I am considering the use of surge testing in a predictive testing program. Are any of you conducting surge testing on large numbers of 480V motors? Have you experienced faults that could be blamed on the test procedure? Have you restarted motors after they have failed a surge test?
Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
-
Congratulations waross on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Motor Winding Predictive Testing 5
- Thread starter testtech
- Start date
-
2
- #2
electricpete
Electrical
I don't have any direct experience with surge testing other than testing done at the shop when we send our motors in.
As you probably know, surge test is like a hi-pot test but applied to the turn insulation. If surge test fails, you may have done damage and it is risky to put the equipment back in operation (it may fail in service).
There is no other practical way to test the turn insualtion.
I have heard one experienced gentlemen (oem/repair shop background) say that surge test is a good test for low voltage motors (480vac).... he ways it will find a variety of faults. I have heard another very experienced utility engineer responsible for motor maintenance at a large electric utility say that he discontinued surge testing because he felt the benefits were low.
Like hi-pot, I guess a determining factor is: how painful is a failure in service as compared to a failure during testing. If failure in service is much more painful, then it may be a good test. If failure in service only cost twice as much as failure in test, don't do the test... you will fail more motors.
I find among electric utilities many more do the dc hi-pot then do the surge test. I'm not really sure why that is... maybe this is due to the fact that surge test is a little more difficult to perform and interpret.
As you probably know, surge test is like a hi-pot test but applied to the turn insulation. If surge test fails, you may have done damage and it is risky to put the equipment back in operation (it may fail in service).
There is no other practical way to test the turn insualtion.
I have heard one experienced gentlemen (oem/repair shop background) say that surge test is a good test for low voltage motors (480vac).... he ways it will find a variety of faults. I have heard another very experienced utility engineer responsible for motor maintenance at a large electric utility say that he discontinued surge testing because he felt the benefits were low.
Like hi-pot, I guess a determining factor is: how painful is a failure in service as compared to a failure during testing. If failure in service is much more painful, then it may be a good test. If failure in service only cost twice as much as failure in test, don't do the test... you will fail more motors.
I find among electric utilities many more do the dc hi-pot then do the surge test. I'm not really sure why that is... maybe this is due to the fact that surge test is a little more difficult to perform and interpret.
I work for a manufacturer and we routinely surge test new build, I haven't heard of it being used on machines in service except in a repair shop. We use a Baker surge tester at the standard 2xline V + 1000 but only once in the lifetime of the winding at full voltage in order to prevent damage. Any subsequent testing is done at 75% of this.
I would say it is a really a pass/fail test either on a new winding or a suspect one. The Baker tester software (an optional extra) does give some quantitative data in the form of EAR values typically between 1 and 10 - a measure of the relative area under each ringing curve for three phases, but I doubt it is of much use in predictive maintenance. You could discuss it with Baker, they have been helpful to us in the past. There are other suppliers, maybe cheaper too I have to say.
I would say it is a really a pass/fail test either on a new winding or a suspect one. The Baker tester software (an optional extra) does give some quantitative data in the form of EAR values typically between 1 and 10 - a measure of the relative area under each ringing curve for three phases, but I doubt it is of much use in predictive maintenance. You could discuss it with Baker, they have been helpful to us in the past. There are other suppliers, maybe cheaper too I have to say.
Surge testing is not a good PM tool since it is potentially destructive and is far from foolproof to understand the results.
I have personally failed many machines that I surge tested. It is essentially an AC hipot. I have also personally seen failed windings with large obvious coil failures and these machines passed a surge test with flying colors.
Baker and PdMA both make pretty good winding analyzers which use P.I., hipot, resistance, capacitance, inductance and rotor influence to trend for winding condition.
I have personally failed many machines that I surge tested. It is essentially an AC hipot. I have also personally seen failed windings with large obvious coil failures and these machines passed a surge test with flying colors.
