Perancangan Model Predictive Control (MPC) Optimasi Level dan Tekanan pada Separator Minyak-Gas di PT. Pertamina Hulu Mahakam

Pramubudi, Irham (2026) Perancangan Model Predictive Control (MPC) Optimasi Level dan Tekanan pada Separator Minyak-Gas di PT. Pertamina Hulu Mahakam. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009221069_Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (14MB)

Abstract

Penelitian ini bertujuan merancang sistem Model Predictive Control (MPC) untuk mengendalikan level cairan dan tekanan gas pada separator minyak-gas serta membandingkan kinerjanya dengan PID. Model plant dibangun menggunakan data operasi berbasis time series dengan pendekatan state-space MIMO 2 × 2. Identifikasi dilakukan menggunakan N4SID sebagai initial model yang kemudian disempurnakan dengan Prediction Error Method (PEM). Model final memiliki orde empat dengan input Pressure_MV dan Level_MV, serta output Pressure_CV dan Level_CV. Hasil validasi menghasilkan fit pressure sebesar 78,659%, fit level sebesar 78,645%, dan fit rata-rata sebesar 78,652%. Model memenuhi kriteria stabilitas diskrit dengan nilai mutlak maksimum eigenvalue sebesar 0,994345754 serta memiliki rank controllability dan rank observability sebesar 4 dari 4. Simulasi dilakukan pada kondisi tanpa gangguan dan dengan gangguan sebesar 5%. Pada PID, pressure mampu mendekati setpoint tanpa gangguan, tetapi level masih berosilasi dan kedua variabel tidak mampu pulih setelah gangguan. MPC menghasilkan galat akhir mendekati nol serta mampu memulihkan pressure dan level masing-masing dalam waktu 150 detik dan 10 detik. Secara keseluruhan, MPC memberikan pelacakan setpoint dan penolakan gangguan yang lebih baik serta mendukung SDG 9 melalui pengembangan inovasi sistem kendali dan otomasi proses industri.
==================================================================================================================================
This study aims to design a Model Predictive Control (MPC) system for controlling liquid level and gas pressure in an oil-gas separator and to compare its performance with PID. The plant model was developed using time-series operational data through a 2×2 MIMO state-space approach. Identification was conducted using N4SID as the initial model, which was subsequently refined using the Prediction Error Method (PEM). The final model has an order of four with Pressure_MV and Level_MV as inputs and Pressure_CV and Level_CV as outputs. The validation results produced a pressure fit of 78.659%, a level fit of 78.645%, and an average fit of 78.652%. The model satisfies the discrete stability criterion with a maximum absolute eigenvalue of 0.994345754 and has controllability and observability ranks of 4 out of 4. Simulations were conducted under disturbance-free conditions and with a 5% disturbance. Under PID control, pressure approached the setpoint without disturbance, but level remained oscillatory, and neither variable recovered after the disturbance. MPC produced near-zero final errors and restored pressure and level within 150 seconds and 10 seconds, respectively. Overall, MPC provided better setpoint tracking and disturbance rejection while supporting SDG 9 through innovation in industrial control systems and process automation.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Separator minyak-gas, Model Predictive Control, PID, state-space, MIMO, N4SID, PEM, SDG 9, Oil-gas separator, Model Predictive Control, PID, state-space, MIMO, N4SID, PEM, SDG 9
Subjects:	T Technology > TJ Mechanical engineering and machinery > TJ217.6 Predictive Control T Technology > TP Chemical technology > TP155.7 Chemical processes.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Irham Pramubudi
Date Deposited:	31 Jul 2026 00:53
Last Modified:	31 Jul 2026 00:53
URI:	http://repository.its.ac.id/id/eprint/140172

Actions (login required)

View Item