Simultaneous 4kV Motor Failures 2

[joepower]

We experienced the simultaneous failures of two 4kV fan motors on a high resistance grounded system. They were on the same switchgear bus, a 6000 HP fan (located outside) and a 600 HP fan (inside) – both less than 300 feet from the bus. The 6000 HP fan has surge protection at its terminals – the only other surge protection is on high side of main transformer. Our control system shows the loss of the 6000 HP fan one second before the 600 HP fan, however, the points have a 0.5 second scan time and it is too close to call if that is real or not. The 6000 HP fan went on instantaneous at 8920 A (740 FLA) and one end coil was visibly blown out, and the 600 HP fan on time overcurrent at 1160 A (76 FLA). No abnormal weather or operating conditions. A similar situation occurred previously wherer we got a 2-fer in failures. Any ideas?

 

[busbar]

electricpete, the quoted resistor values seem logical: 20 amperes and 400Ω correspond to 8kV line-to-neutral. It also figures at 160kW, so that’s more likely a short-time {id est, 10-second} rating and not intended for continuous dissipation. This would suggest inverse-time tripping and not an annunciation-only scheme. A fraction of the resistor current would be desirable for ground-fault relaying...a 5-10A primary pickup for zero-sequence CT scheme is reasonable—or possibly for a residual-CT configuration. This corresponds reasonable relay sensitivity weighed against minimal thermal stress to 5-mil copper-tape-shielded MV cable.

 

[electricpete]

busbar - I'm pretty sure that's what we have (400 ohms 20A). Not sure if it falls exactly in the definition of high-R grounded. We also have a 4kv system with neutral grounding resistor and similar relaying (no alarms). I will double check both resistance values.

Hi Gord. I will try the system with inductance and I expect higher voltages as we have said.

The reason I turn the fault off at the peak of Ea is because that is how I would achieve the highest peak phase-to-ground voltage (Vb=En+Eb) approaching twice peak nominal line-to-neutral voltage (in my simulation without inductance).

Compare two HiR cases in my simulation (ignore the LowR cases):

(Note for the following discussion I have assumed the magnitude of nominal line-to-neutral voltages Ea, Eb, Ec vary between –1 and 1.)

Case 1 – graph labeled 'HiR_3_cycle_short' - represents continued application of the fault for 3 cycles. En(t) remains very nearly equal –Ea(t). Vb(t)=Eb(t)+En(t) = Eb(t) –Ea(t) which has maximum value of the line-to-line voltage which we know is sqrt(3)~1.7

Case 2 – graph labeled 'HiR_Intermittent_Half_Cycle_Short' - represents removal of the short when Ea reaches it's positive peak and En reaches it's negative peak (at t=8.33 msec +k*16.7msec). At approx 60 degrees later (2.8msec later) at time = 11 msec Eb will reach it's negative peak (-1) and Vb will be Vb= Eb +En = -1 + ~-1 ~ -2 where En has decayed only slightly from it's value of –1. (looks like –0.84 in my graph giving –1.84…. would be higher magnitude for slower RC time constant).

Case 2 (where the fault is removed at the peak) gives a higher maximum line-to-ground voltage (for the case with no inductance)

 

[jbartos]

Suggestion: Visit

http://www.ecmweb.com/ar/electric_choosing_grounding_options/

for high resistance grounding.
Above approximately 4.16kV, especially on 13.8kV the high-resistance grounding cannot be applied because of high charging currents. These are obtained by calculations. The medium resistance grounding is used. Sometimes, when the charging current is too high on 4.16kV, the medium resistance grounding is used. This may be the case, if there happen to be surge arrestors/capacitors protecting rotating machinery. In some cases, the surge protection is implemented upstream of the system grounding location, e.g. on the transformer primary side.

 

[jbartos]

Suggestion: Visit

http://www.ieee-pcic.org/Subcommittees/safety/1999_agenda_items/

and click on bowen.ppt
etc. for more info on high and medium resistance grounding

 

[electricpete]

I took a closer look at our system.

