IEEE 112 Field Testing for Large Motors
- Thread starter Shujinko
- Start date
- Status
petersonra
Senior Member
- Location
- Northern illinois
- Occupation
- Semi-retired engineer
I was curious enough about your question to go look up just what ieee112 is. It seems to me that other than some things that might require a test stand environment like a dynamometer, most of the tests could be accommodated in the field if for some reason you really wanted to. I've never seen anybody do much more than insulation or vibration testing, or things like power, voltage, and current. Many large motors have continuous measurement capability of things like rotor temperature via embedded rtd's.
It does not seem to me like most of the tests are particularly useful unless the motor is newly manufactured or remanufactured. In either case the motor would presumably be in a shop where these kind of tests are done routinely.
It does not seem to me like most of the tests are particularly useful unless the motor is newly manufactured or remanufactured. In either case the motor would presumably be in a shop where these kind of tests are done routinely.
112 is for test stands. Unless there is something seriously wrong or there isn’t a big enough dyno that you are willing to pay for, there is no point. The goal of 112 is to get enough data to calculate the standard 5/6 parameter motor circuit model. Those values don’t generally change much.
Unless you need something really detailed like calculating starting time accurately there is no need. This is strictly the realm of the EMTP and SPICE crowd. I’ve used it maybe twice in 30 years. Steinmetz is a 5 parameter model. IEEE 112 adds core losses for a 6 parameter model. You can estimate most of these with name plate data only if it’s a NEMA or IEC motor without IEEE 112 dyno tests.
Our shop goes to 23,000 HP in house, more if we need to. We have the clean room, test stand to 13 kV, etc. We do several wound rotor motor rebuilds over 1,000 HP plus we have at least 1 or 2 others in process all the time with a dedicated large motor crew. I’m the chief electrical service engineer so it’s my job to do most field tests.
The standard battery of tests we run on ANY motor and that most motor shops run.on all motors is the “Baker” or “PdMA” sequence offline. Every incoming motor gets this. If the first test or two isn’t getting results I automatically run this test because it saves a lot of time over doing the same tests by hand. This is ohms on coils, inductance on coils or surge comparison, insulation resistance and PI, capacitance to ground. We check by comparisons (resistive unbalance for instance) or to historic or known data. This catches about 95% of defects. It can be trended in the field with a field test or used for diagnostics. After this test it is known if it is safe to start. Up and running we can look at power, power quality, and impedance. Voltage/current unbalance is a common issue. Harmonics very rare but get blamed for a lot of things. We can also look at rotor bars (side bands) and eccentricity (swirl pattern) while running. Some newer relays show this or external testing can find it. Partial discharge is being touted but I only know of one case where it might be useful.
Finally another basic test is no load currents. Useful as a quick test for installation errors as well as checking rotation at the same time. Things that the offline test can’t verify easily.
Vibration has already been mentioned. This does a better job on mechanical diagnostics. We also do ultrasound on occasion for noise complaints and ultrasonic greasing with a UE 9000 (documenting).
On truly large generators like hydroelectric there are some additional tests that check for things like water leaks. It’s such a specialized business and not in our service area we don’t do this.
There is a little data you can get from a locked rotor test too but as a field test we can do this in across the line starting if we can start that way. It’s not a common test to use. It is mostly looking for contactor issues during startup. I have rarely done this. It’s possible to use it predictively and some of the nicer relays (SEL 749M, 710)’have this test built in.
The rest of the tests are 25/50/75% load tests. With some of the name plate data (112 data) and a power test and/or looking for harmonics like pole pass frequency I can determine for instance RPM accurately and torque via kilowatts since resistive stator losses are very low. So calculating horsepower and/or comparing say % slip to current can check efficiency. Lots of ways to do this calculation but unless I’m looking at the load torque which I rarely need to do (just look at power or current) there is not much value in doing this. Biggest thing is rotor bar testing is only valid under reasonable load (40%+). There is an offline rotor bar test (rotor influence test) but it reacts to ANYTHING, even innocuous casting voids. It is an indicator, not a performance test.
The old hi pot and step voltage tests are rarely used if ever. DC hi
pot in particular causes damage on equipment in service. See IEEE 400 which is the hi pot test that recommends not doing it. It doesn’t show anything that insulation resistance and surge/inductive tests won’t find so it’s not a useful test. Surge can be used to initiate partial discharge safely. Insulation resistance voltages are high enough to overcome system capacitance without risking damage.
Unless you need something really detailed like calculating starting time accurately there is no need. This is strictly the realm of the EMTP and SPICE crowd. I’ve used it maybe twice in 30 years. Steinmetz is a 5 parameter model. IEEE 112 adds core losses for a 6 parameter model. You can estimate most of these with name plate data only if it’s a NEMA or IEC motor without IEEE 112 dyno tests.
