Self-Supervised Learning Data Scarcity Fatigue is a M.Tech project topic for Aerospace Engineering. It gives students a clear starting point for research, implementation planning, and documentation.
Self-Supervised Learning Data Scarcity Fatigue Project Details
| Abstract |
The growing volume of data in Prognostics and Health Management (PHM) has led to significant interest in Deep Learning (DL) methodologies, which often demonstrate enhanced accuracy in Remaining Useful Life (RUL) predictions. However, a primary hurdle for DL approaches is the scarcity of extensively labelled datasets, particularly in industrial applications. To mitigate this data labelling challenge, Self-Supervised Learning (SSL), an emerging paradigm within unsupervised learning, is explored. This research investigates the efficacy of pre-training deep learning models using self-supervised methods on unlabelled sensor data to improve RUL estimation, particularly within a Few-Shot Learning (FSL) context where labelled data is limited. The study addresses a fatigue damage prognostics problem, specifically focusing
on RUL estimation for aluminum alloy panels, representative of aerospace structures, experiencing fatigue cracks, utilizing strain gauge data. Synthetic datasets comprising strain data are employed to comprehensively analyze the impact of dataset scale on predictive performance. The findings indicate that self-supervised pre-trained models achieve substantially superior performance compared to traditional supervised learning approaches when faced with data scarcity, offering a robust solution for PHM applications.
|
| Reference Paper |
Self-Supervised Learning for Data Scarcity in a Fatigue Damage Prognostic Problem |
| Domain |
Aerospace Engineering |
| Sub-Domain |
Structures & Systems / Aerospace Structures |
| PDF Download |
Download / View PDF |
| Get Help |
Get Help on WhatsApp
Message: Hi FE, I need help with “Self-Supervised Learning for Data Scarcity in a Fatigue Damage Prognostic Problem” in “Aerospace Engineering”
|
How to Use This Self-Supervised Learning Data Scarcity Fatigue Topic
This resource helps students understand the project idea, reference paper direction, and next step for implementation. Moreover, students can compare this Self-Supervised Learning Data Scarcity Fatigue topic with related M.Tech project topics.
Additionally, the topic can support synopsis preparation, report writing, and academic documentation. Therefore, students should review the linked reference paper first. For more branches and sub-domains, explore the complete Fried Engineers resource library.
