The need for AI to optimize dual-use PV installations to extract maximum value

There has been increasing application of solar when combined with shared land use. Current examples are PV systems floating on water, and also dual-use land where agriculture is interwoven with PV panels. There are clearly challenges with these types of implementations in that either the geography makes the installation more difficult, or the activity requires a need to share both the land and the light "harvesting" of photons. Because of these challenges, it is necessary to optimize the implementation of the PV and other land-use to extract meaningful value beyond just a demonstration project. But the diversity and complexity of the configurations do not lend themselves to simple analysis. Given this complexity, it is necessary to develop flexible algorithms which can take multiple variables into account where the data may be from multiple complex sources. This is the very configuration where Advanced Neural Net applications can assist and guide the installation and management of the combined land-use activities.This paper will describe the following: (1) the key advantages of AI/ML, along with the current issues, (2) the need for transferrable algorithms development, and (3) the overall methodology for implementation and extension to the broader PV dual-land use segment. Specific data will be presented on the novel AI/ML proposed. A key attribute with this novel AI/ML is one can scale and transfer the template developed without re-customization. The key point of the paper is that – as with any niche application of PV and especially in a dual-use mode, there are substantial challenges which must be overcome, and the use of AI/ML will be a useful tool to exploit the advantages and elevate the application to a more broadly used mode of PV.