Irradiance Estimates and Forecasts via API, Globally: From Resource Assessment to Storage Management with Sky Cameras

In conjunction with the upcoming annual PV Operations USA conference on November 06 & 07, we sat down with Chief Technology Officer at Solcast, Nick Engerer. Last chance to register! Limited passes remain, secure your place here today!

Could you describe some specific examples of when & why solar forecasts are needed in solar farm operations? 

Solar farm operators are increasingly being required by the local ISO or energy market operator to submit power output (MW) or energy yield (MWh) forecasts as a part of their participation in that energy market. These vary from minutes to hours to day-ahead, depending on the local market rules and requirements. For example, 5-minute ahead instantaneous power forecasts are being submitted to the Australian Energy Market Operator in Australia. 15 minute ahead forecasts (and soon to be 5 or 10 minutes ahead) forecasts are also now required in India. Vietnam, Taiwan, the Phillipines and other jurisdictions are likely to follow. Even Ercot in Texas is looking into this method of managing utility-scale solar (and wind) in their system. 

How does Solcast approach the solar forecasting challenge for utility scale sites?

Solcast combines three key pieces of information. First, we have our global, rapid-update solar forecasting system that tracks cloud cover across the globe, creating new forecasts every time we get a new satellites scan (every 5-15 minutes). We use this to produce a first-guess power output forecast at the site. Second, we gather both historical and real-time SCADA readings of the solar farm to locally ’tune’ our power forecasts to capture the individual nature of each facility (e.g. local terrain shading, degradation over time). Third, wherever highly accurate short-term power forecasts of only minutes ahead are required, we integrate data from a low-cost sky-imager device installed on site that watches the local cloud cover. All three together produce highly precise and accurate forecasting. 

I have heard a lot about sky-imagers or cloud cameras, could you tell me about how these work?

These units are a lot of fun. Most are prohibitively expensive because they are made in small batches or require many units to be installed across the solar farm. However, at Solcast, we’ve chosen a low cost, highly scalable option. What these units do, fundamentally, get installed at the solar farm, point up at the sky and attempt to detect clouds and predict their future position. In our case, we combine this information with the real-time SCADA data from the solar farm and our satellite imagery to provide one seamless forecast from minutes out to days ahead. 

I know Solcast does solar forecasting, but what about historical data like TMY files?

Another solid question. The two are increasingly linked, in that forecast, historical and even real-time solar data have a role in PV plant operations. Historical data is most often used for prospecting for new sites, completing due diligence for investment and then tracking the yield of the solar farm at regular intervals to check on the performance of the asset. Solcast also offers historical time-series and typical meteorological year (TMY) data for these purposes, which can be downloaded in just a few minutes from our API Toolkit. 

Where can people who are interested learn more about Solcast?

We’d love for folks to read-up on our website: https://solcast.com and come here to our presentation at PV Operations USA coming up on 6-7 November in San Diego. I hope to see you there!

Find out more on November 6th, 16:30 at #PVOPSUSA2019