Baker and PdMA both make pretty good winding analyzers which use P.I., hipot, resistance, capacitance, inductance and rotor influence to trend for winding condition.
dgallagher
Mechanical
You may want to look at the ALL-TEST III - it is a reliable, portable motor, coil and winding tester:
-
3
- Thread starter
- #6
In reference to rewindr and dgallagher, I have used so-called inductive coil tests for years on thousands of motors. I am unconvinced that this technique can predict a turn short. These techniques might find an existing hard short, but I don't believe it is likely to find one of these in a timely enough fashion to actually predict a fault before one occurs. Certainly, as a port mortem tool, the inductive type testers are effective. However, as a predictive tool, I believe they are sorely lacking. I would certainly be interested in learning of experiences to the contrary.
electricpete
Electrical
I am unfamiliar with what the BJM test is all about.
"The ALL-TEST detects insulation faults between phases, shorts between coil turns, open circuits as well as ground
insulation faults.In addition, phase balance and phase angle can be easily measured and compared.
Don't confuse the ALL-TEST with lesser testers using only inductance, resistance and insulation to ground as their
main reference!...
PATENTED METHOD. Our method used to find a leakage between turns or coils is called the I/F
method. By doubling the frequency of a given voltage, the current is halved. If there is no or little
change, there is a fault present.The instrument reads the change in % and analyzing a problem is
quick and easy. For complete details, see our User Manual."
What is the principle of this test? Is it the same one that testech is talking about?
"The ALL-TEST detects insulation faults between phases, shorts between coil turns, open circuits as well as ground
insulation faults.In addition, phase balance and phase angle can be easily measured and compared.
Don't confuse the ALL-TEST with lesser testers using only inductance, resistance and insulation to ground as their
main reference!...
PATENTED METHOD. Our method used to find a leakage between turns or coils is called the I/F
method. By doubling the frequency of a given voltage, the current is halved. If there is no or little
change, there is a fault present.The instrument reads the change in % and analyzing a problem is
quick and easy. For complete details, see our User Manual."
What is the principle of this test? Is it the same one that testech is talking about?
- Thread starter
- #8
In reply to electripete:
There are two inductance type testers of which I am aware and I own them both. Here is a brief and oversimplified summary: Tester "A" measures inductance imbalance and several other parameters (megohms, resistive imbalance in micro-ohms, capacitance to ground). Tester "B" measures inductive imbalance, resistive imbalance (using a two wire method,rather than Kelvin; this is clearly inadequate), resistance to ground up to 100 megohms and two unique inductive balance measures--power factor and current ratio. Power factor compares the test signal voltage and current phase lag for each motor winding. The Current ratio introduces a test voltage at two frequencies, the second is twice the first and then measures the ratio of the flowing current.
The theory, in a nutshell here, is that 3 phase motors have 3 essentially balanced windings. If a winding has a short the inductance balance will be altered. This in fact is true. If you test a shorted motor (after failure) you will verify these imbalances. Using only inductance imbalance as a predictive tool is nearly useless. This is because the position of the rotor will influence the inductance imbalance as much as or more than a short. The other two inductance based tests-power factor and current ratio are supposed to be independent of rotor position. My test results suggest that this is generally the case.
Here is the basic problem: In order to detect a short with these instruments, the short must be hard, i.e., a part of the circuit. One school of thought claims that motors can operate for extended periods of time with shorts. The other school claims that motor failure ensues shortly after the development of a hard short.
With standard predictive technologies there are reams of studies documenting cases from detection through disassembly to verify the nature of the fault. In addition, many faults may be diagnosed and confirmed by multiple instruments.
Neither is the case with inductive instruments. There are no industry standards on which to base a decision or even conduct a test. There is little or no documentation demonstrating specific shorts that were identified in properly operating motors.
I am in possession of several motors that were replaced when they failed an inductance type test. After they were physically removed from the circuit, they passed! Prior to replacement, these motors were tested at the MCC and local disconnect. I test numerous chiller motors ranging from 100HP to over 1000 HP. Approximately 25% of the motors fail the inductance tests. I have never recommended action based on this test and none of the chiller motors have failed over the years. The acceptance range of the tests can be fairly close. However, repeated testing will often produce sufficient spread in the results that you can select whether you want to pass or fail the motor. I have surge tested one motor that failed the inductance test. The motor passed the surge test. In and of itself, this proves nothing, except the motor is still running.
I hope this provides a brief overview of the tests. If others have had positive tests results, I encourage that descriptions be shared. In the matter of predictive motor winding testing, we end users are at the mercy of the instrument manufacturers because the tests do not have the same level of acceptance, penetration and engineering documentation as other predictive test technologies.
