Yoann Buratti

Ph. D. Candidate
Currently writing my Ph. D. thesis about Machine learning powered solar cell characterization at UNSW

Deep learning on solar cell luminescence images

Dear community, 

I've been working over the past three years in applying machine learning to solar cell characterisation with industry partner's datasets in order to understand, identify faults and improve the performance and reliability of solar cells at the manufacturing level.
eg. on luminescence imaging dataset (basically the reverse ability of a solar cell to convert electricity into light, capturing that light in an image gives insight on the cell's performance): https://onlinelibrary.wiley.com/doi/full/10.1002/pip.3484

As I wrap up my Ph.D. thesis on the subject, I'd like to ask for feedback from the community on:

1/ For IP-protected datasets, I've found sharing the trained models' weights to be a nice workaround to allow other research groups to compare their results and use the models, any other workaround?

2/ Artificial intelligence in manufacturing for monitoring and optimisation seems like a way forward in the solar industry and probably other industries. It's still very early for a full industry 4.0 type of manufacturing. How do you envision this going forward? Any interesting read or already fully AI-controlled manufacturing plant you know about?

3/ I'd welcome any feedback on my recently published papers, please don't hesitate to contact me!

And on a final note, it's great to see the noble prize in physics in climate change science, without which this community wouldn't exist. 
Thanks for reading!
Like Comment