Scholarlink

Predictive maintenance (PdM) models leverage advanced data analytics and machine learning techniques to predict equipment failures and optimize maintenance schedules, enhancing operational efficiency and minimizing downtime. Maintaining stringent safety and reliability standards is paramount in the aerospace industry, particularly within high-technology sectors such as aeronautics and propulsion systems. This study highlights the transition from traditional preventive maintenance to proactive and predictive maintenance strategies, with a specific focus on the maintenance of painting booth filters—a critical yet underexplored aspect of aerospace maintenance. The objective of this paper is to compare the performance of various machine learning approaches in enhancing the predictive maintenance model for this application. This research employs machine learning (ML) methodologies to advance PdM by accurately predicting filter lifespan and scheduling timely maintenance. Ten algorithms: Random Forest (RF), Gradient Boosting (GB), Ridge Regression (RR), CatBoost (CB), XGBoost (XGB), Light Gradient Boosting Machine (LGBM), Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and CNN-LSTM—were evaluated based on their forecasting performance using metrics such as Mean Square Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Symmetric Mean Absolute Percentage Error (SMAPE). The results demonstrate that Ridge Regression (RR) achieved superior performance, closely followed by MLP and LGBM, despite RR and LGBM exhibiting slightly longer inference times. While CNN demonstrated the shortest inference time, it lacked forecasting accuracy. Consequently, MLP emerged as the optimal model, balancing accuracy and efficiency. This research highlights the potential of integrating ML algorithms in PdM. This study establishes a foundation for future advancements in predictive maintenance technology, underscoring its transformative potential in industrial applications.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Predictive Maintenance in Aerospace: Leveraging Machine Learning for Optimized Filter Lifespan

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Syaimak Abdul Shukor,
  • Jia Hui Tang,
  • Iqraq Kamal,
  • Mohd Faizzuan Mohammad,
  • Zalinda Othman,
  • Shahnorbanun Sahran,
  • Nor Samsiah Sani,
  • Mohd Ridzwan Yaakub,
  • Azizi Abdullah

摘要

Predictive maintenance (PdM) models leverage advanced data analytics and machine learning techniques to predict equipment failures and optimize maintenance schedules, enhancing operational efficiency and minimizing downtime. Maintaining stringent safety and reliability standards is paramount in the aerospace industry, particularly within high-technology sectors such as aeronautics and propulsion systems. This study highlights the transition from traditional preventive maintenance to proactive and predictive maintenance strategies, with a specific focus on the maintenance of painting booth filters—a critical yet underexplored aspect of aerospace maintenance. The objective of this paper is to compare the performance of various machine learning approaches in enhancing the predictive maintenance model for this application. This research employs machine learning (ML) methodologies to advance PdM by accurately predicting filter lifespan and scheduling timely maintenance. Ten algorithms: Random Forest (RF), Gradient Boosting (GB), Ridge Regression (RR), CatBoost (CB), XGBoost (XGB), Light Gradient Boosting Machine (LGBM), Multilayer Perceptron (MLP), Convolutional Neural Network (CNN), Long Short-Term Memory (LSTM), and CNN-LSTM—were evaluated based on their forecasting performance using metrics such as Mean Square Error (MSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Symmetric Mean Absolute Percentage Error (SMAPE). The results demonstrate that Ridge Regression (RR) achieved superior performance, closely followed by MLP and LGBM, despite RR and LGBM exhibiting slightly longer inference times. While CNN demonstrated the shortest inference time, it lacked forecasting accuracy. Consequently, MLP emerged as the optimal model, balancing accuracy and efficiency. This research highlights the potential of integrating ML algorithms in PdM. This study establishes a foundation for future advancements in predictive maintenance technology, underscoring its transformative potential in industrial applications.