Prasetyo, Moh. Bayu Aji (2024) Eksplorasi Deteksi Anomali dan Prediksi Kegagalan Equipment untuk Predictive Maintenance (PdM) Menggunakan Machine Learning (Studi Kasus Raw Mill pada PT X). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5010201030-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5010201030-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy |
Abstract
Sistem raw mill merupakan sistem krusial dalam produksi semen di PT X. Saat ini, perusahaan menggunakan sistem pemeliharaan berbasis preventif dan terjadwal. Terdapat potensi pengoptimalan sistem pemeliharaan dengan menggunakan sistem predictive maintenance (PdM) menggunakan machine learning (ML) untuk menganalisis kegagalan sistem serta prediksinya di masa depan berbasis data pembacaan sensor yang ditempatkan pada mesin. Penelitian ini berfokus pada perancangan sistem PdM melalui tiga tahapan, yaitu tahap diagnostik, prognostik, dan perancangan arsitektur pendukungnya. Tahapan diagnostik dilakukan dengan deteksi anomali berbasis K-Means clustering dan klasifikasi anomali berbasis artificial neural network (ANN) dengan tujuan untuk melakukan pemetaan dan klasifikasi potensi kegagalan. Pada proses prognostik, prediksi kegagalan sistem dilakukan dengan model hibrida convolutional neural network (CNN) dan long-short term memory (LSTM) untuk memetakan dan merekonstruksi pola dependensi temporal data pembacaan sensor. Sebagai pendukung, arsitektur penunjang dibentuk untuk melakukan analisis pemodelan maintenance secara komprehensif melalui penjadwalan pemeliharaan, serta mengintegrasikan keseluruhan fungsi PdM dengan melakukan perancangan dashboard PdM yang komprehensif. Hasilnya, model PdM berbasis ML ini mampu melakukan pemetaan dan klasifikasi anomali atau kegagalan dengan baik. Namun demikian, model kurang baik dalam merekonstruksi pola atau status anomali sistem raw mill pada tahap prediksi, dengan galat yang cukup tinggi antara data aktual dan prediksi.
==========================================================================================================
The raw mill system is a crucial system in cement production at PT X. Currently, the company uses a preventive and scheduled-based maintenance system. There is potential to optimize the maintenance system by using a predictive maintenance (PdM) system using machine learning (ML) to analyze system failures and their future predictions based on sensor reading data placed on the machine. This research focuses on designing a PdM system through three stages, namely the diagnostic, prognostic, and supporting architecture design stages. The diagnostic stage is carried out by K-Means clustering-based anomaly detection and artificial neural network (ANN)-based anomaly classification with the aim of mapping and classifying potential failures. In the prognostic process, system failure prediction is performed with a hybrid convolutional neural network (CNN) and long-short term memory (LSTM) model to map and reconstruct temporal dependency patterns of sensor reading data. As a support, the supporting architecture is formed to perform comprehensive maintenance modeling analysis through the maintenance scheduling, as well as integrating the entire PdM function by designing a comprehensive PdM dashboard. As a result, this ML-based PdM model can map and classify anomalies or failures well. However, the model is not good at reconstructing the pattern or anomaly status of the raw mill system at the prediction stage, with a high error between actual data and predictions.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Deteksi Anomali, Machine Learning, Prediksi Kegagalan, Predictive Maintenance, Raw Mill, Anomaly Detection, Failure Prediction |
| Subjects: | Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > TS Manufactures > TS174 Maintainability (Engineering) . Reliability (Engineering) |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis |
| Depositing User: | Moh. Bayu Aji Prasetyo |
| Date Deposited: | 04 Aug 2024 15:46 |
| Last Modified: | 04 Aug 2024 15:46 |
| URI: | http://repository.its.ac.id/id/eprint/109763 |
Actions (login required)
 |
View Item |