Another look at insulation resistance

OF THE MANY WAYS TO check the condition of insulation in a form-wound motor or generator, insulation resistance (IR) measurement is perhaps best known and most widely used. What it can and cannot do, however, is still misunderstood. What's the lowest acceptable value of IR? What's most likely to cause a low reading? What's the influence of winding temperature?

First of all, do industry standards require IR testing? No. The procedure is highly recommended, and we're told how to do it - but we don't have to do it. If we do, however, here's what to keep in mind: Like any other resistance, IR varies with temperature of the insulation being tested. To compare readings taken at different times, essential for the "trending" in a predictive maintenance program, each measurement must be "corrected" or adjusted to the same reference temperature of 40°C. That's made clear in IEEE 43, the Recommended Practice for IR testing.

That adjustment is made by either multiplying the measured value by the factor Kl (see graph) corresponding to the test temperature, or dividing the reading by K2. The K values pub- lished in various documents over the years have not always been the same. What appears in IEEE 43, matching the K2 plot shown here, is based on a presumed change of 2-to-l in IR for every 10° difference in insulation temperature.

As we know from dealing with the familiar "10 degree rule" in evaluating insulation life, the actual temperature difference for a 2-to-l change may vary from 5° to 20°. The K values here are subject to similar variation (hence the example of an alternative Kl plot). Tests on the newer epoxy VPI insulation systems have shown even more divergent results.

Note that the temperature must be that of the insulation itself. Temperature of the surrounding air, or even the machine frame or core, is irrelevant. A further complication arises from the recommendation in NFPA Standard 70B, Chapter 18, that "operating machines should be tested immediately following shutdown when the windings are hot and dry." The reason for that is to be sure the windings are well above the condensation temperature at which a film of moisture can develop on the surface.

The first difficulty with this, however, is the unavoidable variation in temperature throughout the winding as it cools. A measured IR value will therefore be some kind of average. A second difficulty is that the temperature will not remain constant throughout the one-minute measurement period (the change will be even greater if readings are taken for ten minutes to arrive at the Polarization Index).

Care must be taken, therefore, to keep the temperature as uniform and constant as possible during an IR test on a hot winding, recognizing that some distortion of the result is inevitable. Taking readings with the machine "cold" - stable at ambient temperature - promotes accuracy but risks condensation.

Given a valid test, what minimum IR value is acceptable? For many years, a lower limit in megohms was considered to be "1+1" - that is, one megohm for each kV of machine rating, plus one, as the lowest "safe" value (allowing continued operation).

Thus, the minimum for a 4,000 volt winding would be 5 megohms. Table 3 of IEEE 43 still prescribes that minimum for "most" windings made before "about 1970." For formed coils rated below 1000 volts, the IEEE 43 minimum is 5 megohms, whereas NFPA 70B recommends a 2 megohm limit.

However, such limits are seldom regarded as acceptable in maintenance work regardless of winding age or voltage rating. For newer windings, IEEE 43 gives a 100 megohm limit (as adjusted to 40°C). Depending upon the application, past experience, and engineering judgment, a limit of 500 - or even 5,000 - may be specified. Such figures aren't always attainable under all conditions.

What causes a low IR reading? Humidity and chemical contamination are most often cited. Note that a winding need not be wetted by immersion, splash, or drip; above the atmospheric dew point, moisture will condense on insulation surfaces. That will form a conductive path. If any dirt or chemical is also present on the surface, a carbon "track" can quickly develop, leading to winding breakdown. The onset of that will drive IR down.

Here's what IEEE 432 says, in the Guide for Insulation Maintenance for Rotating Electric Machinery 5 hp to less than 10,000 hp: "Insulation resistance tests . . . primarily indicate the degree of contamination of the insulating surfaces . . . and will not usually reveal complete or uncontaminated ruptures. . . . The contaminant may be . . . metal dusts. . . ." Hence the first step in dealing with an unacceptably low value of IR is usually to clean and dry the winding, then repeat the test.

Other testing may also be prudent, and useful guides for such testing are to be found in both IEEE 43 and IEEE 432. Insulation experts have often pointed out that no single test can ever suffice to completely evaluate insulation condition.

As engineers have understood for many years, "Insulation resistance does not provide a direct indication of the dielectric strength or voltage endurance of an insulator," and ". . . the test can be extremely misleading when used to judge remaining life or the comparative quality of an insulation system."

When properly performed, and wisely interpreted, the IR test is a useful component of a maintenance program - but only one.

© 2008 Barks Publications Provided by ProQuest LLC. All Rights Reserved.