There are two inductance type testers of which I am aware and I own them both. Here is a brief and oversimplified summary: Tester "A" measures inductance imbalance and several other parameters (megohms, resistive imbalance in micro-ohms, capacitance to ground). Tester "B" measures inductive imbalance, resistive imbalance (using a two wire method,rather than Kelvin; this is clearly inadequate), resistance to ground up to 100 megohms and two unique inductive balance measures--power factor and current ratio. Power factor compares the test signal voltage and current phase lag for each motor winding. The Current ratio introduces a test voltage at two frequencies, the second is twice the first and then measures the ratio of the flowing current.
The theory, in a nutshell here, is that 3 phase motors have 3 essentially balanced windings. If a winding has a short the inductance balance will be altered. This in fact is true. If you test a shorted motor (after failure) you will verify these imbalances. Using only inductance imbalance as a predictive tool is nearly useless. This is because the position of the rotor will influence the inductance imbalance as much as or more than a short. The other two inductance based tests-power factor and current ratio are supposed to be independent of rotor position. My test results suggest that this is generally the case.
Here is the basic problem: In order to detect a short with these instruments, the short must be hard, i.e., a part of the circuit. One school of thought claims that motors can operate for extended periods of time with shorts. The other school claims that motor failure ensues shortly after the development of a hard short.
With standard predictive technologies there are reams of studies documenting cases from detection through disassembly to verify the nature of the fault. In addition, many faults may be diagnosed and confirmed by multiple instruments.
Neither is the case with inductive instruments. There are no industry standards on which to base a decision or even conduct a test. There is little or no documentation demonstrating specific shorts that were identified in properly operating motors.
I am in possession of several motors that were replaced when they failed an inductance type test. After they were physically removed from the circuit, they passed! Prior to replacement, these motors were tested at the MCC and local disconnect. I test numerous chiller motors ranging from 100HP to over 1000 HP. Approximately 25% of the motors fail the inductance tests. I have never recommended action based on this test and none of the chiller motors have failed over the years. The acceptance range of the tests can be fairly close. However, repeated testing will often produce sufficient spread in the results that you can select whether you want to pass or fail the motor. I have surge tested one motor that failed the inductance test. The motor passed the surge test. In and of itself, this proves nothing, except the motor is still running.
I hope this provides a brief overview of the tests. If others have had positive tests results, I encourage that descriptions be shared. In the matter of predictive motor winding testing, we end users are at the mercy of the instrument manufacturers because the tests do not have the same level of acceptance, penetration and engineering documentation as other predictive test technologies.
jbartos
Electrical
- Jan 15, 2000
- 6,440
- 0
- 0
Suggestion: Visit
or equivalent for more info
or equivalent for more info
electricpete
Electrical
Thanks for the great detailed info, testtech.
I was always under the impression that a turn-to-turn fault would escalate to phase or ground fault pretty darned quickly, which would make detection of "hard-wired" turn faults not too important. But you never know...
I was always under the impression that a turn-to-turn fault would escalate to phase or ground fault pretty darned quickly, which would make detection of "hard-wired" turn faults not too important. But you never know...
The idea behind surge testing which is an improvement upon simple inductance testing is that it applies a significant voltage between turns, so in theory it can identify weaknesses and not just hard shorts. Incidentally, I agree with electricpete, a turn-to-turn fault is most likely to deteriorate into a complete failure pretty rapidly.
But it may only properly stress the insulation on the first few turns from the phase connection, in the same way that PWM/IGBT inverters are supposed to do. If you are testing a machine for possible damage from an inverter this may be adequate, but in my case (testing new build) maybe it's doing more harm than good (anyone want to buy a second-hand surge tester).
But it may only properly stress the insulation on the first few turns from the phase connection, in the same way that PWM/IGBT inverters are supposed to do. If you are testing a machine for possible damage from an inverter this may be adequate, but in my case (testing new build) maybe it's doing more harm than good (anyone want to buy a second-hand surge tester).
jbartos
Electrical
- Jan 15, 2000
- 6,440
- 0
- 0
Suggestion: When it comes to the turn-to-turn shorts, there is a major difference between:
1. A turn-to-next-turn-in-a-layer short, and
2. A turn-to-next-turn-in-other-turn-layer short.
The 1. can eventually become harmless in winding with many turns
The 2. is more critical and probably will escalate to a winding collapse due to the short.
