PERANCANGAN SISTEM PENDETEKSI KERUSAKAN MULTIPLE FAULT PADA KOMPONEN POMPA AIR SENTRIFUGAL DENGAN MENGGUNAKAN PENDEKATAN RESIDUAL ANALYSIS

Anatta, Kelvin (2020) PERANCANGAN SISTEM PENDETEKSI KERUSAKAN MULTIPLE FAULT PADA KOMPONEN POMPA AIR SENTRIFUGAL DENGAN MENGGUNAKAN PENDEKATAN RESIDUAL ANALYSIS. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[img] Text
02311540000091-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only

Download (2MB) | Request a copy

Abstract

Pompa merupakan salah satu instrumen yang banyak digunakan dalam sektor industri, yang dimana 27% penggunaan listrik dialokasikan untuk pompa sehingga apabila terjadi kerusakan yang dapat mengganggu kinerja dan reliability pompa, terutama sampai total breakdown, hal ini pasti akan cukup berbahaya dan sangat merugikan. Oleh karena itu, sistem untuk mendeteksi kerusakan dini (early detection system ) diperlukan untuk maintenance pompa agar tidak sampai mengalami total breakdown. Dalam penelitian ini, dirancang sistem pendeteksi fault pada pompa dengan menggunakan analisa residual. Residual didapat dari membandingkan sinyal dari pengukuran variabel keadaan pompa dengan sinyal estimasi dari hasil model matematis. Adapun dilakukan 50 kali pengambilan data pengukuran dan residual untuk meningkatkan validasi data, dan variabel keadaan yang diamati adalah tekanan, debit, torsi, dan kecepatan. Setelah itu dicari fitur parameter statistik dari nilai residual, dan selanjutnya dibuat model machine learning untuk mendeteksi fault dari fitur yang didapat, yang dilakukan dengan cara membuat model classifier otomatis. Adapun dihasilkan 24 model dengan metode regresi linier yang berbeda, dan classifier terbaik memiliki akurasi prediksi mencapai perkiraan 98.3% dengan menggunakan metode Subspace KNN. Sedangkan, yang terburuk memiliki akurasi 16.7 % dengan menggunakan metode RUS Boosted Trees. =================================================================================================================== Pump is one of the most important and widely used instrument in industrial sector, of which 27% of electricity usage were allocated for pump, which indicated how important pump is, so that if fault or damage happened to pump that can disturbed its performance and reliability, or even total breakdown, this can be dangerous and costly. So,system that can be used for early detection of pump’s fault are needed so that pump can be maintained before experiencing total breakdown. In this research, fault detection system for pump using residual analysis were designed. Residual were obtained by comparing signal from state variabel from signal obtained from mathematical model signal. Data acquisition and residual feature were obtained 50 times to increased data’s vailidity and state variables that were observed were pressure, flowrate, and torque. Then, statistic parameter feature were obtained from residual and then model were made using machine learning for fault detection from obtained feature by making automatic classifier. The were 24 classifier models and the best obtained model has prediction accuracy of 98.3% using Subspace KNN method. The worst obtined model has prediciton accuracy of 16.7% using RUS Boosted Trees method.

Item Type: Thesis (Undergraduate)
Additional Information: RSF 629.895 Ana p-1 • Anatta, Kelvin
Uncontrolled Keywords: Fault Detection, Pompa Sentrifugal, Residual Analysis, Fault, Machine Learning, Fitur Parameter Statistik. ===================================================================================================== Fault Detection, Centrifugal Pump, Residual Analysis, Fault, Machine Learning, Statistic Parameter Feature.
Subjects: T Technology > T Technology (General) > T55 Industrial Safety
T Technology > T Technology (General) > T57.5 Data Processing
T Technology > TJ Mechanical engineering and machinery > TJ910 Electric pumping machinery
T Technology > TJ Mechanical engineering and machinery > TJ919 Centrifugal pumps--Design and construction.
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7870.23 Reliability. Failures
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User: Kelvin Anatta
Date Deposited: 04 Aug 2020 07:42
Last Modified: 17 Nov 2020 12:09
URI: https://repository.its.ac.id/id/eprint/76837

Actions (login required)

View Item View Item