Water treeing in cable insulation-are laboratory tests meaningful?

Laboratory experiments using specially manufactured specimens to determine the liability of a polymer material to the formation of water trees are in many cases erroneous or even misleading in regard to the use of this polymer in real cable insulations. A fundamental problem arises from the fact that in real cables water trees originate at technological defects that cannot be imitated in the same way in laboratory specimens. Furthermore because of differences in the electrode geometry the growth rate and especially its dependence on the actual length of the water trees will be different from that in real cable insulations for most laboratory manufactured specimens. Specific laboratory conditions such as the use of higher than service frequencies often increase the problem. Especially the question of the long time growth rate of water trees in service can be dealt with only by using specimens cut from real cable insulations. To accelerate the creation and growth of water trees in the laboratory an artificially induced supersaturation of water molecules in the polymer can be used. An experimental set-up for this purpose is proposed.<>