We're hiring in our ML team! 💻
Learn more
To give a feel for why we're so excited about sharing energy data, let me tell a science fiction story to illustrate what we hope a 'web of linked energy data' could enable….
It's 2030.
You buy a new electric vehicle. You tell it your home address and security credentials and, just from these, the car automatically offers its services to the distributed energy resources (DER) aggregator who manages your solar array and heat-pump; and also finds the forecast service which provides forecasts for your PV array and home energy demand. These devices & services coordinate - completely automatically - to minimise your energy bill and emissions, whilst ensuring your car is charged when you need it to be; and your house is warm when you need it to be.
But that's only the start...
Your DER aggregator also talks to the substation at the end of the road, and to the other houses on your street. All these coordinate to ensure the substation is never overloaded.
A level still deeper into the grid, all these systems coordinate with the Distribution System Operator and the Electricity System Operator. If the system operator sees a problem, they can quickly and surgically use demand-side flexibility in specific locations and with specific characteristics to help fix the problem.
The electricity system also coordinates with the gas, transport, and water systems.
And that's just what happens in real-time…
Much of this data is kept for posterity. A lot of system-level and anonymised data is shared openly. More sensitive data is only shared with a limited number of organisations. A huge number of web applications and services are built on this web of energy data. Policy makers and grid planners use one of a competing group of web services to help them make multi-billion pound decisions on where to install new infrastructure. Community energy groups use a different set of competing apps to figure out where to install local renewable energy projects. Researchers use historical data to investigate sophisticated new forecasting models.
In conclusion: As a community, we need to figure out how to enable machine-to-machine communication in a hugely complex network, in near-real-time, whilst ensuring the system as a whole never fails even though individual devices will often fail; and we've got to do this on the back of a system that was mostly built decades ago.
Ultimately, what's needed is a distributed digital representation of the structure of the energy system and its dynamical state. A representation that machines can reason about, in order to optimise locally (reducing energy bills for domestic users) and globally.
Now, I'm definitely not suggesting we solve all of this in the next 12 months! Rather, I wanted to sketch what a true 'web of energy data' could enable in a decade or two, if we get the foundations right.
