Rancang Bangun Sistem Deteksi Kesalahan Combined Misalignment Pada Motor Induksi Menggunakan Pengukuran Arus Dengan Metode 2 Dimensional Convolutional Neural Network (2D-CNN)

Al-Karim, Muhammad Iqbal (2023) Rancang Bangun Sistem Deteksi Kesalahan Combined Misalignment Pada Motor Induksi Menggunakan Pengukuran Arus Dengan Metode 2 Dimensional Convolutional Neural Network (2D-CNN). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311940000060-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (5MB) | Request a copy

Abstract

Motor induksi merupakan salah satu jenis motor listrik yang paling banyak digunakan dalam industri modern namun rentan mengalami masalah. Lebih dari 70% masalah getaran mesin berputar disebabkan oleh kesalahan misalignment. Apabila masalah ini dibiarkan akan menyebabkan motor induksi mengalami kerusakan bahkan kegagalan. Pada penelitian tugas akhir ini dibuat sistem deteksi kesalahan combined misalignment yaitu gabungan antara angular dan parallel misalignment. Digunakan pengukuran arus dengan sensor SCT-013 untuk memperoleh data sinyal arus kesalahan misalignment pada motor induksi. Data arus tersebut di-preprocessing dengan metode recurrence plot menghasilkan output gambar. Gambar hasil preprocessing sinyal digunakan sebagai input arsitektur 2 dimensional convolutional neural network (2D-CNN) yang diusulkan. Variasi misalignment yang digunakan adalah normal (tidak ada misalignment), angular misalignment dengan sudut 1o dan 2o, parallel misalignment dengan jarak 2 mm dan 4 mm, serta kombinasi angular dan parallel misalignment (angular 1o, parallel 2 mm; angular 1o, parallel 4 mm; angular 2o, parallel 2 mm; angular 2o, parallel 4 mm). Hasil sistem deteksi dengan arsitektur 2D-CNN yang diusulkan memiliki rata-rata precision sebesar 96,28%; recall sebesar 96,11%; dan F1-score sebesar 96,17%. Secara keseluruhan arsitektur 2D-CNN yang diusulkan memiliki accuracy yaitu 96,29%.
==================================================================================================================================
Induction motors are one of the most widely used types of electric motors in modern industry but it is prone to problems. More than 70% of rotating machine vibration problems are caused by misalignment errors. If this problem is left unchecked, it can lead to damage and even failure of the induction motor. In this final project research, a combined misalignment fault detection system is made, which is a combination of angular and parallel misalignment. Current measurement is used with the SCT-013 sensor to obtain misalignment fault current signal data on the induction motor. The current data is preprocessed with the recurrence plot method to produce an image output. The preprocessed image is used as input to the proposed 2 dimensional convolutional neural network (2D-CNN) architecture. The misalignment variations used are normal (no misalignment), angular misalignment with angles of 1o and 2o, parallel misalignment with distances of 2 mm and 4 mm, and a combination of angular and parallel misalignment (angular 1o, parallel 2 mm; angular 1o, parallel 4 mm; angular 2o, parallel 2 mm; angular 2o, parallel 4 mm). The results of the detection system with proposed 2D-CNN architecture have an average precision of 96.28%; recall of 96.11%; and F1-score of 96.17%. Overall, the proposed 2D-CNN architecture has an accuracy of 96.29%.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Motor Induksi, 2D-CNN, Misalignment, Recurrence Plot, Induction Motor
Subjects:	R Medicine > R Medicine (General) > R858 Deep Learning T Technology > TJ Mechanical engineering and machinery > TJ174 Maintenance and repair of machinery T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK2785 Electric motors, Induction. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK351 Electric measurements.
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Al-Karim Muhammad Iqbal
Date Deposited:	29 Jul 2023 16:21
Last Modified:	29 Jul 2023 16:21
URI:	http://repository.its.ac.id/id/eprint/100345

Actions (login required)

View Item