly. The simulation results show that the inverse dynamics of the ANFIS controller is able to maintain temperature stability. The peformance of controller to tracking reference point at top, middle and bottom temperature distillation coloumn is as follow, 32 minutes, 23 minutes and 23 minutes, and for a time constant, 11 minutes, 8 minutes and 10 minutes. Meanwhile, performance controller for disturbance rejection at temperatures top, middle and bottom temperature distillation column is as follow, for a maximum overshoot, 0.4, 0.5 and 0.5, for a steady state is as follow 45 minutes, 55 minutes and 38 minutes, and for the dumping is as follow 0.21, 0.24 and 0.21.