The growing use of mechatronic systems in Industry 4.0 highlights the need for reliable and interpretable fault diagnosis methods.
These systems, combining sensors, actuators, and dynamic processes, are vulnerable to faults that can affect performance and
safety. This work presents a hybrid fault diagnosis framework tailored for mechatronic systems, integrating digital models with
AI-based fault classification. In the offline stage, a digital replica of the physical system is built using physical laws and system
parameters, where various fault scenarios are injected to generate labeled datasets. These datasets train an ensemble of machine
learning models, including Decision Tree, Support Vector Machine, K-Nearest Neighbors, and Random Forest, using features
from time and frequency domains. In the online stage, the trained models perform real-time fault detection and classification,
supported by a feedback loop for continuous learning. The framework is validated on a UR5e robotic manipulator, demonstrating
high accuracy and robustness across diverse fault scenarios, confirming its suitability for smart industrial systems.

عنوان المؤتمر

7th International Conference on Industry of the Future and Smart Manufacturing

دولة المؤتمر

فلسطين

تاريخ المؤتمر

12 نوفمبر، 2025 - 14 نوفمبر، 2025

راعي المؤتمر

IMDP

معلومات إضافية

موقع المؤتمر الإلكتروني