The typical tests for a capacitor are a megger test(5 or 10kV for your 33kV unit) from the capacitor to the frame, and using a capacitor meter to measure the capacitance and compare it to the nameplate value.
You can also perform:
* HV test across the capacitor plates - usually DC, but can be done AC. (new, test values are roughly 2.15xrated voltage AC, 4.3xrated voltage DC. for older units I would be looking at around 75% of these values).
* HV test from plates to frame - can be done AC as there is very little capacitance (new, for 33kV caps, around 70kV AC - depends on BIL of system - , but decrease to around 75% for old units)
* PD measurement whilst applying HV to plates and between plates.
* DDF measurement (or DLA or tandelta or powerfactor depending where you're from).
When you IR test across the plates of the capacitors, don't forget that if there is a discharge resistor internally, you will probably measure the value of this, rather than IR of the capacitor (not much you can do about this though). IR plates to frame will not see the discharge resistor.
And always keep the termials connected together and to the frame (using sealing wire or something light gauge) after testing and during transport, so that some poor unsuspecting bugger doesn't get a belt from stored charge that will build up on the plates. A number of people have mentioned that it's not comfortable.
I think your means of testing about describes what I was looking for.
We have a single step outdoor stack with 26 cans/phase of which only 20 are in circuit at a time. The other six are spare. Our bank controller keeps indicating an Iub of 14% even when all MV fuses are in tact. So my suspicion is that a couple of the cans are open circuit and thus don't show defect by blowing their fuses.
