Gearbox Condition Data for Fault Detection via Dynamic PCA

An extract from the gear database of the Salesian Polytechnic University (UPS) used in a research article on early fault detection. The data includes measurements from a gearbox in good condition (Phase I) and with low and high severity failures (Phase II). The dataset was published by Antonio Pérez-Torres on figshare in May 2026 and includes an experimental plan and R code for processing.

Use Cases

• Developing fault detection algorithms based on multivariate sensor data from gearboxes.
• Benchmarking dynamic principal component analysis (DPCA) methods for condition monitoring.
• Training models to classify gearbox health states (good, low severity failure, high severity failure).

Strengths

• Includes data for both healthy and faulty gearbox conditions, enabling supervised learning.
• Accompanied by R code for processing, which may aid reproducibility.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file format are unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the specific experimental setup at the source university.

Provenance

Source

Gear database of the Salesian Polytechnic University (UPS).

Collection Method

Experimental data collection from a gearbox test setup, as described in the associated article.

Freshness

Last updated 2026-05-17 01:18:59; freshness should be verified.