PI Test Values AC Induction Motors: Random Wound Versus Form Wound Ratios

[BigMotorGuy]

I work at an induction motor manufacturer and we routinely do PI tests on newly constructed 3 phase stators. All the info I have read on PI ratios says anything below 2 is suspect. However, we routinely test smaller (e.g. 10Hp, 380v 2 pole) random wound stators measure very near 1. It appears when running the PI test, that the resistance reaches its maximum (i.e. current becomes near constant/minimum) well before 1 minutes (20 seconds in many cases), so the 1 minute reading is very near the 10 minute reading. We have never experienced a ratio below 1, but a ratio of 1 is not uncommon. The 1 minute resistance values are usually 50GOhms (or greater since meter maxes out at 50GOhms)so we know the insulation system if good. However, the larger form wound units (>300Hp, >690v) having PI ratios over 4 is quite common and considered acceptable by all accounts.

All the recommendations say anything below 2 is suspect, so is this such an uncommon issue for PI results being 1 on new, small random wound stators?

Thanks to All.....

 

[Zogzog]

Once your IR values are over a few Gig the PI ratio becomes meaningless.

 

[rhatcher]

You cannot get an effective PI reading if your meter maxes out before the end of the test. This is true if the meter maxes out at any time before the ten minute reading.

What Zogzog was referring to is that, per IEEE, the results of a PI test can be disregarded if the one minute reading is greater than 5 GOhms (5000 MOhms). With an insulation resistance of 50 GOhms, I would think that you have nothing to worry about. However, if you can obtain a meter with a high enough range, you will be able to measure the PI value. If the insulation is the same as the larger motors, the PI reading should also be similar but at a much higher level of insulation resistance.

I would suggest that the reason that the larger motors have lower insulation resistance than the smaller motors and therefore can complete the PI test without maxing out the meter can be understood if you consider the surface area of the insulation systems being tested. Imagine the insulation being divided into small portions of a square inch. The insulation resistance of each square inch can be represented as a resistor. As you increase the area of contact between the motor winding insulation and motor frame, specifically the size and number of slots, you increase the number of resistors representing one square inch of insulation that are placed in parallel. Hence, assuming a uniform insulation system, the motor with twice the ground insulation surface area will have half of the apparent insulation resistance value.

I hope that this explanation helps you.

 

[edison123]

PI in low voltage, random wound machines doesn't have any meaning. It only applies to form wound, moulded coils.

Also, as zogzog pointed out, IR values with over 5 Gigaohms do not produce any meaningful PI's.

Muthu

http://www.edison.co.in

 

[BigMotorGuy]

Thanks for the feedback! It makes sense that the meter capabilities limits the ability to capture the increasing resistance profile, especially if the test only looks at the 1 minute and 10 minute mark. It would seem like a good idea to develop a test and equipment which can record data quickly enough to capture the change in resistance irregardless of length of test as long as you record up to steady state.

BigMotorGuy......

 

[petronila]

Hello Big

Agre with above concepts about the insulation values. The IEEE Standard 43 was re-issued in IEEE Standard 43-2000, and states that "If the insulation resistance taked at the first minute is above 5000 Meg-Ohms, PI may not be meaningful". This "new" standard include additional concerns like the use of temperature corrections of readings and I can experiment this in manny modern insulation placed in M.V machines in some times PI = 1.2 and IR = 100 GOhms, the insulation complete dry and with out faults.

For years the PI have been used like a tool that can help to evaluate the winding insulation condition,I.M.O the PI test can be done in random low voltage machines rated above 100 HP and my experience in rewinding shops tell me that is a good practice. The customer receives a report with a PI value (corrected at winding temperature) and in the future can trend and compare wity further tests also could plot a Gohm vs Time graph and watch the slope, this information will help to know the insulation conditions year by year and to show some problem, but the PI alone can tell you a only part, so IR most be done before the PI test, then the winding most be grounded by a 3 or 4 times the IR time test in order to have a good PI readings, once you have this two test then is possible talk about the insulation conditions.

Is not usual that one OEM provides in your data test the PI values, usually the OEM performs many other tests like winding resistance, IR, Hipot and surge for evaluate the winding conditions for it I am surprised that your company is performing PI test to 10 HP small motors.

PI test is based in the insulation polarization so the 10 minutes are intended to have enough time for it and I think your low IP readings are comming from no-polarization due to the insulation material composition, a ratio below one indicate the 10 minute reading is less than 1 minute reading.

Regards

 

[electricpete]

so IR most be done before the PI test, then the winding most be grounded by a 3 or 4 times the IR time test in order to have a good PI reading

I understand the important of grounding to eliminate residual effects of prior testing, but the sequence you described seems to make extra work. Why not just energize once and measure at 1 minute and then again at 10 minutes?



=====================================
(2B)+(2B)' ?

 

[petronila]

Hello Pete,

Thanks for the comments,Yes could be seen like a time waste and I understand what you mind, the modern meggers will do it the job directly collecting all data and plotting curves and all that but I am speaking for all cases including MV machines, not all large machines has only three wires, we can also find six leads type and for this cases is possible take IR readings for each phase-Ground and IR phase-phase and also PI for every phase and the residual effects are very important in end readings.


Regards

Carlos

 

[electricpete]

Thanks for clarifying Carlos - it makes sense now.

=====================================
(2B)+(2B)' ?