ORCA-VFD Dataset: Reliability, Degradation, and Remaining Useful Life Data for Variable Frequency Drives

ORCA-VFD is a multi-domain dataset for reliability modeling and remaining useful life (RUL) estimation of variable frequency drives (VFDs).The dataset integrates physics-derived sequences, processed fault-injection data, and field-informed degradation patterns to create a unified framework for lifecycle analysis and predictive maintenance research.This Figshare record contains the full synthesized ORCA-VFD lifecycle dataset, including:100,000-hour lifecycle trajectories spanning infant-mortality, useful-life, and wearout phases.Anomaly score sequences derived from physics-informed feature engineering.Core-8 reliability features computed from VFD electrical signatures.Processed versions of physics, fault, and field datasets, transformed into the standardized ORCA-VFD format.Training, validation, and test sets used for remaining useful life model development.Metadata files, including feature definitions and lifecycle documentation.Raw third-party datasets (e.g., Hanke physics data, PMSM fault data) are not redistributed here and are available at their original sources as cited in the ORCA-VFD manuscript.This Figshare package includes only newly created or transformed data, compliant with open data licensing practices.The ORCA-VFD dataset supports research in:predictive maintenancephysics-informed machine learningVFD degradation modelingreliability engineeringRUL predictiondomain adaptation and cross-domain validationeconomic optimization of maintenance actionsThe companion GitHub repository provides the full modeling code, lifecycle synthesis scripts, feature engineering tools, and sample files:https://github.com/gencaddy2/ORCA-VFD