Questions about Polarization Index 1

[packdad]

Help out a lowly mechanical guy! I'm learning about PI tests and have a couple questions:

1. What, exactly, is going on in the windings as the PI test is conducted? From what I understand, there are several different types of currents being generated by the test: Current due to capacitance charging (time dependent), polarization current (time dependent), and leakage current (steady). Of these three, I only have a good grasp of the last one! Capacitance I can sort of understand -- sort of. But what, exactly, is polarization current? I guess I thought that polarization current WAS the capacitance current, but apparently I was wrong.

2. The acceptable value for a PI test is 2. Why? Does a value less than this indicate that leakage current, which is clearly bad, is dominant? I think there is an IEEE standard which might explain this, but I don't have the number or a reference source.

Thanks!

 

[jbartos]

Suggestion: A little background which I sent to a similar post; however, it did not appear, i.e.
1. Engineering (theoretical) effects tend to be as follow:
1a. Capacitors
1a1. DC voltage/current
1a1a. Tends to charge capacitor and the capacitor keeps its charge with certain voltage proportional or equal to the DC charging power supply. DC current is minimal when capacitor is fully charged. Initially, it behaves as a short circuit element during it initial transient time close to time t=0sec.
1a1b. If DC supply is disconnected the capacitor is charge is slowly dissipating over the capacitor dielectric resistance and outside capacitor surface resistance.
1a2. AC current/voltage
1a2a. AC current tends to flow through the capacitor similarly as through resistor with 90° lead with respect to the power supply voltage.
1a2b. When AC power supply is disconnected the capacitor discharges its charges similarly as in 1a1b.
1a2c. AC capacitor tends to have larger losses than DC capacitor.
1a2d. AC capacitor behaves as a short circuit element for the AC current and voltage during its initial transient time.
1a2e. AC capacitor reactance is Xcap=1/(1(PI)fC) and in the impedance Z=R + j[1/(2(PI)fC)]

2. Visit

http://www.hills2.u-net.com/tesla/tccap.htm

http://www.steelman.com/

for more info

3. IEEE Std 100-2000 Dictionary, defines the polarization current as time-dependent, decaying current in the specimen, following the instant that a constant voltage is applied until steady-state conditions have been obtained. Notes: 1. Polarization current does not include the conductance current. The sum of the polarization and conductance currents in the specimen is that which is normally observed during measurements. 2. Polarization current includes both polarization absoption and capacitive-charge currents.

4. Visit

http://www.ieee.org

for information on capacitor standards.

 

[rhatcher]

packdaddy, you were pretty close in a couple of respects. First, in my mind at least, polarization current and capacitive charge current are related if not the same. Second, for modern (Class F and H) insulation systems a polarization index reading of less than 2 is due to the leakage current being a dominant factor and yes, that is bad. However, a PI index of less than 2 does not have to be associated with a high leakage current.

The polarization index of a motor's insulation is a way to determine its' suitability for service. This test will detect insulation that is moisture contaminated. The test is performed by applying a high DC voltage between the motor's winding and it's frame and recording the values of insulation resistance at 1 and 10 minutes. The ratio of the two values (10 minute/1 minute) is the PI index. Spot testing can provide useful info, but this test yields the best results when measured as a trend over time.

The PI index is most meaningful on AC motors with Class A or B insulation systems due to the hygroscopic nature (ie. able to absorb moisture) of the organic materials used on those systems and the fact that they are organic materials. For this test the motor's insulation can be modeled as a capacitor. The winding is one plate, the frame another, and the insulation is the dielectric. The amount of capacitance and the leakage current of the capacitor is based on the amount of moisture in the organic insulation materials. (Moisture may be water or oil, both are polar molecules, hence the term polarization...) The current elements present are as follows:

steady state: conductance current or leakage current - The current flow through the insulation based on insulation resistance.

time dependent (decaying): capacitive charge current and polarization current - In my mind these are very related, especially for Class A and B systems. In laymen's terms (which may be the limit of my understanding versus an insulation expert) the amount of current necessary to charge the capacitor formed by the motor under test as described above is a function of the plate size and the dielectric strength and quality. For the insulation systems in question, moisture content is an important factor in dielectric strength and quality. If the insulation contains too much moisture, it will reach it's steady state charge much quicker than one that does not. For oil contamination, the steady state leakage current may be quite low (high insulation resistance) despite the quick charge time and low PI index. For water contamination, the steady state current may be increased due to the conductivity of the water (low insulation resistance). In either case the PI index will be less than 2. Also, note that for an organic insulation, oil contamination is bad.

On the other side of the coin, an organic insulation that is dried out will take longer than normal to reach it's steady state value. Note that for organic materials dried out equates to brittle which is undesirable. This will result in a PI index of greater than 2.

Modern insulation systems (class F and H) do not include hygroscopic (or organic) materials in the insulation system. Since the insulation cannot absorb moisture, its' capacitance remains somewhat constant when contaminated with moisture. However, water contamination will increase the leakage current to the point that it is dominant, resulting in a PI of less than 2 and a low insulation resistance. However, for these systems, a PI of greater than 2 may not have any significance.

