Reading the meter against the plan.
Every unit’s week, half hour by half hour
One block is one battery unit. Each row is a day, the first day at the top, and each column is a half hour from midnight to midnight. Grey means the meter was within 2 MWh of the plan. Blue means it was not, but the unit’s frequency response contract accounts for the difference. Red means it was not and nothing published explains it. A hatched row is a day Elexon has not metered yet. Click a block to open it.
Reading the meter.
Click any block to see that unit’s plan, its contracts and its meter, day by day.
What still needs explaining
The stretches of an hour and a half or more where a unit was off its plan and its contracts do not account for it, largest gap first. The total gap adds up how far the meter was from the plan in each half hour, in either direction, so a battery that charged and discharged while its plan said idle shows a large gap beside a small net meter reading. A battery that trips, runs out of charge, runs a test or trades without updating its plan would all show up here. The data cannot say which.
The fleet, day by day
Meter minus plan, added up across every unit, beside what the response contracts predicted. Below zero means the fleet took in more power than its plans said.
The same thing as a table
Click a column to sort. Hours are unit hours, so two units off plan in the same half hour count twice.
The plan is built exactly as the live page builds it. The meter is Elexon’s dataset B1610. Contracts are NESO’s response auction results by unit, and the grid frequency is Elexon’s, every fifteen seconds. The expected response is the only worked out figure here. It measures what a contract required; what each unit was paid is on its own page. How this is worked out, and what it cannot tell you