Sistem Deteksi Kerusakan Bearing Pada Motor Induksi Tiga Fasa Menggunakan Metode Convolutional Neural Network

Imaninsa, Adelia Dwi (2022) Sistem Deteksi Kerusakan Bearing Pada Motor Induksi Tiga Fasa Menggunakan Metode Convolutional Neural Network. Diploma thesis, Institut Teknologi Sepuluh Nopember.

Text 10311810000039-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (2MB)

Abstract

Salah satu bagian dari motor induksi tiga fasa yang berpotensi untuk mengalami kerusakan adalah bearing. Dari survey yang dilakukan oleh Electric Power Research Institute dan Motor Reability Working Group IEEE ada dua jenis kerusakan yang banyak ditemui, yaitu kerusakan bearing, khususnya pada inner race dan outer race bearing, dengan presentase sekitar 41-44%. Kerusakan bearing dapat menyebabkan terjadinya getaran dan noise berlebihan, peningkatan suhu, serta timbulnya bunga api. Identifikasi kerusakan bearing dapat dilihat melalui banyak parameter seperti getaran, suhu, arus, tegangan dan suaranya. Pada penelitian proyek akhir ini, parameter yang digunakan untuk mengidentifikasi kerusakan dari motor induksi adalah getaran. Data getaran yang diperoleh, diolah menggunakan metode Convolutional Neural Network (CNN). Data dari sensor akan digunakan sebagai proses training dan pengujian sistem, sehingga mampu merepresentasikan kondisi dari bearing. Dari hasil pengujian yang telah dilakukan, akurasi terbesar, yaitu 99,90% dihasilkan menggunakan model CNN dengan konfigurasi hidden layer sebanyak 16. Dari masing-masing konfigurasi layer memiliki akurasi di atas 99% yang mana menunjukkan bahwa model CNN yang digunakan dapat mengklasifikasikan kondisi bearing dengan benar.
==============================================================================================================================
One part of a three-phase induction motor that has the potential to be damaged is the bearing. From a survey conducted by the Electric Power Research Institute and the IEEE Motor Reliability Working Group, there are two types of damage that are commonly encountered, namely bearing damage, especially in the inner race and outer race bearings, with a percentage of around 41-44%. Bearing damage can cause excessive vibration and noise, increase in temperature, and cause sparks. Identification of bearing damage can be seen through many parameters such as vibration, temperature, current, voltage, and sound. In this final project research, the parameter used to identify the damage to the induction motor is vibration. The vibration data obtained were processed using the Convolutional Neural Network (CNN) method. Data from the sensor will be used as a training process and system testing so that it can represent the condition of the bearing. From the results of the tests that have been carried out, the greatest accuracy, which is 99.90%, was generated using the CNN model with 16 hidden layer configurations. Each layer configuration has an accuracy above 99% which indicates that the CNN model used can classify bearing conditions correctly.

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	Deteksi kerusakan, Convolutional Neural Network, Bearing, Motor Induksi Tiga Fasa. Fault detection, Vibration, Convolutional Neural Network, Bearing, Three Phase Induction Motor.
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK3070 Automatic control
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Mr. Marsudiyana -
Date Deposited:	15 Jul 2026 07:33
Last Modified:	15 Jul 2026 07:33
URI:	http://repository.its.ac.id/id/eprint/135077

Actions (login required)

View Item