Matched

Open-data ratings

What can be determined from public data?

From public data we can track the power that suppliers buy from renewable generators and the power that they sell to their customers, and both can be characterised with half-hourly fidelity.

Renewable supply

When suppliers buy renewable power it is certified by Ofgem1 with a ‘Renewable Energy Guarantee of Origin’ certificate (REGO) which is tracked in a centralized database. We use these certificates to track where each supplier buys their renewable energy from, and in what volume.

Certificates track energy in monthly or annual blocks so we use grid data to infer the supplier’s half-hourly power delivery from the volumes of wind, solar, biomass, and hydro generation that were on the grid at large. For example, given certificates for 100 MWh of wind energy in October we distribute the energy across each half-hour of the month proportional to the wind power on the grid at that time.

We are not yet handling transmission losses but will do so in future iterations, alongside other improvements2.

Customers’ demand

Feed P114 from Elexon provides the regional half-hourly metering data of all parties that are registered to manage energy transactions with the grid. This feed has been in the public domain since 2022.

With the P114 feed we can track the aggregate demand served by each supplier to all of their customers. In most cases this is straightforward, but there are some important considerations when suppliers have embedded generation.

Renewable deficit

This is whatever power a supplier sells its customers that they can’t cover with the renewable power that they’ve bought.

A supplier will have a renewables deficit when they underestimate the demand of the customers, overestimate the output of generators, or contract insufficient renewable generation.

To cover a renewable deficit a supplier must buy from power markets which are predominantly fossil-fueled.

The frequency and extent of deficits determine how far a supplier is from being renewable: suppliers that are exclusively renewable have no renewable deficit.

Aside: when suppliers have renewable surpluses they must sell the excess to other suppliers. Since it’s consumed elsewhere a surplus has no effect on the quality of electricity that a supplier sells its customers.

Temporal-matching score

From the renewables deficit we can define a matching score which aggregates the deficit over all periods of time and (so that we can compare suppliers of different sizes) normalises by the demand.

\[\begin{equation*} \mathrm{M_{t}} = \left(1 - \dfrac{\sum_{t} \mathrm{renewables\_deficit_{t}}}{\sum_{t} \mathrm{demand_{t}}} \right) \times 100\% \end{equation*}\]

Or, said another way, the temporal-matching score is the fraction of customer’s energy that comes from renewable sources.

Volumetric-matching score

The industry currently markets electricity using a volumetric-matching score which is simply the ratio between the renewable power bought and the power consumed over the course of a year.

Definition $$ \begin{equation*} \mathrm{M_{vol}} = \left( \frac{\sum_{t} \mathrm{renewables_{t}}} {\sum_{t} {\mathrm{demand_{t}}}} \right) \times 100\% \end{equation*} $$

The volumetric score ignores the timing of power supply or whether it could be physically delivered to customers and - as a result - will always be bigger than the temporally-matched score.

Code

Our code is open source and questions and comments are welcomed!

Footnotes

  1. Office of Gas and Electricity Markets ↩︎

  2. The half-hourly decomposition of renewable supply can be improved by:

    • directly using BM volumes directly for assets >50 MW
    • applying a geographical calibration to account for regional wind speed and solar irradiance
    • applying apply an asset-level calibration based on the type, vintage, and performance-curves of the asset
    • (where available) use metering data directly

    ↩︎