Does anyone have experience in performing the DC Absorption Test & its analysis?

I know it involves applying 2.5/5 KV DC to the winding insulation & recording the charging current for a set period (typically 1000 seconds) & then discharging through an external resistor for the same period, again recording the same. These are then plotted on log scales as a function of time.

I believe it is possible to calculate various time constants (ion migration, slow relaxation, interfacial polarization) from these, but I do not know how (& the basis for analyzing this data). Any information would be very helpful.

Regards,

Aditya

Do I have have any experience on that test - NO.
(Maybe someone else can provide a better response.)

I haven't heard the test you describe applied to rotating equipment.

The terms you mention (ion migration, slow relaxation, interfacial polarization) are a little unfamiliar to me.

I have heard a simplified model of insulation described by Greg Stone from Iris Power Engineering. He described that the response to a step-like change in voltage is something similar to the following:

i(t) = Ic*exp(-t/RC) * Ia * exp(exp-t/tau)+ Iresistive

Where Ic is capacitive charging current, Ia is absorption current, and Iresistive is the leakage current (the one that concerns us most).

The capacitive time constant t/RC may be a few seconds and the absorption time constant tau is a few minutes.

At t=0, capacitive current is the largest, but dies out quickly at which point Ia is the largest. By 10 minutes only Iresistive remains

Thus the theory of the Polarization Index (P.I.) test is that at one minute we have Ia and at 10 minutes we have Ir. The ratio R10/R1 = I1/I10 = Ia / Ir. A high value is supposed to be good because Ir is then low relative to Ia. (Ia somehow provides a useful benchmark against which to ratio Ir.)

If we take the same model and attempt to determine the components from a step test, it should not be difficult. Draw i vs t on a log-log scale. The original current is Ic. The slope will be steep for the first few seconds (-RC) and then steady out for several minutes (at slope=-tau). If you take this constant slope portion of the curve and extend it back to t=0, then you should have Ia. Then of course after a long period of time, the steady current reached is Ir.

Those three components each have some phsysical meaning. Exactly what the relationship between those three three components really mean in terms of insulation condition (beyond the ratio we find from P.I.), I'm not sure.

I do a textbook "Dielectrics in Electric Fields" which provides many details that I haven't studied yet.

Again this is not a standard rotating equipment test to my knowledge. Can you explain a little more about your question and what you would want to accomplish.

Actually, I think I made an error and the absorbption current should be a power-law form rather than exponential.

Hi Pete,

I am familiar with the determination of capacitive, absorption & conduction currents; these were explained in a Megger paper.

However, the test decribed by me is different & is supposed to supplement the standard PI test. It is a requirement of a RLA study on a 30 year old 220 MW generator that has been awarded to us.

ABB & Rotary UK (see www.rotaryelectrical.co.uk/downloads/rla_services_brochure.pdf) refer to this test and now the client insists that we perform it.

I have attached the test results for reference.

Regards,

Aditya

Generator_Data.xls (110 Kb, 21 downloads)

Again it's not something I'm familiar with, but I did a little review and found a relevant article.
Attached is an excerpt from
DIELECTRIC TEST METHODS FOR ROTATING MACHINE STATOR INSULATION INSPECTION
by Goffeaux, R. Krecke, M. Comte, B. Cottet, M. Fruth, B. Centrale de Vianden; from 1998 Conference on "Electrical Insulation and Dielectric Phenomena"
The complete article available here for a fee:
http://ieeexplore.ieee.org/xpl/freeabs_all.jsp?arnumber=732952

Looking at the second column on the first page we see description of polarization mechanisms similar to the terms you used in your original posts:
"These three polarisation mechanisms are:
- movement of slow ions in the impregnating resin [ion migration]
- relaxation of electrical charge being present in the HV insulation [slow relaxation]
- interfacial polarisation, due to the presence of different materials (e.g. mica and resin) [ interfacial polarization]

Then below that is an equation for the current as a function of time, representing the components of the current. Apparently the technique is to take the measured test results and do curve fitting to determine the values of the constants required to make that equation match the data. Each of these constants has some physical meaining related to relaxation etc I assume. It appears to me similar to what I described in my first post, except just that they are fitting the data to a more complicated equation with more terms.

Here is the citation for reference 11 which I assume provides more detail on the equation being fitted:
I didn't include the references in the excerpt. Here are the one's mentioned:

Reference 11 - 11. B. Fruth, Th. Baumann, F. Stucki, “Space Charge Injection and Partial Discharge Inception at the
Metal Polymer Interfact:”, Proc. 3rd Int. Conf. on Conduction and Breakdown in Solid Dielectrics, Trond-heitn, 1989, pp. 35-39"
12. R. Goffeaux, “L’indice de conduction, une autre possibilite d’evaluation de la qualit6 eelectrique de l’isolation HT du bobinage statorique de grosses machines CA-HT”, Bull. Scient. AIM, 110 annee, no. 4, 1, 1997
13. J.S. Johnson, A.W. Ziener, “DC Testing experience on rotating machines insulation”, Trans. AIEE, no. 419, March 1989

I assume that reference 11 would have more detail on the physical interpretation of the components of that equation, but I don't have that reference.

This message has been edited. Last edited by: electricpete,

DielectricTestMethodsExcerpt.pdf (345 Kb, 39 downloads)

wello i have some doubts about the test, since this is the current assessment of absorption but did not identify the constant and the document is referred to as isolationism but not all of the machines behave in the manner shown there.If they could send information (nrodriguez@udec.cl)


I leave them here some related link:

http://ieeexplore.ieee.org/iel2/545/4227/00162569.pdf?t...ber=&arnumber=162569

http://ieeexplore.ieee.org/iel4/5934/15805/00732952.pdf?arnumber=732952