We're mentoring as part of Google Summer of Code! ☀️
Read more and apply here 💻
Solar Photovoltaics (PV) is one of the most significant sources of uncertainty in the UK’s power forecasts. To mitigate against the risk of a large cloud sweeping across the country (and hence the grid losing Gigawatts of PV generation), the Electricity System Operator (National Grid in the United Kingdom) keeps lots of natural gas generators operating at less than full capacity, so they have headroom to ramp up quickly (spinning reserve).
The physics of the grid dictate that - at every instant - supply must precisely match demand. So, any loss in PV supply must be immediately replaced. The spinning reserve, usually gas turbines, are kept idling because they take several hours to start up from cold, costing a lot of money and pumping out a lot of CO2.
If National Grid had better PV forecasts, even for the next few hours, they could reduce the amount of spinning reserve required, and hence reduce emissions (by about 100,000 tonnes per year for the UK [Details]).
Further these forecasts will make solar easier to invest in and to integrate into homes and businesses. You can read more about how forecasts will save carbon and increase solar penetration here.
Nowcasting is a bit of an odd term. It means "forecasting for the next few hours". In general, nowcasting doesn't use the hugely complex and computationally-expensive numerical weather models that run on supercomputers and constitute the bread-and-butter for most forecasting agencies. Instead, nowcasting usually uses not just numerical weather models, but also statistical or machine learning techniques to take recent observations and roll them forwards.
It turns out that most nowcasting research up until now has been done on rainfall (because predicting floods saves lives and property). Relatively little research has been done on nowcasting sunlight.
Our main interest is in trying to build machine learning models to forecast solar PV (which basically boils down to trying to predict the movement and evolution of clouds). We will spend the majority of the next year or two writing code to experiment with new ways to predict sunlight for the next few hours. Inputs to the model may include satellite images of clouds, numerical weather predictions, vertical cloud profiles, and geographical information. We will utilise the team’s experience in Machine Learning as well as contributors from the open source community to accelerate progress.
We already have an early-stage prototype PV forecasts displayed in the National Grid control room, and with a couple of other energy market participants, validating our model’s effectiveness. Through 2023 and beyond we are continually improving our forecasts with National Grid and measuring the impact on cost and carbon.
As well as our national forecast, in 2022 we started working on creating super-local solar forecasts. We are using similar machine learning techniques with satellite, weather and PV data to forecast the solar energy generated by an individual solar farms or installations.
We are a successful applicant to the Google.org Impact Challenge on Climate. The Google.org Impact Challenge on Climate commits €10M to fund bold ideas that aim to use technology to accelerate Europe’s progress toward a greener, more resilient future.
We have received support from National Grid Electricity System Operator and Innovate UK “Open Source Solutions for Net Zero” to produce working forecast solutions for national and super-local level respectively.
European Space Agency Business Applications awarded us an AI Kick-Start co-funding for six months. Kick-Start activities are compact Feasibility Studies to explore new service concepts that use space tech.
