PV DETECTOR

#4 PV DETECTOR

Using Artificial Intelligence to assess the quality and performance of solar panels.


THIS PROJECT IS PRESENTED BY EDF GROUP TEAM

Sara Benyakhlef

Sara Benyakhlef EDF Renewables

Eric SANDRE

Eric SANDRE EDF Renewables

Julien DUPUIS

Julien DUPUIS R&D

In a word

An innovative and efficient AI system has been developed to automate the quality-control of solar panels.

The project

EDF Renouvelables has risen to the challenge of combining digitisation, rationalisation and energy transition.

And the result is DeteCtor. Through the use of algorithms and the application of artificial intelligence to image processing, DeteCtor outperforms the human eye and brain. It assesses the quality of solar plants more accurately, more quickly and more cost-effectively.

Designed to be used as a benchmark by the Group’s operating and maintenance subsidiaries, it is a big step towards digitisation of the solar-plant quality-control process. To date, it is the only system that uses artificial intelligence whilst conducting a two-tier assessment: photovoltaic cells and modules. It is able to detect faults with a precision of 98.4%.

The system assesses the condition of solar panels in operation and prior to installation on site, hence the significant operational and financial benefits for the Group.

The PV DeteCtor team is now preparing for the large-scale manufacture and roll-out of this system.

Why is this a game changer ?

98,40%
Precision with which the system classifies faults on PV panels.
4
EDF Renouvelables subsidiaries already have the system. The aim is now to roll it out more widely.
8 seconds
The time currently required to process one module as against 10 minutes for the human brain/eye (60 times faster than the human).
A strong contribution to the EDF Group’s strategy
A digital and technological innovation is being used to support successful completion of the energy transition.

EDF Pulse Internal Awards Calendar

1

6 January 2020

Launch of the 7th edition of the EDF Pulse Internal Awards.
2

31 January 2020

Registration closes. 330 files received.
3

11 March – 13 March 2020

Selection of the 24 finalist projects by the selection committees.
4

23 March 2020

Announcement of the 24 finalist teams
5

2 November – 1st December 2020

Employees vote for their favorite project.
6

12 January - 13 January 2021

Finalists coaching.
7

13 January 2021

Auditions before the members of the Group's Executive Committee. Awards ceremony.
8

13 January 2021

Revelation of the 7 winners!