I heard that there is a method to measure the short circuit level at the bus. Technically it appears to be difficult and risky. I have not come across such measurements. Does anybody have clue on this?
I have never done that in the direct sense. That would, indeed, be risky
But I have done a few estimates based on voltage drop during heavy starts. The results are in quite good accord (less than 10% error, mostly) with the results you get when using transformer and infeed impedance numbers.
You need a really heavy start so that the voltage drop can be measured with good accuracy. You also need to record the waveforms so you can find out phase relations and also make sure that switching transients and DC components have died out so that you use clean sine waves for the calculations.
I have seen low voltage voltage applied to the primary with a shorted secondary. The problem with that approach is that the upstream impedance usually differs a lot between the two supplies - the ordinary supply usually has a lot less impedance than the LV supply has.
As Skogsgurra indicates, a think one can measure the short-circuit level (Scc) at a busbar just measuring the voltage change (deltaV) when switching a reactors or a capacitor bank (Mvar) onto that busbar.
So, the Scc will be numerically and approximately Scc = (V/deltaV)*Mvar.
Example: Capacitor bank = 20 Mvar, deltaV = 2%, V = 1pu
Scc = (1/0.02)*20 = 1,000 (MVA)