| Ref.No: | 61401800 |
| Start date: | 13.03.2024 |
| End date: | 31.01.2026 |
| Approval date: | 13.03.2024 |
| Department: | NAVAL ARCHITECTURE & MARINE ENGINEERING |
| Sector: | MARINE ENGINEERING |
| Financier: | ΕΛΛΑΔΑ 2.0 ΕΘΝΙΚΟ ΣΧΕΔΙΟ ΑΝΑΚΑΜΨΗΣ &ΑΝΘΕΚΤΙΚΟΤΗΤΑΣ, ELLHNIKO IDRYMA EREYNAS KAI KAINOTOMIAS |
| Budget: | 70.000,00 € |
| Public key: | ΨΖΘΞ46ΨΖΣ4-ΔΞ3 |
| Scientific Responsible: | Prof. IOANNIS GEORGIOU |
| Email: | georgiou@central.ntua.gr |
| Description: | THE CURRENT PROPOSAL FOCUSES ON THE DEVELOPMENT OF AN INNOVATIVE AND FULLY AUTONOMOUS DIGITAL AI-BASED PLATFORM FOR REAL TIME DIAGNOSIS OF EARLY-STAGE FAULTS, INCLUDING FAULT DETECTION AND IDENTIFICATION (TYPE / LOCATION / SEVERITY), AS WELL AS REMAINING USEFUL LIFE ESTIMATION (RULE) IN DRIVETRAIN SYSTEMS UNDER VARYING ENVIRONMENTAL AND OPERATING CONDITIONS (EOCS) AND UNCERTAINTY. DRIVETRAINS ARE CRITICAL PARTS OF IMPORTANT ENGINEERING STRUCTURAL SYSTEMS SUCH AS WIND TURBINES, SURFACE AND AERIAL VEHICLES, ROBOTICS, BIG INDUSTRIAL INFRASTRUCTURE AND OTHER. THE CONCEPT OF THE EEDRIVEN IS UNIQUE IN THE SPECIFIC SCIENTIFIC AREA AS FOR THE FIRST TIME PHYSICSBASED MODELS AIDED BY SURROGATE DATA-DRIVEN MODELS WILL BE DEVELOPED FOR THE ACCURATE & REALISTIC DYNAMICS REPRESENTATION OF DRIVETRAIN SYSTEMS USING ONLY A LIMITED NUMBER OF EXPERIMENTS FROM FEW PHYSICAL SENSORS AND ADVANCED MODEL UPDATING TECHNIQUES IN ORDER TO TRAIN STATE-OF-THE-ART MACHINE LEARNING (ML) METHODS – CURRENTLY NUMEROUS MEASUREMENTS AND SEVERAL SENSORS ON A DRIVETRAIN ARE NEEDED – CAPABLE OF ACHIEVING DIAGNOSIS OF EARLY-STAGE FAULTS AND RULE IN COMPLETE DRIVETRAIN SYSTEMS – WITHOUT NON-REALISTIC ISOLATION AND TREATMENT OF THEIR COMPONENTS INDEPENDENTLY – UNDER VARYING EOCS AND UNMEASURABLE OPERATIONAL UNCERTAINTY. |