 

[jack6238]

Some comments on Polarization Index.

Background

Electronic Polarization

When atoms are placed in an electric field, the electrons are displaced toward the anode (positive) from their unstressed positions, and the nucleus is displaced toward the cathode (negative). Due to the much higher mass of the nucleus, this displacement is less than that of the electrons. This off center displacement is call an induced dipole because the charges return to their previous position when the field is removed. This mechanism is considered to be perfectly elastic. Although it is not perfectly elastic because not all of the displacement energy is returned to the system. This loss is referred to as the dielectric loss in insulation systems and it is also a function of frequency. In the case of zero frequency (DC) this loss is insignificant.


There are other types of polarizations and these are atomic, orientation and space charge. Space charge is the type of polarization caused by cosmic radiation or thermal deterioration.

Dielectric Absorption

When making a DC insulation test on a material used in the insulation system of motors, generators or cables, there are three components of currents- the leakage current, the capacitive charging current and the dielectric absorption current. The electronic polarizations (dipoles) result in polar molecules and molecule chains within the insulation system. As mentioned before the energy used in creating the dipoles is reversible and this energy can be released upon the removal of the applied voltage. It is this energy that causes the voltage to reappear at the electrodes (plates) after the stored energy of the capacitance has been dissipated by a short circuit and the short circuit removed.

The storage of energy in the dielectric (absorption current) occurs at a much slower rate and it is this reason that the test specimen should be short circuited for at least as long as the test and the recommended time for the short is four times the length of the test. This is particularly important if the leakage current is small. In other words there is a lot of polarizations occuring in the insulation system and unless it is discharged properly, a DC voltage will reappear in the system.


The time constant of the polarization for a material is a function of the composite nature of the test specimen. It should be emphasized that the the polarization mechanism are only small fractional responses. However they are the basis for the time dependence decrease in current for the test specimen. The longer the item is under test, the smaller the value of current flowing and creating dipole moments. Now of course a large value of leakage current(low resistance) will mask this absorption current.

Time-Resistance Method of Testing Electrical Equipment

Basically the condition of the insulation system is determined by comparing the insulation reading at two different times during the test. This comparison makes use of the fact that relatively "good" or "clean" insulation systems will show an increase in resistance with time. The longer the time, the less energy is going into arranging the dipole moments in the insulation system. This comparison is also useful because it is a dimensionless quanity. (resistance/resistance or current/current)

The theory is that good and clean insulation will show an increase in resistance with time-that is the absorption current will not be masked by the leakage current. On the other hand if the absorption current is masked by moisture or other contamination of the insulation system, the insulation resistance will have a relatively flat characteristic.

In summary, if the insulation system is clean and dry, the insulation resistance will increase with time( a smaller current). An insulation system that is contaminated with water, oil or is weak will show little if any increase with time.

James G. Biddle (now AVO Corp. of Dallas, TX) publishes the following table for the Dielectric Absorption Ratio (Polarization Index)

Condition 10:1 Min Ratio

Dangerous < 1
Poor <1.5
Questionable 1.5-2.0
Fair 2-3
Good 3-4
Excellent > 4

I have given the ratios for the 10/1 min. test. A 60sec/30 sec. test is also made and the corresponding values arefrom top to bottom) <1.1, 1.1-1.25, 1.25-1.4, 1.4-1.6 and > 1.6. Many technicians prefer the shorter test times especially if they have a hand crank test set.

Some References

&quot;A Stitch In Time...&quot; by James G. Biddle pub. 21-P8
Insulation Testing by D-C Methods by James G. Biddle pub. 22t1a-1975
Principles of Dielectric Engineering, by Andrew Blanck, Rutherford Research Products Comapany
The Characteristics of Insulation Resistance by Sydney Evershed, presented to Institute of Electrical Engineers, London, 27 Nov., 1913

 

[packdad]

This is great stuff. Thanks for posting!

 

[electricpete]

The relevant standard is IEEE 43-2000.

To the guy who kept saying that class A and B are old and hygroscopic whereas class F and H are new and not hygroscopic.... You're a little off-base. Class A, B, F, H are temperature ratings. The older style that are hygroscopic are asphaltic windings... also called thermoplastic. The newer style are synthetic resin (epoxy or polyester), also called thermosetting. Changeover for most manufacturers occured in the 70's. But you'll find plenty of the newer style (synthetic resin) on motors whose nameplates indicate class B.

Absorption current and polarization current are different. The fellow who talked about the dipoles gave a good explanation.

One good thing about the PI test (as compared to the 1-minute reading alone), is that it is &quot;normalized&quot;. By dividing the ratio of two resistance measurements, the effects of temperature variation is cancelled out. You can try to temperature correct using a chart but chances are there will be large errors. Therefore PI is likely to be an easier number to &quot;trend&quot; over time than IR since there won't be &quot;noise&quot; due to temp variation. Also PI eliminates a lot of volume effects related to size of the machine... so easier to develop a single thumbrule with good applicability to a lot of machines.

