Scholarlink

In a smart manufacturing setting, maintaining the stability and uninterrupted operation of machinery and equipment is a fundamental goal of the production process. Historically, equipment dependability has been preserved by monitoring usage duration and compliance with operational schedules to reduce the likelihood of unforeseen system breakdowns. Recent breakthroughs in artificial intelligence algorithms have proposed several machine learning and deep learning models as viable methods for evaluating equipment’s remaining usable life. This essential method enables predictive maintenance. This research utilizes a Deep Convolutional Neural Network (DCNN) model to forecast the Remaining Useful Life (RUL) of motors in the aerospace sector. Moreover, improving the transparency and dependability of deep learning models is essential for assuring trustworthy decision-making. Consequently, Explainable Artificial Intelligence (XAI) methodologies are employed to enhance the interpretability of the DCNN model. The efficacy of the suggested strategy is shown using a reputable dataset, namely the C-MAPSS dataset by NASA.

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

Deep CNN for Remaining Useful Life Prediction: An XAI Approach Using SHAP for Model Interpretation

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Le Hoang Nguyen,
  • Quoc-Thông Nguyen,
  • Kim Duc Tran,
  • Huu Du Nguyen,
  • Sébastien Thomassey,
  • Xianyi Zeng,
  • Kim Phuc Tran

摘要

In a smart manufacturing setting, maintaining the stability and uninterrupted operation of machinery and equipment is a fundamental goal of the production process. Historically, equipment dependability has been preserved by monitoring usage duration and compliance with operational schedules to reduce the likelihood of unforeseen system breakdowns. Recent breakthroughs in artificial intelligence algorithms have proposed several machine learning and deep learning models as viable methods for evaluating equipment’s remaining usable life. This essential method enables predictive maintenance. This research utilizes a Deep Convolutional Neural Network (DCNN) model to forecast the Remaining Useful Life (RUL) of motors in the aerospace sector. Moreover, improving the transparency and dependability of deep learning models is essential for assuring trustworthy decision-making. Consequently, Explainable Artificial Intelligence (XAI) methodologies are employed to enhance the interpretability of the DCNN model. The efficacy of the suggested strategy is shown using a reputable dataset, namely the C-MAPSS dataset by NASA.