L R AS Published on Sunday 10 October 2021 - n° 377 - Categories:PV building
Computer model for fault detection in residential installations
Researchers at the University of Massachusetts Amherst have developed SunDown, a computer model for detecting outages in residential solar installations. The model avoids
instruments and sensors. Instead, it exploits correlations between the power produced by adjacent panels to detect deviations from the normally expected behaviour of a panel.
The model is able to handle simultaneous faults in multiple panels and classify the fault to determine the likely cause. The model is capable of detecting faults with an accuracy of 99%. It can also detect multiple simultaneous faults with an accuracy of 97%.
https://www.pv-magazine.com/2021/10/08/computer-model-for-residential-solar-fault-detection/
PV Magazine of 8 October 2021
Editor's note: This is probably not the first software to detect the deficiencies of a particular panel or part of a panel. A comparison with other software is needed to determine its qualities and defects