Solar photovoltaic fault diagnosis with thermal mapping

When a solar panel stopped producing the expected amount of power, the company decided it was time to take a look with a drone's eye view.

The well-known company contacted drone service company Oblivion Aerial SA to investigate why one of the solar panels on its roof wasn't producing the expected energy amount. Based on the temperature difference in the thermal map, created using a FLIR Vue Pro thermal camera and Pix4Dmapper, exact problems and locations could be noted noticed.

Two thermal images side by side. One has been adjusted, allowing problem areas to stand out more.
Pix4Dmapper-generated thermal map using non-adjusted and adjusted images

Project details

CompanyOblivion Aerial SA.
IndustryThermal Inspection
Project DateOctober 2016
Project ManagerDamiano Maeder
Project Size1,600 Square-meter Rooftop
Number of Images211
GSD~10 cm

Two issues were found in the first inspection. A large region of the rooftop had a higher temperature, which indicated possible connection issues for the inverter. Upon receiving this information, the client decided to shut down the power supply for the detected region and recheck the connectors. This issue normally happens when connectors are oxidized, due to high humidity and low-PH weather conditions.

The other problem was a hot spot. The thermal map helped the company estimate the cause of the hot spot. For example, a large temperature difference between a hot spot and surrounding modules (32 degrees Celsius for this project) may indicate a broken module, while a smaller difference can simply be dirt or soil covering the module.

A thermal image of a purple figure hosing a yellow solar panel.
Cleaning the panels

A second inspection was conducted while the installation was connected to the inverters. There was clearly a misconnection between multiple groups of modules. The issue was resolved by replacing the connectors.

Equipment for a drone's eye view

A quadcopter multi-rotor drone equipped with a FLIR Vue Pro (640) camera, 9mm focal length, was flown manually to acquire NADIR images, which were triggered every second.

A drone ready for takeoff
Drone with Gimbal airframe for aerial photography use

The first inspection was conducted with 211 images with a rough resolution of 10 centimeters. However, the second inspection included only 11 images, which took less than five minutes to process. Despite being so quick to process, the results gave the client sufficient information for the quick examination after the replacement work.

Fully powered

From the first inspection result, we can see some dark regions, which were affected by the camera calibration during the flight. Each image was manually adjusted for constant reflectance values. Along with some on-site temperature sampling, a high accuracy thermal map was created in Pix4Dmapper.

A thermal drone image of a roof, showing the location of each image captured.
Inconsistency in values on thermal images due to camera drifting and re-calibration during the flight

For this project, some issues were found apart from the major connection problem. An overheated cell indicates a possible damage on the north-west, and some scattered dirt coverage which resulted in minor temperature rise.

A small hot spot due to shading from stock dirt
hotspot appearing on thermal images and the physical damage on the glass
A hotspot found on thermal images, the cells appear overheated and cracked (left) and a visual inspection will confirm the physical breakdown of the module (right).

This gives information on location and condition of solar modules. The output can be overlaid with other existing building data for further analysis, and archived for repetitive monitoring.

Why take to the skies

The most obvious advantages of thermal mapping for this project is the speed and flexibility of response, such as hot spot detection, and hot spot type estimation, based on the information thermal images provide. This information cannot be gleaned from color images. The ease of locating problematic modules over a huge area is an obvious benefit, over a large area of modules that look exactly the same.

Thanks to the Pix4Dmapper, it is now easy to orientate through an entire installation without the risk of having the wrong location of malfunctioning modules or problematic components.

thermal inspection of thermal panels
Incorrect reflectance and temperature values for facade on oblique image
Discover Pix4Dmapper
Map purely from images.

Related articles