# Multi-mode Fault Diagnosis Datasets of Gearbox Under Variable Working Conditions

arXiv: https://arxiv.org/abs/2403.12521 · HTML: https://mlsanity.com/paper/2403.12521/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 19 Mar 2024.

## Built on by (most cited first, 5 of 5)

- [Multi-output Classification using a Cross-talk Architecture for Compound Fault Diagnosis of Motors in Partially Labeled Condition](https://mlsanity.com/paper/2505.24001/) (2025)
- [Rethinking the Role of Operating Conditions for Learning-based Multi-condition Fault Diagnosis](https://mlsanity.com/paper/2506.17740/) (2025)
- [Disentangling Damage from Operational Variability: A Label-Free Self-Supervised Representation Learning Framework for Output-Only Structural Damage Identification](https://mlsanity.com/paper/2604.19658/) (2026)
- [Asymmetric Adaptation-based Real-time Fault Diagnosis Under Transitional Operating Conditions](https://mlsanity.com/paper/2605.24457/) (2026)
- [When Labels Are Scarce: An Oscillatory State Space Model for Vibration Diagnosis](https://mlsanity.com/paper/2609.27411/) (2026)

## Abstract (arXiv)

The gearbox is a critical component of electromechanical systems. The occurrence of multiple faults can significantly impact system accuracy and service life. The vibration signal of the gearbox is an effective indicator of its operational status and fault information. However, gearboxes in real industrial settings often operate under variable working conditions, such as varying speeds and loads. It is a significant and challenging research area to complete the gearbox fault diagnosis procedure under varying operating conditions using vibration signals. This data article presents vibration datasets collected from a gearbox exhibiting various fault degrees of severity and fault types, operating under diverse speed and load conditions. These faults are manually implanted into the gears or bearings through precise machining processes, which include health, missing teeth, wear, pitting, root cracks, and broken teeth. Several kinds of actual compound faults are also encompassed. The development of these datasets facilitates testing the effectiveness and reliability of newly developed fault diagnosis methods.