This is why some motor and transformer manufacturers carefully wind coils.
1. A turn-to-next-turn-in-a-layer short, and
2. A turn-to-next-turn-in-other-turn-layer short.
The 1. can eventually become harmless in winding with many turns
The 2. is more critical and probably will escalate to a winding collapse due to the short.
This is why some motor and transformer manufacturers carefully wind coils.
- Thread starter
- #13
It appears that UKpete and Rewinder have done lots of surge testing.
I am aware that surge testing can be potentially destructive. However, I am curious how the instruments you have used compare with the most current models. In these, the injected energy is limited to about 15 watts. Older models achieved 180 watts. I am not sure of the relationship between watts or joules and the inherent destructiveness of high voltage, ie, as current is reduced at a given voltage, how will the level of destruction be impacted? In older models, I believe two phases were constantly pulsed during the test. In the latest models, one phase is pulsed at at a time using a step voltage process. The electronics can now detect waveform defomation resulting from a single pulse as the voltage increases. It detects the arc as it forms and then immediately ends the test. This, of course, is according to Baker, I obviously don't have field experience that would support or deny the claims of reduced destructiveness. The test standards vary signicantly by organization and country, with the European standards calling for substantially higher voltages than US standards. For 480V for in service testing, IEEE522 recommends 1029 volts, EASA recommends 1970V and IEC3415 recommends 4498 volts or 6902 volts. I think test voltage as well as characteristics of the winding determine the level of voltage reaching through the windings.
One way to place this in perspective is to consider pulses produced by PWM/IBGT drives. An excellent reference on this subject is found at IRIS Power Engineering:
The paper indicates that a 480V PWM drive can produce about 10,000 surges per second over 1,000 volts with 1,200 surges being recorded. Motors in study failed over days to months in actual operation. Thus, it took billions of surges to produce failure in the motors in the paper. This level of surging will obviously never be achieved in a surge tester.
It is certainly possible that the test is safe in the sense that it will not create a fault. However, it seems to me that the method of exposing a potential fault will by definition result in a penetration of the insulation. In this condition will the motor restart at operating voltage after failing the surge test? If the answer is yes, then the test may be reasonable. If the answer is no, then there are risks to be weighed by the owner of the motor to be tested. So, an important question that some of you may have experience with is whether a motor that fails a surge test can be restarted.
I am aware that surge testing can be potentially destructive. However, I am curious how the instruments you have used compare with the most current models. In these, the injected energy is limited to about 15 watts. Older models achieved 180 watts. I am not sure of the relationship between watts or joules and the inherent destructiveness of high voltage, ie, as current is reduced at a given voltage, how will the level of destruction be impacted? In older models, I believe two phases were constantly pulsed during the test. In the latest models, one phase is pulsed at at a time using a step voltage process. The electronics can now detect waveform defomation resulting from a single pulse as the voltage increases. It detects the arc as it forms and then immediately ends the test. This, of course, is according to Baker, I obviously don't have field experience that would support or deny the claims of reduced destructiveness. The test standards vary signicantly by organization and country, with the European standards calling for substantially higher voltages than US standards. For 480V for in service testing, IEEE522 recommends 1029 volts, EASA recommends 1970V and IEC3415 recommends 4498 volts or 6902 volts. I think test voltage as well as characteristics of the winding determine the level of voltage reaching through the windings.
One way to place this in perspective is to consider pulses produced by PWM/IBGT drives. An excellent reference on this subject is found at IRIS Power Engineering:
The paper indicates that a 480V PWM drive can produce about 10,000 surges per second over 1,000 volts with 1,200 surges being recorded. Motors in study failed over days to months in actual operation. Thus, it took billions of surges to produce failure in the motors in the paper. This level of surging will obviously never be achieved in a surge tester.
