Perancangan Sistem Deteksi Kesalahan Pompa Torak Berdasarkan Ekstraksi Ciri Arus Motor, Laju Aliran Dan Tekanan Fluida

Kaltsum, Ummu (2023) Perancangan Sistem Deteksi Kesalahan Pompa Torak Berdasarkan Ekstraksi Ciri Arus Motor, Laju Aliran Dan Tekanan Fluida. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 02311640000109-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2025. Download (3MB) | Request a copy

Abstract

Peningkatan keamanan operasi dengan deteksi kesalahan sebagai bagian dari predictive maintenance khususnya pada pompa torak sebagai salah satu peralatan mekanik yang memiliki peranan penting di industri. Lingkungan kerja yang keras seperti offshore dengan tekanan dan temperature tinggi, mudah terbakar dan meledak, serta rawan korosi membuat pompa torak memiliki berbagai jenis kesalahan seperti kerusakan pada komponen, kebocoran serta kesalahan pada bearing. Tujuan dari penelitian ini adalah mengetahui pengaruh kesalahan pada performansi pompa torak dan melakukan perancangan sistem deteksi kesalahan dengan menjaga performansi pompa torak berbasis ekstraksi ciri. Perancangan sistem deteksi kesalahan ini menggunkan software Matlab R2022a dengan data simulasi pada kondisi healthy dan faulty dari pemodelan pompa torak. Data hasil simulasi dipreproses dengan reduksi data dan transformasi fourier sebelum masuk ke esktraksi ciri. Hasil perancangan sistem deteksi kesalahan berbasis ekstraksi ciri arus motor, laju aliran dan tekanan dengan 9 parameter domain waktu dan 5 parameter domain frekuensi serta classifier Support Vector Machine (SVM) dengan fungsi kernel Polynomial dan K-Fold Cross Validation iterasi lima kali memiliki nilai akurasi baik yaitu 91,8%
====================================================================================================================================
Improving the safety of operations by detecting errors as part of predictive maintenance, especially in reciprocating pumps as one of the mechanical equipment that has an important role in the industry. Harsh working environments such as offshore with high pressure and temperature, flammable and explosive, and prone to corrosion make thoracic pumps have various types of errors such as damage to components, leaks and errors in bearings. The object of this study is to determine the influence of errors on thoracic pump performance and design an error detection system by maintaining the performance of thoracic pumps based on extraction characteristics. The design of this error detection system uses the Matlab R2022a software with simulated data in healthy and faulty conditions from thoracic pump modeling. The simulated data is preprocessed by data reduction and fourier transformation before entering the characteristic extraction. The results of the design of an feature extraction based fault detection system for motor current, flow rate and pressure with 9 time-domain parameters and 5 frequency-domain parameters and a Support Vector Machine (SVM) classifier with Polynomial kernel functions and K-Fold Cross Validation iteration five times have a good accuracy value its 91,8%

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Pompa Torak, Ekstraksi Ciri, classifier, Support Vector Machine (SVM), Arus Motor, Laju Aliran, Tekanan, Reciprocating Pump, Feature Extraction, Classifier, Support Vector Machine (SVM), Motor Current, Flow Rate, Pressure
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T57.62 Simulation T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Industrial Technology and Systems Engineering (INDSYS) > Physics Engineering > 30201-(S1) Undergraduate Thesis
Depositing User:	Ummu Kaltsum
Date Deposited:	30 Jan 2023 08:46
Last Modified:	30 Jan 2023 08:46
URI:	http://repository.its.ac.id/id/eprint/95807

Actions (login required)

View Item