Quick answer
Ask for about 12 months of half-hourly data for every meter the battery will serve, plus recent bills. Check that the file covers the right meters, has no major gaps or strange readings and clearly shows whether values are kW or kWh. Also ask about future loads such as EV chargers or new machinery.
Aim for twelve months of half-hourly data for every MPAN or MPRN the battery will serve, but if the client cannot get a full twelve months straight away, work with what they have to keep the proposal moving. When you contact them, tell them what you still need and update the model when you have it. Check it covers the right meters, shows whether each reading is in kWh or kW, and lines up with their bills before you size the battery.
How to get the right half-hourly data
Get the latest bill for each supply first. On a GB site, use it to check the MPAN. On an Irish site, use it to check the MPRN. The bill should also show the meter serial number, supplier, account number and any separate import or export meters.
In most cases the client can download the raw CSV or Excel file from their supplier portal. If you're getting it for them in Great Britain, the supplier or data collector will normally ask for a signed letter of authority. EDF, for example, asks for the account details, written permission and an authorised signature before it releases any smart-meter data.
For an Irish site, ask the client to link the MPRN to an ESB Networks Online Account. They can then download half-hourly import and export data where it is available.
Ask them for:
- Twelve months if they have it, or as close as they can get
- A second year as well, if they can get it
- Separate import and export files
- Any field or note showing whether readings are actual or estimated, where the file includes this
- The site’s opening hours, shutdowns and operating calendar
A full year is ideal because it covers a full cycle of the things that push a site's demand up and down, like winter heating, summer cooling, holidays and changes to production cycles. But the risk with a single year is that it won't show you whether anything unusual threw that year off, such as a long shutdown, a new shift pattern or a dip in output.
What to check before you size the battery
First, make sure the half-hourly data covers all the electricity demand the battery is there to reduce. A site with several MPANs, tenant supplies or separate switchboards may need more than one set of data. And grid import on its own can understate the real demand, because solar or an existing battery may already be covering part of it.
After that, take a quick look for obvious problems before you use the file to size the battery. If the data is wrong, the full quote you build from it will be wrong. Warning signs to look out for include:
- Whole stretches missing, or the same reading repeated for hours
- Spikes that are clearly wrong for the site, such as a reading several times higher than the site could physically draw
- Long runs of estimated readings rather than actual ones
- A supplier or meter change part way through
Just look for discrepancies. You don’t need to analyse every half-hour one by one. And don't be surprised if a couple of days a year have 46 or 50 half-hours instead of 48. That is the spring and autumn clock changes, not a fault in the data.
The one thing always worth checking is whether each reading is in kWh or kW. The two are easy to mix up, and getting it wrong doubles or halves the demand your sizing is built on.
Ask what they are planning to add to the site. New EV chargers, heat pumps, machinery or extra shifts all add demand. More solar works the other way. It covers some of the site demand itself and may charge the battery, so the battery is left handling a different load.
Once you’re happy with the data the client has provided, use the height of the peaks to work out the likely kW requirement and the length of those peaks to work out the usable kWh requirement. Then test each battery and inverter size against the site's load and the client's tariff. Give the client a shortlist to compare, with the assumptions beside each option.
Alternatively, give the client access at your next appointment to an interactive simulator like GridVolt's, where they can test different battery sizes and tariffs against their own half-hourly data. So when the FD takes the proposal to the board, they present their own calculations, which will always be more trusted than a sales rep’s.
Related questions
Do I need half-hourly data to quote a commercial battery to a client?
How do I size a commercial battery for the best payback?
Can I use a blended electricity rate to model commercial battery savings?