The renewable energy sector is facing a serious hurdle. How will Australia dispose of 100 million solar panels in an environmentally friendly way when they reach their end of life? The industry is now grappling with the anticipated waste generated by over 100,000 tonnes of solar panels due to be dismantled in Australia from 2035.
PV Industries’ goal is to stop solar panel waste from being sent to landfill. They do this by collecting and recycling as many end-of-life solar panels as possible and supplying the recovered material to local manufacturers. But some solar panels can be reused, rather than sent to landfill, and PV Industries has found a clever way to do this, but not at scale. Yet.
So they engaged Mondo Ventures to find out how.
The process of checking the panels to find out if they need to be recycled or reused is tedious and time-consuming and often inaccurate and expensive. PV Industries needed a way to automate the inspection process to provide rapid recommendations about each solar panel at scale.
PV Industries responded by developing, in partnership with the University of New South Wales (UNSW), an AI algorithm that determines if a solar panel is due for replacement or reuse. This AI model, when given a photo of a single solar panel (often part of a much larger array), reports on whether or not it needs to be recycled, or more importantly, if it can be reused.
Despite being a great automation tool, the AI algorithm could only process one image at a time through a simple standalone interface. This was a highly manual and time consuming process for sites with large arrays of solar panels. PV Industries needed a solution that would be sleek and visually engaging for investors and customers while allowing rapid automated processing of a high volume of solar panels.
The primary aim of this project was for Mondo to integrate into the AI model and build the functions and features required to test panels in a production environment with real users and large solar panel arrays.
The Mondo team did this by developing a bespoke, front-end dashboard allowing users to drag and drop thousands of solar panel images taken from multiple commercial sites. The application can batch analyse each of the images and report if they can be reused, retained or recycled in a matter of minutes.
The dashboard allows users to easily map out different solar array setups, configured for different customers and run the AI analysis in a matter of seconds producing a clear heat map of recommendations for easy action and reporting to customers.
Mondo started the 3-week engagement with two simple outcomes that would define the success of the project:
These two outcomes were delivered to production in only 1.5 weeks! This meant that the team could use the remaining 1.5 weeks to focus on bonus enhancements.
PV Industries and potential customers can now see at a glance the overall performance, at the solar panel level, for each of their locations. Green is good, and the rest can be actioned. This end-to-end process used to take weeks, and now it’s done in a matter of hours.
The application, called Helios, can now deliver more efficient solar arrays for customers, more recycling, and a greener planet for all.
Time savings for a medium to large solar site of 700 panels. Reducing 17.5 working hours (2 days) to under an hour.
% Reduction in human error.
PV Industries needed
Drop us a note and we’ll be in touch as soon as we can be.