I'm still looking at this thread, but I thought I'd throw out a
comment that may or may not be related. I remember many years ago
learning that high voltage capacitors are hazardous because thermal
processes will cause them to become polarized over time even if they
are initially "discharged". The voltage they can reach depends on
their construction, and can be potentially lethal for units with large
capacitance designed for operation at many kV. Such capacitors are
always stored with their terminals shorted together to prevent such
spontaneous polarization. We even have a bunch of them sitting around
here that were part of a pulsed laser that had been donated many years
ago. The smaller ones are 100 uF @ 3 kV, and the large ones are 100 uF
@ about 20 kV. Both types are stored with shorting bars between their
terminals.
I have a student doing a senior research project on series and
parallel combinations of capacitors (a project inspired by discussions
on phys-l!). We might look into this problem, too.