High-Speed Train Bogie Vibration & Fault Diagnosis

The High-Speed Train Bogie Vibration Dataset for Condition Monitoring and Fault Diagnosis Using Deep Learning provides synthetic vibration signal data generated to simulate the operating conditions of high-speed train bogies. This dataset is designed to aid in the development of deep learning models for condition monitoring and fault diagnosis, allowing researchers and engineers to detect, classify, and localize faults in critical bogie components.

Key Features:
Vibration Signals: The dataset contains vibration signals generated for various bogie components, simulating real-world mechanical vibrations in longitudinal (X), transverse (Y), and vertical (Z) directions. The signals incorporate noise to replicate realistic train operating conditions.

Multiple Fault Conditions:

Dataset 1: Simulates 7 operating conditions, including 1 normal state and 6 fault states (single and mixed faults), across different speeds (80, 120, 140, 160, and 200 km/h).
Dataset 2: Includes 15 operating conditions, with 1 normal state and 14 specific single-fault states, all captured at a fixed speed of 200 km/h.
Data Structure: Each sample contains 486 time-series data points representing the vibration signals from the bogie under different conditions. Each condition (normal and faulty) contains 1000 samples, offering sufficient data for training and evaluation.

Fault Types:

Single Faults: Specific failures of components such as air springs, lateral dampers, antiyaw dampers, wheelsets, and drive motors.
Mixed Faults: Combinations of faults affecting multiple components simultaneously, providing complex fault diagnosis scenarios.
Dataset Composition:
Dataset 1:
7 conditions: 1 normal + 6 fault states.
Different speeds (80-200 km/h) to simulate real-world variability.
1000 samples per condition, each with 486 data points.
Dataset 2:
15 conditions: 1 normal + 14 single-fault states.
Fixed speed (200 km/h) for focused fault analysis.
1000 samples per condition, each with 486 data points.
Applications:
Deep Learning-based Fault Diagnosis: This dataset is ideal for developing and testing deep learning models such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and hybrid architectures, focusing on fault detection, classification, and localization.

Condition Monitoring: The dataset allows for the continuous monitoring of bogie components, enabling predictive maintenance and early fault detection, critical for ensuring the safety and reliability of high-speed trains.

Target Audience:
This dataset is intended for researchers, data scientists, and engineers working in the fields of:

Mechanical condition monitoring
Intelligent fault diagnosis
Railway system safety
Deep learning applications in industrial systems