Partial Discharge in Molded MV Products

[BigJohn1]

I can't see any rhyme or reason for what will cause partial discharge in medium voltage cabling systems:

I have noticed that a lot of molded splice and termination products have very significant internal gaps between the conductor and the molding as part of their design. These seem to work fine for years with no serious partial discharge.

Yet, when hand-taped splices and terms are put on a tremendous amount of effort is put into excluding even the smallest voids or spaces, for fear of destructive partial discharge.

What is the difference here? Why is PD not as much of a problem with molded products?

Further, I have seen 35kV elbows where PD was happening very seriously but after looking at dozens of identical installations there again seemed to be no rhyme or reason behind where it was occurring (sometimes near air spaces in the kit, sometimes not, some with zero evidence of PD, some almost completely destroyed).

Why are there "different rules" for molded products than tape for excluding partial discharge?

 

[noneavailable]

Probably has something to do with the science behind the design. But I would assume this is because the actual insulation covers those gaps you are talking about.

 

[stevenal]

I've noticed this as well. Seems odd to use an air gap in a confined space that way, subject to pressure/volume variations with loading.

 

[Compositepro]

Often the polymer in direct contact with the conductors in high voltage cable is made slightly conductive to mitigate the partial discharge issue. Voids in this polymer layer would not be a problem, so voids may be intentionally introduced to to save cost.

 

[itsmoked]

Could be the chemistry too. Ozone is highly active. A tape or other material could be relatively rapidly attacked while the molded connections could be of chemically inert materials not subject to ozone attack. Also, ozone half-life is greatly affected by humidity so identical splices at different sites may age radically differently.



Keith Cress
kcress -

http://www.flaminsystems.com