Low-Voltage Arc Detection and Sensor Placement
IIT Kharagpur, Prof. Aurobindo Routray
2014 - 2015
This project implemented an arc classification algorithm, and optimal arc sensor placement algorithm in an electrical grid.
- Developed an algorithm for low-voltage (230 V, 50 Hz) arc characterization and classification from other spikes using K-means clustering of PSD spectral distance measurements, achieving a high accuracy of >96% on test data collected from the experiment.
- Developed an algorithm to estimate optimal sensor number and location for arc detection using a bipartite graph approach with high accuracy, which showed up to 24.75% cost reduction capability in smaller electrical networks.