ABB IRB 1100 Robot Synthetic Fault Data with 215 Features

A dataset of 2,000 synthetic instances simulating multisensory signals from an ABB IRB 1100 industrial robot's six axes. Created by Mohamed Lamine Soudani, it includes 215 explanatory characteristics monitoring electrical, mechanical, thermal, and vibration parameters for predictive maintenance research.

Use Cases

• Training fault detection models based on simulated electrical, mechanical, thermal, and vibration sensor signals.
• Evaluating feature selection methods based on the 215 explanatory characteristics.
• Benchmarking ensemble tree methods against SVM approaches for industrial robot maintenance.
• Developing severity-level classification models based on three levels per defect type.
• Analyzing class-specific performance for defects like overheating, misalignment, and gear wear.

Strengths

• Contains 2,000 simulated instances, providing a structured synthetic dataset.
• Includes 215 explanatory characteristics derived from multisensory signals.
• Demonstrates specific model performance metrics, such as LightGBM achieving an F1-macro score of 0.7625.
• Explicitly validates severity levels for defects, confirming feasibility for graded maintenance decisions.

Limitations

• Data is synthetic, generated via simulation, and requires further validation on real-world industrial datasets.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is known (2,000), but the dataset size is very small (904.0 B), indicating limited scope or compressed format.

Provenance

Source

figshare

Collection Method

Generated via a sophisticated simulation model of the ABB IRB 1100 robot.

Freshness

Last updated 2026-05-22 12:48:59; freshness should be verified.