Do You Believe in Ghosts? Ghost Water Trees: Part I
by Glen Bertini, CEO, Novinium
Water trees are ubiquitous in solid dielectric insulated cables that fulfill three requirements …
1. There must be water present. (Note: Water is always present, see .)
2. There have to be ions. (Ions such as salt are found wherever there is water. Vintage conductor and insulation shields are loaded with them. The environment provides still more. See .)
3. Alternating electrical potential above a threshold value. (All energized medium voltage cables operate above the threshold.)
One American southwest circuit owner, smack in the middle of a desert wondered if their cable might be spared water tree degradation due to the scarcity of water. In fact, some folks insisted that there could not be water trees and refused to even look. With just a tinge of covert action my firm persuaded an inquisitive soul at the desert firm to surreptitiously provide samples of cable. On our dime, we sent the samples to an independent laboratory for analysis at . Three exemplary pictures of non-existent water trees are shown in Figure 1.
Figure 1. Desert-born vented tree (1) and bow tie trees (2) imaged by CTL using methylene blue stain.
In 30 wafers, each 25-30 mils thick and stained with methylene blue, CTL identified 39 vented trees and 56 bow tie trees longer than 10 mils. Those values equate to 193 bow tie trees per cubic inch of insulation and 49 vented trees per square inch of insulation/shield interface. Desert or not, water trees are ubiquitous and prolific. See  to learn why even cables in a desert are wet.
Staining with methylene blue, which paradoxically results in the insulation ending up with a red or purple hue, is a common practice to help visualize water trees. In most cases, this staining process is required to visualize water trees. However, in some circumstances water trees may be visible without staining. I call these ghost trees because they look like Casper against a field of snow – visible but with a low contrast. Figure 2, shows two of these ghost trees in an insulation wafer.
Figure 2. Occasionally water trees are visible without staining. In this image from a Canadian cable two ghost trees are visible. One is quite prominent and the second is barely visible just to the left of its more prominent cousin. There are likely other trees that would become visible if this sample were to be stained. Also visible in this photograph is the halo, a water-rich region at the mid-point of the inner and outer surfaces of the insulation. It’s the presence of water with its much greater light diffraction coefficient than the surrounding polymer that makes these three features faintly visible.
If the Figure 2 sample was allowed to dry, ghost trees appear to disappear – ghosts can do that you know. They may not disappear entirely, because when they dry the ions (salt) that were dissolved in the water are left behind. Enough ionic contamination may leave barely visible clues. But the backbone of all trees, the truly invisible part, is oxidized polymer. In chemical speak the oxidized polymer includes a very large number of carbonyl groups (i.e. a carbon with a double bond to an oxygen). These carbonyl groups never go away whether or not the cable is dried and are only visible if they are stained.
Water trees are the overwhelming cause of medium voltage cable failure as shown in . Cable rejuvenation specifically targets the chemistry of water trees and has been shown to recover the dielectric strength of cables riddled with water trees (even water trees that span 100% of the insulation width) to like-new dielectric strength – about 40 kV/mm (1000 volts/mil) as shown in .
On occasion when qualified workers are re-terminating vintage cables, ghost trees may be visible to the naked eye after the insulation shield is removed. Figure 3 is just such a case. In Figure 3 there was an obvious cable failure on the right edge and the less obvious ghost trees to the left portion of the failure point. In Part II of this story we will examine these ghost trees in more detail. We will determine if they are real, or not! We will try to determine what distinguishes ghost trees from those that are invisible without staining.
Figure 3. The white splotches are presumed to be vented water trees emanating from the insulation shield. Part II will explore these splotches in more detail.
In the meantime, it came as a surprise to me when a well-intentioned engineer from my firm, the global leader in cable rejuvenation technology, urged our field operations personnel to not attempt rejuvenation of cables with ghost trees near the termination. His reasoning flowed from his observation of Figure 3 and proceeded as follows.
1. Figure 3 is a treated cable that had failed. By the way: In the failure report the engineer properly concluded, “The thickness of the insulation layer is out of tolerance, which created areas of increased electrical stress.” He determined the cable did not meet the ICEA S-94-649 specification. Below spec insulation thickness caused the cable to fail prematurely.
2. Ghost trees are proximate to the failure and hence are potential culprits in the failure.
3. Because of the proximity identified in 2, ghost trees are problematic and should be avoided.
This logic is flawed. The cable failed because of a manufacturing defect – its insulation thickness was below the relevant specification. But even if the cable had not been defective, water trees are proximate to every cable failure. As demonstrated by the Figure 1 exemplar, water trees are routinely found in the scores or hundreds per cubic inch. The distinction that a very few water trees are visible to the naked eye (specifically those ghost water trees initiating at the insulation shield), while the vast majority of water trees are invisible without staining is a distinction without a meaningful difference. The spooky charge transporters in water trees are the invisible carbonyl groups that exist regardless of our ability to visualize or not visualize the macro structure. The carbonyl groups form a dendritic structure where each carbonyl group acts as a capacitor. Countless carbonyl capacitors coupled in series and parallel grade the electrical field around the tree structure.
We are not yet through with these ghost trees. In “Do You Believe in Ghosts? Ghost Water Trees Part II” we will examine these ghost trees and attempt to identify the difference between the ghosts and their more prolific and more invisible cousins, the latter being visible only with a suitable staining procedure.
References Bertini, “Molecular Thermodynamics of Water in Power Cables,” IEEE Electrical Insulation Magazine, Dec. 2006.  Mashikian, Groeger, et al, “Jacket for Insulated Electric Cable,” U.S. Patent 6,005,192, Dec. 21, 1999.  Cable Technology Laboratories, Report 07-037, Feb. 28, 2007.  Steennis and Montfort, “Water Treeing is Service Aged Cables …,” IEEE Trans. on Power Delivery, V5N1, January 1990.  Mokry, et al., “Cable Fault Prevention Using Dielectric Enhancement Technology.” Jicable, 1995.
Did you enjoy this post? Read this article and more by Glen Bertini on Linkedin!