I have designed a 200-500kV power supply to be used in the laboratory. Is it possible so immerse coils and brass parts in distilled water instead of transformer oil?
In my experience, no matter how pure the water, it'll eventually start conducting. Trace amounts of metals will solvate and you'll get breakdown under the water.
Go with dried (water free) oil, and pull a good vacuum on the oil too. It'll save you problems in both the short and long term.
Due to electrolyte process, liberate oxygen will form the oxides, and will pollute water, hence will not provide the insulation needed, as compare to oil for long run.
Deionized water may be used as insulator only in pulsed mode. At longer pulses or DC, water will conduct.
There is a time constant limit, epsilon*resistivity. At room temperature, you cannot have water resistivity higher than 20 Mohm.cm (limited by natural ionization H2O = H+ plus OH-), i.e., the time constance cannot be larger than 150 us.
One can submerge his equipment in a SF6 gas filled tank or a high pressure tank, to avoid oil. But the gas pressure needs to be large enough so that no breakdown occurs.
In DC, one says epsilon and resistivity. In AC, one often refers to epsilon and tan(delta). One thing said in two ways. Water has large tan(delta), so that large loss, due to current conduction.
So the answer is the same as for DC, the time voltage applied should be as short as the loss is acceptable, for AC frequency up to ~1 GHz.
Thank you all. I decided that the best option is SF6 with gas recovery system. I read its breakdown is 150kV for 6 bars at 5mm. It is not much pressure, I hope all the electronics will resist it.