I did see that motor surge caps used on 13.2kv motors are fractions of a millifarad… approx 1000 times more than motor capacitance which are on the order of microfarads.

13.8 kv system – 20ohm resistor, 400A ground current, 10 sec rating. Since ground current >10A I believe this is considered a LOW-resistance grounded system => We have surge caps on our 13.2kv motors. We have no calculation to demonstrate Rn<Xc0. Perhaps this is not required for a LOW-resistance grounded system ?

4.16 kv system – 240ohm resistor, 10A ground current, 60 minute rating. Hi-resistance grounded system => We have Calcuation to demonstrate Rn<Xc0. We have no surge caps on our 4 kv motors, which is why I previously thought we left surge caps out of our calc.

Ou 13.2kv motors have 51G ground trip. Our non-critical 4kv motors have 51G ground trip. Our critical safety 4kv motors have only 50G ground alarm. I’m not sure the reason we use time overcurrent 51G vs 50G for our motor ground trip functions. Also I was surprised to see the long time rating on our 4kv neutral resistor... I haven't looked closely at protection applied to the bus/transformer.

 

[joepower]

Original poster here. For the high resistance grounding, the transformer feeding the bus has a single 10 kVA, 1-phase, 4160/120V transformer in its neutral with a 0.866 ohm, 8 kW resistor on its secondary with a 64 relay. The bus the motors are fed from has a similar setup but with three single phase trasnformers connected grounded wye - open delta with a 4.3 ohm, 10 KW resistor for selectivity on ground faults with the transformer. Still researching surge protective device parameters.

 

[electricpete]

Two different grounding devices on the same bus?

 

[DanDel]

joepower, I have some questions about your last explanation.
What you describe for high resistance grounding appears to be a neutral grounding transformer with a resistor in the secondary circuit. Typically in this setup, there is a 59(overvoltage) relay across the resistor to sense a GF in the primary. A 64 relay senses current. Is the relay in series or parallel with the resistor?
For the three single-phase transformers connected grounded-wye, open-delta; open-delta is for a two transformer system. Possibly, is the resistor in series with the three secondary windings of the delta connection? Where and what type of GF protection is present?

 

[electricpete]

Looking at the bus transformrs, I suspect they are for relaying (grounded wye primary, broken-delta secondary). If primary voltages to ground are balanced then the sum of the three voltages induced in secondary are balanced.

I think the ground fault current produced by that set of but pt's will be very low because the secondary relay is likely very high impedance (only magnetizing curent flows in primary if secondary is open-circuited).

 

[electricpete]

Referring again to your bus pt's:
IEEE 242-1986 (Buff Book) section 4.12 - Overvoltage relay (59) - Section (2) - Ground Fault Detection - &quot;One method measures the zero sequence across the corner of a broken delta secondary of three voltage transformers that are connected grounded wye..... a resistor may be required accross the relay to prevent damage due to ferroresonance.&quot;

Can you confirm you have a broken delta, not an open delta?

Is there a 59 relay associated with the bus PT's?

Is the 4.3 ohm resistor in parallel with that 59 relay?

 

[RRaghunath]

I agree for systems with large capacitance that would rule out the possibility of limiting the earth fault current magnitudes to a low value such as 10A, it is still possible to go in for high resistance grounding system. The difference is that the current magnitudes could be 20A or so as the design demands and the motors in the system have to be tripped on detection of earth fault, as there is risk of core burnout.

The above method fails to offer the benefit of continued process and thus is like low / medium resistance grounding system. In fact, it is less preferred considering the difficulty in detecting reliably such low fault currents compared to the other systems. Raghunath

 

[busbar]

joepower, a single-phase grounding transformer is typically applied where there is a physical neutral connection, such as a serving wye-secondary bank, or a dedicated outboard zigzag or grounded-wye/broken-delta grounding bank applied to a delta system that has no physically-accessible neutral bushing.

