Is it a self-contained or Slave, central battery System?
You need to take care when you choose to test the system.
BS5266 recommends if it’s a self-contained system the testing is staggered/alternate fittings tested, with the remainder tested the following day or once the previous lights are fully recharged.
If they are slave lights with a central source there are two options.
Say it’s a 3hr duration, test the battery to 2hrs and by means of observing the central battery voltage & calculation, determine if it would stand up to a full 3hr duration test.
It’s also good practice when design central battery systems to allow for a 25% spare capacity, this is for future expansion and to always ensure the system is never fully discharged after testing.
Only problem with the above method is the discharge curve is exponential so calculation is unreliable.
Or another method recommend by BS5266 is to do a full duration test at a time lease likely to cause problems.
The reason for this is, say you did a full duration test on a self-contained system and shortly after there was an explosion, fire whatever, there would be no emergency lighting!!!