Fault Classification Dataset for Two-Level VSI-Fed Induction Motor Drives

1210 records compiled for 22 distinct fault categories in a two-level voltage source inverter driving an induction motor load. The dataset was acquired on an experimental bench and includes features derived from 3-phase current signals via empirical mode decomposition. Raja Singh R published the dataset on figshare under a CC-BY-4.0 license.

Use Cases

• Training ML models for switch fault prediction based on derived features from inverter current signals.
• Benchmarking fault detection algorithms for different switching frequencies and load torque conditions.
• Analyzing the impact of open circuit faults on drive system performance based on experimental data.
• Developing condition monitoring systems for low-power industrial drives based on intrinsic mode functions.

Strengths

• Dataset covers 22 distinct categories, including a Normal condition and 21 specific switch fault conditions.
• Features were derived from 3-phase currents captured under different switching frequencies and load torques.
• 1210 records provide a basis for machine learning model development.

Limitations

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

Provenance

Source

Experimental bench built around a two-level three-phase voltage source inverter.

Collection Method

Data acquired from an experimental setup, with empirical mode decomposition applied to current signals.

Freshness

Last updated 2026-05-16 11:24:25; freshness should be verified.