I believe the differentiation of high-resistance versus low- or “medium”-resistance grounding lies in having a continuous-rated resistor and ground annunciation, and generally excluding tripping functions—such that manual intervention is needed to reset the annunciation.

The applicable overvoltage relay should be able to withstand 1.73 p-u continuausly, but be able to pickup at under 0.1 p-u, and typically restrained for third voltage harmonic.

 

[joepower]

It is broken delta, and the resistor is in parallel, but teh relay is called out as a 64. However, it appears from further research that only the transformer neutral ground was in the circuit as the bus tie to teh section of gear with the broken delta configuration was open and the buses were independently fed.

 

[busbar]

A device 64 in protective relaying systems is intended to be a generator-field ground relay per ANSI C37.2. For the application the OP describes, there has been either a misapplication or typo. A usual/slang term for ground detection in ungrounded/high-resistance-grounded systems is “59G”, for a [ground]-overvoltage relay.

Now, if bus transfer occurs between a non-high-resistance-grounded and a high-resistance-grounded system, unreliable {or certainly varying} operation may occur.

 

[electricpete]

We have the same situation on our generator isophase bus. When powered from the generator, the system is high-resistance grounded. When backfed through the generator stepup (delta-winding on low-side), it is ungrounded. We have ground protection sensing generator neutral current, and an additional broken-delta scheme to provided continued ground proteciton on the isophase bus for use when backfeeding with generator off-line.

We also call ours 64B, by the way. And we have 64R and 64S on the generator neutral and 64F on the field. I don't know what's up with the numbering.... call it what you like.

 

[electricpete]

But I see now that Joe is saying that those bus pt's and relays were not connected during the fault. Sorry for the tangent.

 

[jbartos]

Suggestion: Reference:
J. Lewis Blackburn, Protective Relaying Principles and Applications, 2nd Ed., Marcel Dekker, Inc., 1998,
Section 7.4 Ground-Detection Methods for Ungrounded Systems
Section 7.4.1 Three-Voltage Transformer
For ballast resistors that are used to reduce the shift of the neutral from either unbalanced excitation path of the voltage transformers or from ferroresonance between the inductive reactance of the voltage transformers and relays and the capacitive system. Typical resistance values across the secondary windings are derived from experience and are shown in Table 7.1.

 

[electricpete]

It does seem like Gord and busbar have identified a likely problem.

I have read where 0.5 micro-Farads per phase is typical value for the 3-pole surge caps in 4kv, as Gord says.

And going through the calc, you get
V_LG = 4KV/SQRT(3) ~ 2300
xc = 1/2/Pi/c = 5300
I per phase = 2400/Xc ~ 0.5
Ic total = 3*I ~ 1.5A.

Compare to your 0.866 ohm resistor. On the primary it looks like 0.866*(4160/120)^2 = 1040 ohms.
Resistive curent = 2300/1040 ~ 2.3A.

You have used up more than half of your allowed capacitive current already, exactly as Gord said. So that does sound like a very fruitful area to pursue.

Some other wandering thoughts:

Simultaneous failure of two motors must be connected through the power system. Sustained voltage excursion, surge (from external source of from first motor failure) or one motor feeding the other as a generator and fault not cleared quick enough. If you have 0.5 sec voltage magnitude recordings on the computer might help shed some small light (waveforms from fault recorder or digital relay are a lot nicer of course).

Even though you say you have ruled out power system problems, are there capacitors nearby which may have been switching? Also utility may be able to provide you info on known activites/trips at the time of your event.