It is certainly possible that the test is safe in the sense that it will not create a fault. However, it seems to me that the method of exposing a potential fault will by definition result in a penetration of the insulation. In this condition will the motor restart at operating voltage after failing the surge test? If the answer is yes, then the test may be reasonable. If the answer is no, then there are risks to be weighed by the owner of the motor to be tested. So, an important question that some of you may have experience with is whether a motor that fails a surge test can be restarted.
testtech,
In my motor repair shop, we routinely surge test about hundred machines (from 1 HP LV motors to 11 KV, 60 MW generators) every year as a part of rewind/overhaul program. All the rewound machines (both random and form wound) are surge tested at 2 rated voltage + 1 KV and all are working for more than ten years without any electrical failures.
During overhaul of machines, we surge test stator and rotor windings at 75% of the above stated level, wherein we found two results. In machines having “hard” faults, the surge test reveals shorts at voltages lower or near the rated voltage in which case we advise our clients for a rewind.
In case of “incipient” turn faults, the surge test reveals shorts at much higher voltage levels. During such test “incipient” turn fault tests, we have found that when we lower the surge voltage, the fault vanishes thus confirming the incipient fault has not turned into a hard fault due to surge test. In such cases, we let the client decide whether to rewind or take some risk in running the machines in that condition. In the cases of clients taking such a risk, the machines have not failed immediately but have failed after 6 to 12 months. This again confirms (1) that surge test is not a potentially destructive test if one knows to stop the test at the fault incipient level and (2) it is a good predictive test too.
Of course, the above is based on my experience and others may differ.
In my motor repair shop, we routinely surge test about hundred machines (from 1 HP LV motors to 11 KV, 60 MW generators) every year as a part of rewind/overhaul program. All the rewound machines (both random and form wound) are surge tested at 2 rated voltage + 1 KV and all are working for more than ten years without any electrical failures.
During overhaul of machines, we surge test stator and rotor windings at 75% of the above stated level, wherein we found two results. In machines having “hard” faults, the surge test reveals shorts at voltages lower or near the rated voltage in which case we advise our clients for a rewind.
In case of “incipient” turn faults, the surge test reveals shorts at much higher voltage levels. During such test “incipient” turn fault tests, we have found that when we lower the surge voltage, the fault vanishes thus confirming the incipient fault has not turned into a hard fault due to surge test. In such cases, we let the client decide whether to rewind or take some risk in running the machines in that condition. In the cases of clients taking such a risk, the machines have not failed immediately but have failed after 6 to 12 months. This again confirms (1) that surge test is not a potentially destructive test if one knows to stop the test at the fault incipient level and (2) it is a good predictive test too.
Of course, the above is based on my experience and others may differ.
- Thread starter
- #15
edison123, Thank you for the reply. You have not only provided the type of "field" evidence that I need, but a useful test procedure for assessing the condition of the winding after it fails the test.
One question. Regarding the motors where a hard short was revealed, were these machines running before they came to your shop, or were they already failed to the point they were inoperable? The issue concerns whether a motor with a hard short can operate for a significant period of time or whether the motor fails shortly after the incipient fault becomes a hard fault?
Thanks for your response!
One question. Regarding the motors where a hard short was revealed, were these machines running before they came to your shop, or were they already failed to the point they were inoperable? The issue concerns whether a motor with a hard short can operate for a significant period of time or whether the motor fails shortly after the incipient fault becomes a hard fault?
Thanks for your response!
testtech,
In the case of machines with hard shorts, majority of them come to us with winding failures. But in others, they were undetected during running and were revealed only during testing under overhaul program in our shop.
In my opinion, turn faults (hard or incipient)are difficult to detect in the normal motor/generator operation till they develop into significant ground/phase to phase faults. This fault 'maturity' time depends on other factors like no. of start / stops, type of switching device (VCB's, VFD's etc), operating environment (polluted atmosphere aid in fault development), operating temperature (higher temps result in faster insulation deterioration) etc.
Finally, in my experience, I have found that incipient turn faults (either detected earlier in PM program or undetected) will invaribaly grow into a ground/phase to phase fault.
I have also learnt from your posts in this thread and thank you for the same.
In the case of machines with hard shorts, majority of them come to us with winding failures. But in others, they were undetected during running and were revealed only during testing under overhaul program in our shop.
