Compact models classify electrical faults nearly perfectly with limited data
One-Cycle Fault Classification and Faulted-Line Identification on the PROTECT-90 Dataset: An Initial Application Benchmark
Machine Learning
Summary
Finding electrical faults quickly and accurately in power lines is key to keeping electricity safe and stable. The authors tested small neural networks on a new dataset called PROTECT-90, focusing on detecting fault types and the exact faulty line. Their models achieved almost perfect accuracy using short one-cycle time windows when full data was available. However, accuracy dropped when only voltage data was used, showing data type matters more than the model design. The work sets a benchmark for applying machine learning to electrical fault detection using this dataset.
What this means in practice
- •For power system engineers: Use compact neural networks to rapidly and accurately detect the type and location of faults on high-voltage lines with minimal data latency.
- •For smart grid technology teams: Incorporate the benchmarked models and timing metrics from PROTECT-90 to improve automated protection response under varying sensor availability.
Tested on one dataset.
Authors
Emad Abukhousa, Abdulaziz Qwbaiban, Saman Zonouz, A. P. Sakis Meliopoulos
Abstract
Open electromagnetic-transient datasets are beginning to make reproducible learning-based protection studies possible, but the practical use of these datasets still requires application-level benchmarks that define timing, sensing, and validation assumptions. This paper presents an initial application benchmark on the recently released PROTECT-90 dataset for two protection-oriented tasks: fault-type classification and discrete faulted-line identification. A compact one-dimensional convolutional neural network (CNN) is evaluated using post-inception windows of 0.25, 0.5, 1, and 2 cycles under strict episode-wise splitting. A non-convolutional multilayer perceptron (MLP) is also trained as an architecture-control baseline. The results show that both tasks are nearly saturated under full observability, with one-cycle test accuracies of 99.84% for fault type and 100.00% for line identification. The main performance variation appears under reduced observability: current-only inputs preserve line identification accuracy at 100.00%, whereas voltage-only inputs reduce line identification accuracy to 53.09% with the CNN and 50.57% with the MLP. This indicates that the limiting factor is measurement information rather than neural architecture. Additional stratified checks show stable performance across topology states and fault-resistance bins, while CPU inference contributes only 0.528 ms to the one-cycle total decision time of 20.53 ms.