Photovoltaic Fault Detector with Keras and Tensorflow

Model Detection

Fault detection methods for photovoltaic systems are numerous (electrical characterization, visual inspection, ultrasonic inspection, infrared imaging, imaging…). Some methods use appropriate equipment (thermal camera…).

That is what the Rentadrone team, have been developing during these last months, to make an open-source solution to achieve that task.

The models used for detection are SSD SSD: Single Shot MultiBox Detector and YOLOv3 with some improvements and modifications. You can see more about in YOLOv3: An Incremental Improvement

Example of type of data

The images used for the design of this model were extracted by air analysis, specifically: FLIR aerial radiometric thermal infrared pictures, taken by UAV (R-JPEG format). Which were converted into .jpg images for the training of these detection models. Example FLIR image:

Example of output data

Soiling Fault Detector

Affected Cell Detector

Diode Fault Detector

Panel Disconnect Detector

In our GitHub repository, you can find all the pertained models detection models that point out where the panel faults as wells as some examples and how to train & test the models.

All contributions are welcome, whether to add new features, fix existing bugs, or add support.

Want to know more about taking a look Rentadrone.cl, or follow us on GitHub.