As far as the explanation with the three (sometimes 4) types of currents. I've heard it before... but understanding those components still doesn't give me any good basis for why PI works so well (why should remaining resistive leakage current be judged on a ratio to initial capacitive current etc... why not just use the 10 minute reading as direct indication of resistive leakage after other components have died away?). The simple fact is that PI has been around forever and has proven itself empirically.

 

[rhatcher]

A star to jack6238 for his excellent post.

For electricpete, I do agree that:
-modern solventless epoxy and polyester based varnishes are not hygroscopic
-the change to &quot;modern&quot; varnishes took place sometime around the 70's
-since that time the older solvent based varnishes or other winding treatments (asphaltic does not ring a bell right now ) are generally not used on any insulation class.
-motors with Class B insulation are still being produced and they do use epoxy or polyester based varnishes.

Now the disagreement:
Insulation class is not merely a temperature rating. That is putting the cart before the horse. The basis for the insulation classifications is the materials used, with the develpment of more heat resistant materials leading to the subsequent insulation classes over the years. (show me a Class H from before the 70's...?)The varnish (resin) is but one component of the insulating system. Any other component that is organic can be considered hygroscopic. I found a site with good info on insulation classes and the materials used at:

http://www.electro-wind.com/Service/system.htm

You can see that Class A is composed of exclusively organic materials (&quot;Class A: Materials or combinations of materials such as cotton, silk and paper when suitably impregnated or coated. Limiting temperature of 105 C.&quot. Class B is mostly inorganic materials but does show rag mylar laminate for slot liner and coil separators. This is constructed of a sheet of mylar bonded with with a sheet of rag paper (an organic). Note that the varnish types listed for both A and B are modern types though.

Another web site with much greater detail showing the actual UL standards on Class B though H is:

http://www.pleo.com/ulsystem/pleosys.htm

On this site you will see various organic products listed as acceptable for Class B insulation systems. Specifically Classes B81 and B82 lists rag paper, fish paper, and kraft paper for minor sheet insulation.

Note that these are relatively current standards listing currently available materials by manufacturer. A listing of the Class B standards from prior to the 70's would probably have more organics used for other purposes since many of the materials listed in the current standard would not have been developed. For example, some of the older motors I have seen used wood as filler, blocking, and wedges on AC stators and as spacers and forms on DC pole pieces. The old standards would also list older solvent type varnishes or perhaps asphaltic treatments that are no longer used. Finally, some of the higher insulation classes (Class H and possibly Class F) would not have been listed as they were not developed then.

Anyway, if you have a Class B motor of any vintage, you may have organic insulation materials. As well, any older motor will include hygroscopic varnishes or other treatments as well as a variety of organic insulating materials throughout the insulating system, all of which can be considered hygroscopic.

 

[electricpete]

rhatcher - You're obviously very knowledgeable in insulation systems. Since you have framed the conversation as a disagreement, I feel compelled to justify my position.

The intent of my first post was merely to clarify the following statement of yours: &quot;...Class A or B insulation systems due to the hygroscopic nature (ie. able to absorb moisture) of the organic materials used on those systems and the fact that they are organic materials&quot;

As you yourself pointed out in your second response, &quot;motors with Class B insulation are still being produced and they do use epoxy or polyester based varnishes.&quot; (non-hygroscopic). So I assume you would agree there was some need for clarification?

My suggestion for determining type of insulation (hygroscopic or non) was therefore NOT to rely on the insulation class (which IS a temperature rating, not a material specification), but instead to use the vintage of the motor, with 1970's being the cutoff. I can't take credit for that suggestion.... it's the approach used by IEEE in standard 43-2000.

Your response has illustrated that there are a lot of general inferences about insulation material that can be drawn based on insulation class. That's some good info.
Not one bit of it contradicts anything in my post. Where's the disagreement?

 

[electricpete]

rhatcher - thanks for not losing your cool and firing back at me. It strikes me my posts were a little abrasive.... Sorry about that. I think that the first time around I missed your point that insulation with synthetic resins might still be hygroscopic due to other organic materials in the insulation system. I'm not sure if I've heard of that but it sounds plausible.

packdaddy - The bureau of reclamation has a fairly thorough discussion of insulation resistance testing and PI at:

http://www.usbr.gov/power/data/fist/fist3~1/3~1_2.htm

By the way they have a TON of other good stuff on electrical equpment maintenance at:

http://www.usbr.gov/power/data/fist_pub.htm

 

[jbartos]

Suggestion/Commnent on rhatcher (Electrical) May 5, 2001:
marked by <<>>.
Now the disagreement:
Insulation class is not merely a temperature rating. That is putting the cart before the horse. The basis for the insulation classifications is the materials used, with the develpment of more heat resistant materials leading to the subsequent insulation classes over the years. (show me a Class H from before the 70's...?)
<<Suggestion: See Reference:
1. Fink G. D., Carroll J. M. &quot;Standard Handbook for Electrical Engineers,&quot; Tenth Edition, McGraw-Hill Book Co., 1968.
Reference 1 Page 18-40 Section 85 &quot;Temperature Rise&quot; Class Insulation for Integral Horsepower induction motors A, B, H.>>