Our shop goes to 23,000 HP in house, more if we need to. We have the clean room, test stand to 13 kV, etc. We do several wound rotor motor rebuilds over 1,000 HP plus we have at least 1 or 2 others in process all the time with a dedicated large motor crew. I’m the chief electrical service engineer so it’s my job to do most field tests.
The standard battery of tests we run on ANY motor and that most motor shops run.on all motors is the “Baker” or “PdMA” sequence offline. Every incoming motor gets this. If the first test or two isn’t getting results I automatically run this test because it saves a lot of time over doing the same tests by hand. This is ohms on coils, inductance on coils or surge comparison, insulation resistance and PI, capacitance to ground. We check by comparisons (resistive unbalance for instance) or to historic or known data. This catches about 95% of defects. It can be trended in the field with a field test or used for diagnostics. After this test it is known if it is safe to start. Up and running we can look at power, power quality, and impedance. Voltage/current unbalance is a common issue. Harmonics very rare but get blamed for a lot of things. We can also look at rotor bars (side bands) and eccentricity (swirl pattern) while running. Some newer relays show this or external testing can find it. Partial discharge is being touted but I only know of one case where it might be useful.
Finally another basic test is no load currents. Useful as a quick test for installation errors as well as checking rotation at the same time. Things that the offline test can’t verify easily.
Vibration has already been mentioned. This does a better job on mechanical diagnostics. We also do ultrasound on occasion for noise complaints and ultrasonic greasing with a UE 9000 (documenting).
On truly large generators like hydroelectric there are some additional tests that check for things like water leaks. It’s such a specialized business and not in our service area we don’t do this.
There is a little data you can get from a locked rotor test too but as a field test we can do this in across the line starting if we can start that way. It’s not a common test to use. It is mostly looking for contactor issues during startup. I have rarely done this. It’s possible to use it predictively and some of the nicer relays (SEL 749M, 710)’have this test built in.
The rest of the tests are 25/50/75% load tests. With some of the name plate data (112 data) and a power test and/or looking for harmonics like pole pass frequency I can determine for instance RPM accurately and torque via kilowatts since resistive stator losses are very low. So calculating horsepower and/or comparing say % slip to current can check efficiency. Lots of ways to do this calculation but unless I’m looking at the load torque which I rarely need to do (just look at power or current) there is not much value in doing this. Biggest thing is rotor bar testing is only valid under reasonable load (40%+). There is an offline rotor bar test (rotor influence test) but it reacts to ANYTHING, even innocuous casting voids. It is an indicator, not a performance test.
The old hi pot and step voltage tests are rarely used if ever. DC hi
pot in particular causes damage on equipment in service. See IEEE 400 which is the hi pot test that recommends not doing it. It doesn’t show anything that insulation resistance and surge/inductive tests won’t find so it’s not a useful test. Surge can be used to initiate partial discharge safely. Insulation resistance voltages are high enough to overcome system capacitance without risking damage.
StarCat
Industrial Engineering Tech
- Location
- Moab, UT USA
- Occupation
- Imdustrial Engineering Technician - HVACR Electrical and Mechanical Systems
Excellent information.
The only other person I ever hear mention the name Steinmetz is Eric Dollard. Very key.
I’ve always heard it called the Steinmetz model or “induction motor equivalent circuit” and most people use the terms interchangeably but there is a subtle difference. It matters when the customer asks for the circuit parameters.
Motor shops generally are EASA members and follow their standards. They may be EASA certified shops or they may be third party inspected by others. For instance we were ISO 9000 for automotive plants before EASA certification existed. UL Listing and nuclear are two others (we have both). EASA is based on motor shop standard procedures and acceptance criteria are based often on extensive empirical testing. IEEE 112 is used more for the very long calculation procedure than the test procedures.
As with a lot of IEEE standards it generally describes things at a high level. EASA procedures and results are most often used in a more direct way (diagnostics, performance testing, and quality control). In our shop EASA IS our standard procedure. You can’t use IEEE 112 in that way. So when I describe what we do I’m really referencing EASA. But if your intent is to plug the data into one of the circuit solvers (EMTP derivatives or Spice derivatives) such as PS-CAD or EMTP-ATP, you need the IEEE 112 or Steinmetz calculated values. That software will not accept EASA results directly. You could theoretically monitor the values over time and flag any significant change in for instance measured stator inductance or resistance but in practice we do this based on comparative metrics. So we can calculate for instance resistive or inductive unbalance (NEMA method) to detect turn insulation failures as opposed to say a change in the 3 phase average values used by IEEE.
- Status