Exploring Dataset Availability and Methods for Enhanced Automated Fault Detection in Induction Motors

Induction motors are widely used in industrial applications due to their reliability, efficiency, and cost-effectiveness. However, unexpected faults still occur, leading to unplanned downtime and increased operational costs. In the absence of proper data, it is not feasible to come up with efficient diagnostic processes. Fault-detection systems are most effective when trained on high-quality and representative datasets. This paper evaluates all publicly available datasets of induction motors and investigates the existing research on signal identification, signal processing and machine-learning methods that can help detect faults. It is the initial attempt to review the datasets on the topic of induction-motor fault detection in the whole and the problems that were reported by authors. We investigated datasets acquired through experimental studies and analyzed sensor-based measurements, focusing on vibration, current, and acoustic modalities. The main identified issues are the lack of real-life data, standardization, selective samples of fault, and less than complete records. Multiple datasets are lab-based, they vary in size, layout and sensor positioning which reduces their applicability to machine-learning. The shortcomings in dataset design are identified in the review and some of the suggestions include open-access data sharing, introduction of multimodal sensors, and improved fault labeling. It offers a starting point on the development of standardized, all-encompassing datasets, which will allow testing of fault detection in induction motors reliably in the future.