Implementasi Condition Based Maintenance Dengan Pendekatan Data Mining Untuk Prediksi Waktu Kegagalan Equipment Kritis Mesin Raw Mill (Studi Kasus: PT ABC)

Riantama, Rafif Nova (2020) Implementasi Condition Based Maintenance Dengan Pendekatan Data Mining Untuk Prediksi Waktu Kegagalan Equipment Kritis Mesin Raw Mill (Studi Kasus: PT ABC). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 02411640000002-Undergraduate_Thesis.pdf]
Preview
Text
02411640000002-Undergraduate_Thesis.pdf

Download (3MB) | Preview

Abstract

Raw mill merupakan salah satu mesin utama dalam proses produksi semen pada unit V PT ABC. Tercatat bahwa availabilitas raw mill pada tahun 2015 – 2019 berada pada rentang 72,64% hingga 75,66%. Angka tersebut jauh di bawah target availabilabilitas perusahaan 90%. Perusahaan telah menerapkan preventive maintenance dengan aktivitas overhaul dan condition-based maintenance (CBM) melalui pemantauan kondisi equipment secara real time menggunakan sensor, namun pemeliharaan dilakukan ketika equipment mengalami kegagalan yang berpotensi menyebabkan tingginya unplanned downtime. Terdapat potensi untuk mengoptimalkan sistem pemeliharaan dengan menerapkan prinsip prognosis pada strategi CBM untuk prediksi waktu kegagalan equipment melalui perhitungan Mean Residual Life (MRL). Penelitian ini bertujuan untuk membuat dan menentukan model terbaik dalam prediksi MRL equipment pada mesin raw mill. Selain itu dashboard pemeliharaan akan dibangun untuk menunjang penerapan praktis di lapangan. Pendekatan regresi dalam data mining digunakan untuk prediksi MRL mesin. Model data mining yang digunakan adalah model Artificial Neural Network (ANN) dan Support Vector Regression (SVR). Parameter pembanding yang digunakan untuk penentuan model terbaik adalah Root Mean Square Error (RMSE) dan standar deviasi. Wilcoxon test, digunakan untuk membuktikan model terpilih layak untuk digunakan sebagai model prediksi. Selain itu digunakan power BI software untuk membangun dashboard pemeliharaan. Hasil menunjukkan bahwa tidak ada model yang dinyatakan layak untuk digunakan sebagai model prediksi kegagalan (MRL) equipment mesin raw mill, ditunjukkan dengan hasil Wilcoxon test lebih kecil dari 0,05. Adapun ketidaklayakan tersebut disebabkan data sensor tidak menunjukkan adanya degradasi fungsi pada equipment.
====================================================================================================================
Raw mill is one of the important machines in cement production process at unit V of PT ABC. The availability of raw mill in 2015 – 2019 are recorded in the range of 72.64% to 75.66%, which is far below the company's target of 90%. Company has implemented condition-based maintenance (CBM) using sensor to monitor machine condition, however maintenance done when failure occurs which potentially lead to low availability. Preventive maintenance also done through overhaul activity. Prognosis process can be applied to optimize the CBM strategy to predict equipment’s failure time through Mean Residual Life (MRL) calculation. The purposes of this study are to build and determine the best model to predict MRL of raw mill equipment. Moreover, maintenance dashboard will be developed to support practical implementation. Regression approach using Artificial Neural Network (ANN) and Support Vector Regression (SVR) model can be used to predict machine’s MRL. Comparison parameters used to determine the best model are Root Mean Square Error (RMSE) and standard deviation. Wilcoxon test help to prove the chosen model are valid to be used as predictive model. Moreover, power BI software used to design maintenance dashboard. This research shows there is no model that feasible to be used as failure prediction model on raw mill machine’s equipment, indicated by Wilcoxon test result smaller than 0,05. The model’s impropriety caused by the sensors used in this research do not indicating function’s degradation process on equipment.

Item Type: Thesis (Other)
Uncontrolled Keywords: Raw mill, Mean Residual Life (MRL), Condition based maintenance (CBM), Artificial Neural Network (ANN), Support Vector Regression (SVR)
Subjects: Q Science > QA Mathematics > QA76.9.D343 Data mining. Querying (Computer science)
T Technology > TH Building construction > TH3351 Maintenance and repair
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User: Rafif Nova Riantama
Date Deposited: 27 Aug 2020 01:17
Last Modified: 23 Jun 2023 07:54
URI: http://repository.its.ac.id/id/eprint/80005

Actions (login required)

View Item View Item