A fault in the first motor creates a steep-front wave which travels and can cause other damage. (it is well known that if you have a flashover during a hi-pot test you stand a good chance of damaging motor or cable at other locations from the traveling wave). One question that crosses my mind... if the large motor failed first internally, are its surge caps effective at preventing the internally-generated surge from LEAVING the motor to the power system? Since the standards make a big point of requiring a surge cap to be connected very close to the motor, I would say maybe not, because those caps on the large motor are not close to the 2nd motor that failed.

Surge caps generally have to be close to the protected equipment and have good soild ground connection to the motor frame ground (don't rely on cap frame ground).

If a turn or ground failure appears to be on the first coil (especially first turn, connection end), that increases the likelihood of cause being surge-related.

Turn failures typically result in a lot of copper melting because they don't trip until they go to phase or ground.

Can they narrow down turn insualtion failure or ground instulation failure and how far in from the terminal? If multiple failures look at how they relate to each other in physical space and electrical connection.

If you see heavy evidence of movement of end-turns in one phase of the smaller motor, I would say that may would support a theory that the 2nd motor failed due to the fault current that it supplied to the fault in the first motor.

Some types of motors are very susceptible to turn insulation problems, possibly even in the presence of surge protection. Those are form wound motors which do not have dedicated turn insulation (they use the strand insulation to serve as the turn insulation).

Our 4kv high-R grounded system does meet the criteria for capacitive current < resistive (at least on paper). We have also had simultaneous failure of two 4kv motors on that system. Cause unknown. (I'll have to look at the report to refresh my memory… it was before my time).

 

[electricpete]

Some more discussion.

My scenario of 2nd motor acting as generator feeding fault in first motor (not cleared promptly) would be expected to lead to instantaneous trip of 2nd motor, if not both motors. That does not appear to be the case for yours.

It will be interesting to know how far these two motors are from the bus. Cable can act somewhat like a surge capacitor. If there is enough capacitance in those cables, maybe we rule out most of the scenario’s associated with surge?

Also, if you are giving thought to removing surge capacitors from your motor, there are two things that may help support that decision:
#1 – the capacitance of the cable between motor and the bus.
#2 – the use of dedicated turn insulation on the replacement motor.
Also the type of breakers used may affect your decision (vaccum contactors suspected to create worse transient than electromechanical breakers).

One other thing to note is that complete surge protection requires not only capacitors (to reduce dv/dt) but also arresters (to reduce the peak). Capacitors need to be close to the load but arresters can be at the source of the switching or anywhere betweeen there and the motor. I think most installations provide arresters on the transformer hi-side and possibly low side to protect from surges coming into the facility.

I think that the type of voltage oscillation resulting from excessive capacitance connected to your system results in temporary overvoltages threatening the ground insulation, but not what we would call surges that threaten the turn insulation. The rise time of surges affecting turn insulation would be below approx 2-5 microsec. LC oscillation will have a resonant frequency probably much lower. At least that’s what I’m thinking now. So examine your 2nd smaller motor carefully to see if you can discern turn insulation damage (signifcant melting of copper).

 

[electricpete]

For 0.5 sec scan time on both motor points, and difference in time recorded at 1.0 sec, I would expect the actual time between signals reaching the computer was between 0.5sec and 1.5 sec. 0.5 sec is a long time considering typical relay and breaker response times and does suggest strongly that the large motor went first.

To round out your calc of capacitances you want to add in your cables and motors.

Cable capacitance can be found from physics of cylinder within cylinder.
Substituting in unit conversions gives the following formula from Okonite:
C= 7 * 35*S1C / LOG(D/d)
Where
C = Cap in picoFarads per foot
S1C = dielectric constant from 2.8 to 3.5 (use higher value for conservatism).
D = Outer Diameter
D = inner diameter (usually conductor diameter).

Our 4kv cables range from 0.15 to 0.122 microfarads per 1000 foot per phase.

Motor capacitance to ground can be determined during Doble test. Westinghouse provided us some typical values ranging from 0.037 microfarads for 400hp 2-pole to 0.149 microfarads for 1500hp 6-pole motor (4kv motors).