Perancangan dan Implementasi Extended Kalman Filter untuk Position Sensorless Control Berbasis Field Oriented Control pada Motor Brushless DC Kondisi Ramp-Up

Budianto, Rama Suryansyah (2026) Perancangan dan Implementasi Extended Kalman Filter untuk Position Sensorless Control Berbasis Field Oriented Control pada Motor Brushless DC Kondisi Ramp-Up. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5009221003-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (6MB) | Request a copy

Abstract

Motor Brushless DC (BLDC) semakin banyak digunakan pada kendaraan listrik, Unmanned Aerial Vehicle (UAV), dan otomasi industri karena keunggulan efisiensi dan keandalannya. Teknik Field Oriented Control (FOC) meningkatkan performa kendali torsi motor BLDC namun membutuhkan informasi posisi rotor yang presisi. Pendekatan sensorless control berbasis Extended Kalman Filter (EKF) merupakan solusi yang menjanjikan untuk mengeliminasi sensor mekanis pada poros, namun estimasi posisi pada kondisi ramp-up dari standstill menghadirkan tantangan khusus karena sinyal back-EMF yang lemah pada kecepatan rendah menyebabkan estimator tidak dapat berkonvergensi. Penelitian ini menyajikan perancangan dan implementasi algoritma EKF-FOC untuk position sensorless control motor BLDC pada kondisi ramp-up. Model matematis motor BLDC diturunkan dalam kerangka referensi A-B-C kemudian ditransformasikan ke referensi α-β, dan konstanta BEMF (k_{e}) diidentifikasi menggunakan pendekatan grey-box dua tahap dengan estimasi bersama terhadap offset fasa sensor, menghasilkan k_{e}=0.008565V·s/rad dengan model fit 89.38% dan 89.25% untuk komponen i_{\alpha }dan i_{\beta }. Validasi model terhadap hardware menghasilkan NRMSE arus fasa rata-rata 7.20% dan posisi sudut 0.14%, keduanya memenuhi ambang batas keberterimaan <10%. Analisis observabilitas nonlinier menunjukkan bahwa model BLDC bersifat locally observable hanya ketika {\omega }*{e}\neq 0, sehingga diformulasikan arsitektur switching open loop ke closed loop sebagai strategi ramp-up. Sistem kendali FOC dirancang dengan tiga loop kendali, menghasilkan ess*{\omega } sebesar 0.82% pada implementasi hardware kondisi sensored. Analisis sensitivitas matriks kovariansi process noise (Q) dan measurement noise (R) mengungkap mekanisme kegagalan yang saling melengkapi antara keduanya, serta menemukan bahwa akurasi estimasi dan keberhasilan switching merupakan aspek performa yang independen. Implementasi hardware pada kecepatan 3, 6, dan 9 rad/s menunjukkan bahwa konfigurasi baseline mencapai batas bawah kecepatan operasional pada 5 rad/s (OL) menuju 2 rad/s (CL), yang berhasil didorong menjadi 3 rad/s (OL) menuju 2 rad/s (CL) melalui kombinasi kovariansi silang.
====================================================================================================================================
Brushless DC (BLDC) motors are increasingly used in electric vehicles, Unmanned Aerial Vehicles (UAVs), and industrial automation due to their efficiency and reliability advantages. Field Oriented Control (FOC) improves the torque control performance of BLDC motors but requires precise rotor position information. Extended Kalman Filter (EKF)-based sensorless control is a promising approach for eliminating mechanical shaft sensors; however, position estimation during ramp-up from standstill presents a particular challenge, as the weak back-EMF signal at low speed prevents the estimator from converging. This research presents the design and implementation of an EKF-FOC algorithm for position sensorless control of a BLDC motor under ramp-up conditions. The mathematical model of the BLDC motor was derived in the A-B-C reference frame and then transformed into the α-β reference frame, with the BEMF constant (k_{e}) identified using a two-stage grey-box approach that jointly estimates the sensor's phase offset, yielding k_{e}= 0.008565 V·s/rad with a model fit of 89.38% and 89.25% for the i_{\alpha }and i_{\beta }components, respectively. Model validation against hardware produced an average phase-current NRMSE of 7.20% and an angular-position NRMSE of 0.14%, both meeting the <10% acceptance threshold. Nonlinear observability analysis showed that the BLDC model is locally observable only when {\omega }*{e}\neq 0, leading to the formulation of an open-loop-to-closed-loop switching architecture as a ramp-up strategy. The FOC control system was designed with three control loops, achieving an ess*{\omega }of 0.82% in the hardware implementation under sensored conditions. Sensitivity analysis of the process-noise (Q) and measurement-noise (R) covariance matrices revealed complementary failure mechanisms between the two, and further found that estimation accuracy and switching success are independent performance aspects. Hardware implementation at speeds of 3, 6, and 9 rad/s showed that the baseline configuration reached a lower operational speed limit of 5 rad/s (OL) to 2 rad/s (CL), which was successfully pushed down to 3 rad/s (OL) to 2 rad/s (CL) through a cross covariance combination.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	brushless DC motor, Extended Kalman Filter, Field Oriented Control, Embedded System
Subjects:	Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > QA Mathematics > QA402.3 Kalman filtering. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK1007 Electric power systems control T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK2681.O85 Electric motors, Brushless.
Divisions:	Faculty of Industrial Technology > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Rama Suryansyah Budianto
Date Deposited:	01 Aug 2026 06:25
Last Modified:	01 Aug 2026 06:25
URI:	http://repository.its.ac.id/id/eprint/141665

Actions (login required)

View Item