Mechanical Components Vibration Dataset for Cross-Component Fault Diagnosis

Forgis's Mechanical Components Vibration Dataset is a multi-source collection designed for the Mechanical-JEPA project. It contains approximately 12,000+ samples across 16 sources and 5 component types, totaling 9.5 GB. The dataset was last updated on Hugging Face in April 2026.

Use Cases

• Train fault diagnosis models based on vibration data from multiple mechanical sources.
• Develop prognostics models for predicting component failure based on vibration patterns.
• Benchmark cross-component generalization for condition monitoring algorithms.
• Analyze vibration signatures across five different mechanical component types.

Strengths

• Approximately 12,000+ samples provide a substantial basis for model training.
• Data aggregates from 16 different sources, suggesting potential diversity.
• Covers 5 distinct mechanical component types for cross-component analysis.
• Dataset size of 9.5 GB indicates significant data volume.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Description metadata is limited; actual data quality requires manual inspection after download.
• Row count is an approximation ('~12,000+'), which may limit precise suitability assessment.

Provenance

Source

Forgis on Hugging Face

Collection Method

Aggregated multi-source collection, likely from experimental or industrial setups.

Time Range

null

Freshness

Last updated 2026-04-02 22:15:58; freshness should be verified.

Geography

null