Special

Solar Panel Inspection Employs Drone, AI, Automatic Reporting (1) (page 2)

Mosaic abnormalities caused by lightning can be detected

2019/07/24 18:13
Shinichi Kato, Nikkei BP Intelligence Group, CleanTech Labo
Print Page

Generally speaking, there is a delay in the improvement of the efficiency of the analysis process of thermal distribution images for solar panel inspection using drones. Nearly all of the inspections are performed manually and the processes require a lot of time and effort.

When the process is divided into the pre-process, where thermal distribution images of solar panels are shot from the air using infrared cameras by flying drones, and the post-process where obtained images are analyzed for reporting, the efficiency is more advanced for the pre-process up to shooting.

In the services offered by the company, the efficiency has been improved by replacing shooting while walking with shooting from the sky while efforts to automate drone operation and aerial shooting are being made ahead of other companies. Meanwhile, the post-process, where obtained images are analyzed for reporting, has large room for efficiency improvement.

Thermal distribution images obtained by shooting panels in certain ranges from the sky look the same regardless of the place. The location of each image in the panel layout drawing and the situation that caused an abnormality, if any, are generally included in the service menu and are reported. However, they are estimated by human-wave tactics. For example, in the case of a power plant with about 2MW of output, it takes about one week for multiple service providers to carry out these operations alone.

NESIC made it possible to identify the position of each thermal distribution image in the layout drawing, classify the phenomenon that is causing abnormal thermal distribution and determine the degree of urgency of the phenomenon by automatic analysis by utilizing the AI and other knowhow of the NEC Group. Reports submitted to customers were also automated.

Following the automation, the time required for acquisition and analysis of thermal distribution images, identification of abnormal panel positions and situations and preparation of reports can be reduced to one-fourth, compared with the time required for carrying out the pre-process and post-process manually, while the cost can be reduced by about 60%.

The cost per inspection from aerial shooting to preparation of a report generally starts from 400,000 yen (approx US$3,702) for mega solar power plants with 2MW output and 1.1 million yen for plants with 10MW output, excluding tax and assuming the topography to be standard.

Energy Solutions (Chiyoda-ku, Tokyo) also announced that it commercialized automatic inspection services including the post-process.