In my opinion, turn faults (hard or incipient)are difficult to detect in the normal motor/generator operation till they develop into significant ground/phase to phase faults. This fault 'maturity' time depends on other factors like no. of start / stops, type of switching device (VCB's, VFD's etc), operating environment (polluted atmosphere aid in fault development), operating temperature (higher temps result in faster insulation deterioration) etc.
Finally, in my experience, I have found that incipient turn faults (either detected earlier in PM program or undetected) will invaribaly grow into a ground/phase to phase fault.
I have also learnt from your posts in this thread and thank you for the same.
I don't think surge testing is very good for field testing. The rotor has a big influence on the shape of the traces. I think it is much better for shop tests, or at least when the motor is disassembled. I won't even address the destructive aspect of surge. Look at PdMA.com and see their approach to testing. We have been using their tester for over 4 years with good results and work in a chemical plant with 25,000+ motors.
electricpete
Electrical
I am not a huge proponent of surge testing.
But I do beleive that the concern voiced about possible influence of the rotor can be partially addressed by using a voltage-increase method in conjunction with phase comparison method.
Phase comparison method - compare waveforms between phases... expect similar, influenced by rotor.
Voltage increase method - slowly increase the voltage and compare successive waveforms at different voltages. There will be change in magnitude but there should be no change in the zero crossings (unless insulation breaks down between those voltage steps). This test is not affected by rotor position.
But I do beleive that the concern voiced about possible influence of the rotor can be partially addressed by using a voltage-increase method in conjunction with phase comparison method.
Phase comparison method - compare waveforms between phases... expect similar, influenced by rotor.
Voltage increase method - slowly increase the voltage and compare successive waveforms at different voltages. There will be change in magnitude but there should be no change in the zero crossings (unless insulation breaks down between those voltage steps). This test is not affected by rotor position.
- Thread starter
- #19
utvol:
I have an offline PDMA test instrument. Besides megger and capacitance testing (I have yet to convinced of what capacitance to ground actually can measure), the PDMA test has two tests for winding condition--micro-ohmmeter resistance balance and inductive balance. Clearly, with this test, changes in the rotor position can result in large changes in the inductive imbalance. These changes can easily exceed the limits set by PDMA. In my book, this renders the test of little use--it simply cannot highlight a developing problem with sufficient certainty.
The PDMA megger has been very at times. The micro-ohmmeter test has been good (but not as good as infrared) for finding poor connections at the motor peckerhead. However, I cannot say that the PDMA instrument has ever identified a developing winding fault.
I have nearly no experience surge testing. However, when I witnessed a step voltage test, the sinewave simply collapsed when a fault occured. There was no need to compare the three phases.
Perhaps you cold describe the types of problems you have found with the PDMA instrument.
I have an offline PDMA test instrument. Besides megger and capacitance testing (I have yet to convinced of what capacitance to ground actually can measure), the PDMA test has two tests for winding condition--micro-ohmmeter resistance balance and inductive balance. Clearly, with this test, changes in the rotor position can result in large changes in the inductive imbalance. These changes can easily exceed the limits set by PDMA. In my book, this renders the test of little use--it simply cannot highlight a developing problem with sufficient certainty.
The PDMA megger has been very at times. The micro-ohmmeter test has been good (but not as good as infrared) for finding poor connections at the motor peckerhead. However, I cannot say that the PDMA instrument has ever identified a developing winding fault.
I have nearly no experience surge testing. However, when I witnessed a step voltage test, the sinewave simply collapsed when a fault occured. There was no need to compare the three phases.
Perhaps you cold describe the types of problems you have found with the PDMA instrument.
electricpete
Electrical
I am interested as well. It is an issue I have followed for awhile... previous thread here:
thread237-10535
I have heard one or two people say that the plot of resistance (or current) vs time is a little more discriminating then a simple megger and pi. Explanation was something like: a jumpy rather than smooth plot seems to indicate momentary discharges during the test due to contamination?
thread237-10535
I have heard one or two people say that the plot of resistance (or current) vs time is a little more discriminating then a simple megger and pi. Explanation was something like: a jumpy rather than smooth plot seems to indicate momentary discharges during the test due to contamination?
- Status
Similar threads
- Locked
- Question
- Locked
- Question
- Locked
- Question