Publications Scientifiques

Permanent URI for this communityhttps://dspace.univ-boumerdes.dz/handle/123456789/10

Browse

Subject: search.filters.subject.Remaining Useful Life (RUL)

Search Results

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Interpreting NAS-Optimized Transformer Models for Remaining Useful Life Prediction Using Gradient Explainer
    (Warszawa: Polskie Towarzystwo Informatyczne, 2025) Nekkaa, Messaouda; Abdouni, Mohamed; Boughaci, Dalila
    Remaining Useful Life (RUL) estimation of complex machinery is critical for optimizing maintenance schedules and preventing unexpected failures in safety-critical systems. While Transformer architecture has recently achieved state-of-the-art performance on RUL benchmarks, their design often relies on expert tuning or costly Neural Architecture Search (NAS), and their predictions remain opaque to end users. In this work, we integrate a Transformer whose hyperparameters were discovered via evolutionary NAS with a gradient-based explainability method to deliver both high accuracy and transparent, perprediction insights. Specifically, we adapt the Gradient Explainer algorithm to produce global and local importance scores for each sensor in the C-MAPSS FD001 turbofan dataset. Our analysis shows that the sensors identified as most influential, such as key temperature and pressure measurements, match domain-expert expectations. By illuminating the int ernal decision process of a complex, NAS-derived model, this study paves the way for trustworthy adoption of advanced deep-learning prognostics in industrial settings.
  • No Thumbnail Available
    Item
    Turbofan Engine RUL Prediction using ICA and Machine Learning Algorithms
    (Institute of Electrical and Electronics Engineers Inc, 2023) Aribi, Yacine; Boutora, Saliha; Boushaki Zamoum, Pr. Razika; Menasria, Hafid; Abdellaoui, Abdelkader; Kouzou, Pr. Abdellah
    This paper takes an approach to the determination on the Remaining Useful Life (RUL) on a real-life turbo engine model selected from a set of data provided within the public domain for research purposes from the Prognostics Data Repository of NASA. The RUL analysis algorithm uses Independent Component Analysis for data dimensionality reduction and data processing simplicity due to the large number of involved sensors, then a model is trained to predict the remaining useful life for the turbo engine using Random Forest (RF) and Gradient Boosted Machine Algorithms (GBMA). The final RUL data is compared to the real RUL vs. time provided within the original data date for algorithm validation.
  • No Thumbnail Available
    Item
    Remaining Useful Life Prediction in Turbofan Engines: PCA and Machine Learing Approach
    (Institute of Electrical and Electronics Engineers Inc, 2023) Boutora, Saliha; Aribi, Yacine; Boushaki Zamoum, Pr. Razika; Kouzou, Pr. Abdellah; Menasria, Hafid; Abdellaoui, Abdelkader
    In this paper, a study of prediction of the Remaining Useful Life (RUL) of an aircraft's turbofan engine is explored by analyzing a data set of a real-life turbofan engine from the Prognostics Data Repository of NASA and using the Principal Component Analysis (PCA) for data reduction and preprocessing, then selecting machine learning algorithms, mainly the Random Forest (RF) and Gradient Boosted Machine (GBM) so that a model can be trained to predict the possible failures through developing a model to estimate the RUL of the functionality of the turbofan engine