25. August 2011 20:47
Tanδ ex post facto
I have sent you some before-and-after diagnostic data (0.1Hz tan-delta) on two cables treated with Novinium® Ultrinium™ 732 fluid? The results are extraordinarily positive. What say you?
The data you provided is reproduced in the chart nearby. A few of my historical postings provide evidence that I am not a big fan of diagnostics. For example check out the postings below ...
2010-11-12 Diagnostic Testing – Should I do it?
2010-12-12 Electrical Treeing and the Principle of Parsimony
2011-05-11 Middle East Query – Diagnostic Testing Timing
2011-05-19 On-line Diagnostic Testing
Check out the before-and-after dielectric spectroscopy data presented in another post when I addressed tan delta measurements specifically at …
2010-09-10 Cable Rejuvenation Impact on Loss Factor (tan-Delta)
The data from the University of Connecticut’s Electrical Insulation Research Center (EIRC) in that post leaves no doubt that rejuvenation has an impact on tan delta measurements, but it creates considerable doubt as to the meaning of that impact. It is also true that there can be no doubt that rejuvenation with Novinium fluid technology provides a reliable post-injection life extension – over 99.6% of all treated cables are providing failure-free performance.
It is tempting, whether human or amphibian, to embrace data when it reinforces what you already know. As a disciplined scientist, I, however, am able to resist that temptation. The 17 month improvement demonstrated by the Illinois Cable 1 and Cable 2 data suffers at least two shortcomings. First, there were precisely two cables measured two times about 17 months apart. That is not a statistically significant data set. Second, without following the tan-delta over time, it is not possible to correlate the “improvement” with a reliability-metric like AC breakdown performance, which has been measured and modeled extensively. We can rectify these two issues with a more comprehensive analysis that includes more cables and more frequent periodic re-testing. Such a program is in the works with Greater and we will report on the results as the data becomes available. The testing itself carries risks for the cables to be tested. Knowledge is not free and you can’t make an omelet without breaking some eggs.
Interested, but skeptical,
T. B. Frog
11. August 2011 16:06
Dear Frog of Knowledge,
Can you explain the Crow-AMSAA graph on Novinium’s web site at novinium.com/lessons.aspx? How do I compare Novinium’s post-injection reliability to that of UTILX?
Not wanting to eat crow,
What a great play on words! My fans that compose creative questions get to move right to the front of the queue. You are justified in “c’raving” reliability. I took the opportunity your inquiry provided to update the Novinium failure statistics through August 10, 2011. I present those statistics in the chart nearby. Crow was a guy that worked for the “Army Material Systems Analysis Activity” or AMSAA. Crow developed the statistical model that now bears his name and that of his employer. Crow-AMSAA or “C-A” for short is widely recognized as a preferred model to predict the reliability of complex systems that experience multiple failure mechanisms.
The x-axis is the product of the feet of cable that have been treated by Novinium and the years that have elapsed since treatment. For example, if a 328 foot (100 meter) length of cable was treated three year ago, its contribution to the cumulative treated would be 984 feet*years. The x-axis is logarithmic. Plotted against the y-axis are the failures – 56 in total. A least squares regression of the failures provides a slope, or beta, of 0.64. A beta less than 1 means the failure rate is decreasing. Process and chemistry improvements, together with the improving mastery of the Novinium’s craft workers, make Novinium technology more and more reliable. That’s not to say that when Novinium began injection operations over six years ago post-injection performance was unacceptable. Novinium started where the old technology, invented by Novinium founders, reached a reliability plateau. When I did this same C-A analysis nine months ago (November 2010) the beta was 0.72. So, not only is the failure rate decreasing, but the rate of decrease is accelerating! About 99.4% of all the cables, which Novinium have treated, remain in reliable service. This is at least twice as good as the other guys!
With regard to how you can compare Novinium reliability with that of UtilX, I can only provide you with some frog-advice. As a circuit owner you should demand that UtilX publish its total failure statistics – not just a few select circuit owners, the whole data set. Then circuit owners would be able to make an apples-to-apples comparison. Don’t hold your breath, when NEETRAC, their sponsoring circuit owners, and other industry leaders invited UtilX to participate in a side-by-side laboratory experiment, UtilX helped craft an experiment, but withdrew their participation when the experiment actually began. By the way, that experiment is complete and included the only rejuvenation firm willing to share their post-injection results in a truly independent experiment – that would be Novinium. UtilX demurred, citing “business and commercial reasons.”
No need to crow when you can croak,
T. B. Frog