Quarterly Progress Update, Q4 2020
We’ve been quiet for a while, but that’s not because we were enjoying the summer on a small beach island (I wish!), but because we have been super busy with all our projects. Now that things are a tiny bit calmer - not much - we felt like it’s time for another progress update.
First of all: We will be making some changes to our newsletter. From now on, we will publish a general progress report on what’s going on with OCF roughly every three months. If you want to be kept up to date, make sure to subscribe to our newsletter now! It’s four emails per year.
Alright, now that’s out of the way, let’s talk about content. A lot has happened with our two existing projects since the last update and we even took on a new project: Open Energy. We are working on enabling the sharing of energy data, but more about that later.
For our nowcasting project to predict PV generation for any point in the UK we made the following progress over the last months:
- James did his PhD placement with us for three months and did major work on predicting PV yield. You can check out his work in our predictpvyield repo on GitHub. Thanks, James!
- We had a great UCL MSc student join us for a summer research project on predicting solar PV power output using a couple of satellite images. Here's a 5-minute lighting talk video. Gracias, Neil and Marc!
- Tom has been doing some awesome work on helping us consume numerical weather predictions from the UK Met Office and turning them into baseline forecasts. You can find his work in our metoffice_ec2 repo. Merci, Tom!
- Jack is laser-focused on building on Neil & James' great research work and continues to research ML approaches to solar PV Nowcasting
- Our six-month feasibility study together with the European Space Agency came to a close last month. We are now talking with various funders to build a full, open-source PV Nowcasting product. If you are interested, please reach out!
To estimate how much photovoltaic energy will be produced, we first need to know where the PV systems are. At the moment this is not well known in the UK for all but the largest PV systems - particularly domestic systems are poorly documented.
Here is the progress we made with the PV mapping project over the last months:
- Based on results of a major crowd-sourcing campaign to create open geographic data, Dan Stowell has just released a data set and journal paper covering over 230,000 solar PV installations across the UK! That’s huge! Thanks so much for all your work, Dan!
- Two volunteers have been kindly donating their time to develop an app to further boost the crowd-sourcing of solar panel locations. Thanks, Karan and Sol!
Over the last months we had the following achievements:
- Together with Icebreaker One and PassivSystems, we've successfully participated and won the first two phases of a three-phase Innovation competition, concluding the second phase at the end of November.
- Built a basic data discoverability prototype (based on CKAN) to help people find relevant datasets in the energy sector.
- Won support from the Open Data Institute to help us open up solar PV datasets
Over the winter we will continue our work on getting some new research results for solar PV Nowcasting using Machine Learning. We are also writing a load of grant applications and talking to potential partners to help us co-fund the development of Nowcasting! We also hope to release a satellite dataset soon, to enable all you guys to get more involved in our work as well. Exciting stuff. Stay tuned.