Term 252

AI-Driven Automated Gearbox Failure Diagnostic Rig

Project Type: Self-Initiated

Project Description

This project develops an AI-driven automated gearbox diagnostic rig for a single-stage spur gearbox, designed to classify operating conditions as Healthy, Chipped Tooth, or Missing Tooth. The goal is to detect faults early, reducing repeat failures and costly reinstallation. The system integrates mechanical design, motor control, embedded data acquisition, and machine learning into a unified platform. It features a motor-driven gearbox with controlled load, PID-based speed regulation, and synchronized data acquisition for vibration, speed, and torque measurements. A custom software application performs signal processing, feature extraction, and fault classification, with a graphical interface for inputting operating conditions and visualizing waveform and FFT data. The model has been validated on labeled datasets. Mechanical, electrical, and control subsystems have been analytically designed and partially implemented, including safety features such as guarding and emergency shutdown. Full system integration and experimental validation are currently in progress. Overall, the project demonstrates a multidisciplinary approach to vibration-based gearbox fault diagnosis using AI in a controlled and reproducible testing environment.