PV AI

The PV AI tool is a surrogate-model-empowered Rhino/Grasshopper workflow that arms designers with real-time Photovoltaic (PV) generation analysis. 

​

First, I used grasshopper, python, and the PvWatts web API to generate a large database of PV generation analysis for any orientation and climate file. I then trained a shallow artificial neural network regressor on this data to develop a surrogate model that could replace the slow web API with an instantaneous Rhino  plugin. I then built a Rhino/Grasshopper workflow to implement this surrogate within a typical Rhino-based conceptual design workflow.

​

This workflow supports real-time collaborative design charettes between our sustainability and design teams. By the end of a session everyone can understand which design decisions impact PV generation the most. Armed with knowledge, the team can choose a path towards a successful solution.