Perancangan Sistem Kontrol Temperatur Model Sistem Identifikasi ARMAX Pada Coil Outlet Furnace Menggunakan Integrasi PID Dengan Logika Fuzzy

Putri, Salsabilla Rizkita (2026) Perancangan Sistem Kontrol Temperatur Model Sistem Identifikasi ARMAX Pada Coil Outlet Furnace Menggunakan Integrasi PID Dengan Logika Fuzzy. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009221011-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Furnace merupakan salah satu peralatan utama pada industri kilang minyak yang berfungsi memanaskan fluida proses hingga mencapai temperatur operasi sebelum memasuki unit berikutnya. Kinerja furnace sangat dipengaruhi oleh kemampuan sistem kendali dalam mempertahankan coil outlet temperature pada nilai setpoint. Namun, pengendali PID konvensional memiliki keterbatasan dalam menghadapi karakteristik furnace yang bersifat nonlinier, memiliki time delay besar, dan inersia termal tinggi. Oleh karena itu, penelitian ini bertujuan merancang sistem kendali temperatur furnace menggunakan kontroler PID Self-Tuning berbasis Fuzzy Logic untuk meningkatkan kestabilan respon, akurasi temperatur, dan kemampuan menghadapi gangguan. Pemodelan plant dilakukan menggunakan metode Auto-Regressive Moving Average with Exogenous Input (ARMAX) berdasarkan data historis operasi Crude Oil Preheating Furnace, dengan fuel gas flowrate (FR040) sebagai variabel manipulasi dan coil outlet temperature (TRC-119) sebagai variabel terkendali. Model ARMAX3211 terpilih dengan nilai best fit sebesar 79,06% dan fit to estimation data sebesar 96,61%, sehingga layak digunakan sebagai representasi plant. Kontroler PID konvensional dan PID–Fuzzy Self-Tuning dirancang menggunakan 147 rule base dengan dua skenario fungsi keanggotaan keluaran yang dibedakan melalui iterasi parameter Ki, lalu diuji tanpa gangguan serta dengan disturbance T_in (tag 11TR060) berupa data riil fluktuasi temperatur umpan crude oil yang diinjeksikan setelah sistem mencapai kondisi tunak. Hasil penelitian menunjukkan PID konvensional memiliki rise time tercepat (230,71 detik) serta steady-state error terendah, baik tanpa gangguan (0,21%) maupun saat disturbance aktif (2,20%). PID–Fuzzy menghasilkan rise time lebih besar (261,81–324,21 detik) dan steady-state error sedikit lebih tinggi (0,32–0,37%) tanpa gangguan; 2,78–3,03% saat disturbance aktif), namun tetap menghasilkan overshoot 0% dan mampu memulihkan temperatur mendekati kondisi tunak semula setelah disturbance meluruh. Seluruh kontroler mempertahankan temperatur tunak pada kisaran 347,71–348,28°C terhadap setpoint 349°C dengan steady-state error di bawah 5% pada seluruh pengujian. Hasil ini menunjukkan PID–Fuzzy Self-Tuning, khususnya Skenario 2, memberikan aksi kendali yang lebih halus dan konservatif sesuai karakteristik termal furnace, sehingga layak diterapkan pada sistem pengendalian temperatur coil outlet furnace.
============================================================================================================================================
Furnace is one of the primary units in petroleum refinery industries that functions to heat process fluids until they reach the required operating temperature before entering subsequent processing units. Furnace performance is highly dependent on the control system's ability to maintain the coil outlet temperature at its setpoint. However, conventional PID controllers have limitations in handling the nonlinear characteristics of the furnace, its large time delay, and high thermal inertia. Therefore, this study aims to design a furnace temperature control system using a Fuzzy Logic-based Self-Tuning PID controller to improve response stability, temperature accuracy, and disturbance-handling capability. Plant modeling was carried out using the Auto-Regressive Moving Average with Exogenous Input (ARMAX) method based on historical operating data from a Crude Oil Preheating Furnace, with fuel gas flowrate (FR040) as the manipulated variable and coil outlet temperature (TRC-119) as the controlled variable. The ARMAX3211 model was selected with a best fit value of 79.06% and a fit to estimation data of 96.61%, making it suitable as the plant representation. The conventional PID and PID–Fuzzy Self-Tuning controllers were then designed using 147 rule bases with two output membership function scenarios distinguished through iteration of the Ki parameter, then tested under no-disturbance conditions as well as with a T_in disturbance (tag 11TR060), consisting of real fluctuation data of the crude oil feed temperature injected after the system reached steady state. The results show that the conventional PID controller achieved the fastest rise time (230.71 seconds) and the lowest steady-state error, both without disturbance (0.21%) and with disturbance active (2.20%). The PID–Fuzzy controller produced a larger rise time (261.81–324.21 seconds) and a slightly higher steady-state error (0.32–0.37% without disturbance; 2.78–3.03% with disturbance active), while still achieving 0% overshoot and successfully recovering the temperature close to its original steady state after the disturbance decayed. All controllers maintained the steady-state temperature within a range of 347.71–348.28°C against a setpoint of 349°C, with a steady-state error below 5% across all tests. These results indicate that the PID–Fuzzy Self-Tuning controller, particularly Scenario 2, provides a smoother and more conservative control action consistent with the furnace's thermal characteristics, making it suitable for application in coil outlet furnace temperature control systems.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	ARMAX, Furnace, Fuzzy Logic, PID Self-Tuning, SDGs 9,ARMAX, Furnace, Fuzzy Logic, PID Self-Tuning, SDGs 9
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions:	Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Salsabilla Rizkita Putri
Date Deposited:	30 Jul 2026 16:53
Last Modified:	30 Jul 2026 16:53
URI:	http://repository.its.ac.id/id/eprint/139678

Actions (login required)

View Item