Production forecasts for wind and solar farms, every five minutes
Machine-learning production forecasts for a wind farm and a solar farm, computed at the edge and served to the market operator's API.
Challenge
In the Australian National Electricity Market supply and demand are matched every five minutes, so generators must report production forecasts at the same cadence. Accurate wind and solar forecasts are hard, and failure to report can cost an average-sized farm AUD 300 thousand to 1 million in penalties.
Solution
In partnership with Worley and Monash University, Arundo developed a forecasting algorithm for wind and solar production. Arundo Edge connects to the farms' assets, ingests the data and runs the algorithm, exposing results through an API to the market operator.
Impact
Up to AUD 1 million in grid penalties avoided by reporting to the market operator accurately and on time. Deployed on one wind farm of 43 turbines and one solar farm of 32 inverters, reporting every five minutes.