Impacts of loss of excitation on small hydro unit 3

[bacon4life]

For a 2 MVA synchronous hydro generator, what are the impacts of running without excitation for 15 minutes with the wicket gates closed? How much reactive power would you expect to see flowing into the generator. How much real power losses might occur? Would you expect an in service failure for this size unit from overheating of rotor retaining rings, rotor wedges, rotor iron and rotor winding?

I found typical advice for tripping on loss of excitation with only a 0.2 second delay, so I am a bit confused as to whether it would take many minutes for damage to occur. A previous thread mentioned a steam turbine saw about 0.65 PU for 45 minutes, but also mentioned hydro units would be different than steam units.

The wicket gates have a mechanical failsafe closing system. Although the generator is above the tailrace level, I am unclear whether closing the wicket gates would typically also dewater the unit.

 

[Gr8blu]

Wicket gates are used to manipulate the water flow as it approaches the generator. To produce additional energy, the gates are opened wide. To limit the water flow and decrease the production of energy, the gates are closed. Typically, a sequence of wicket gates will encircle the turbine of a hydroelectric plant. With the gates securely closed, no water will pass through the blades of the turbine. Once the wicket gates open, water flows, rotating the turbine’s blades and driving the operation of the generator.

The cage of a synchronous machine (even a generator at the 2 MVA level) is not designed for continuous operation; it's designed to smooth out transients related to step changes in electrical loading. That means it does not have the thermal capacity to operate for very long before something in the circuit (typically the braze material used to join the shorting ring to the individual bars) fails. This type of thermal damage occurs in seconds, not minutes.

Rotor wedges aren't as likely to fail from a "loss of field" condition - and the rotor iron should be good as long as the rotor does not contact the stator directly. Rotor winding might see damage related to overvoltage (still spinning a wire loop in the stator's magnetic field), depending on how long/fast the coast down takes.

Depending on how quickly the wicket gates close off the flow, the generator may come down to a fairly smooth stop similar to a coast down or "bottom out" on the brake surface and stop rather more abruptly. Once the main excitation is removed, the power factor is going to be abysmal (down around 0.15-0.25 per unit), at least as long as the rotor continues to spin.

 

[waross]

With loss of excitation only won't the alternator continue to motor as an induction motor?
That is until enough brazed connections fail.

 

[dpc]

If the generator trips off line while running, it will go into overspeed for however long it takes to close the wicket gates. It has to be designed to withstand this. If you are talking about leaving the machine on-line, and losing the excitation, it will become an induction motor and will normally have pretty limited capability in this mode of operation. You should consult the instruction manual for the machine and review the capability curves.

 

[bacon4life]

From what I can piece together so far, the generator circuit breaker stayed closed for 15 minutes after the exciter tripped. As far as I can tell from limited very limited data, the unit drew 1.4 MVA. I am trying to figure out how it could have stayed online that long if damage to brazing occurs within seconds.

The water flow appears to have stopped within seconds, with the generator continuing at approximately 60 Hz for the whole 15 minute duration.

The capability curve shows -0.8 MVA at 0 MW. Is the failure mode simply operating at -1.4 MVAr rather than -0.8 MVAr? Or is there a different heating mechanism that occurs once the generator shifts to being an induction motor?

I had heard a rough rule of thumb that inductions draw 0.2 to 0.4 PU reactive power at idle. Does it make sense that a this generator would draw more like 0.7 PU when operating as an induction motor.

 

[edison123]

IEEE 492

7.6.4 Asynchronous operation (field maintained) - (I have personally seen a 50 MW hydro machine completely destroyed in this operating condition)

Operation of a generator out-of-synchronism with partial or full-field excitation maintained, places the most severe type of duty on the unit. Such operation produces heavy surge currents in the stator windings whose magnitude may exceed those associated with the machine short-circuit requirements of ANSI C50.12-1982 and cause serious damage to the winding. Such operation also produces torque reversals that create in many parts of the unit high mechanical stresses of magnitudes that may be several times those produced by rated torque. High induced voltages and currents in the field may cause flashover of the field coils to ground, the collector rings, and of the commutator of an associated exciter and thyristors in static exciters.

For these reasons, although it may be difficult to detect the out-of-synchronism machine, it must be identified promptly and the condition must be remedied. Possible corrective action includes removal of the unit from the system without reclosure. Amortisseur winding damage can be expected if slip frequency operation is prolonged.

7.6.5 Loss of field excitation

Complete loss of excitation on an operating generator results in dangerous overheating of its rotor within a few seconds unless the machine is disconnected from the system. The degree to which this heating will occur depends on the initial load on the generator and the manner in which the generator is connected to the system. When excitation is lost, the generator tends to overspeed and operates as an induction generator. This overspeed normally results in a reduction in load due to the characteristics of the turbine and governor, an increase in stator current associated with low voltage at the generator terminals, and high rotor currents. These rotor currents will flow through the field winding (provided the field circuit is not open), and also through the amortisseur windings and rotor pole faces. The amortisseur winding and rotor pole face currents will cause high and possibly dangerous temperatures in a few seconds.

 

[RRaghunath]

From what I can piece together so far, the generator circuit breaker stayed closed for 15 minutes after the exciter tripped. As far as I can tell from limited very limited data, the unit drew 1.4 MVA. I am trying to figure out how it could have stayed online that long if damage to brazing occurs within seconds.

The water flow appears to have stopped within seconds, with the generator continuing at approximately 60 Hz for the whole 15 minute duration.

The capability curve shows -0.8 MVA at 0 MW. Is the failure mode simply operating at -1.4 MVAr rather than -0.8 MVAr? Or is there a different heating mechanism that occurs once the generator shifts to being an induction motor?

I had heard a rough rule of thumb that inductions draw 0.2 to 0.4 PU reactive power at idle. Does it make sense that a this generator would draw more like 0.7 PU when operating as an induction motor.

Hydro turbine generators draw about 0.2% of rated power from the power system, when made to motor. This is mentioned in IEEE 242. Loss of excitation causes MVAr to be drawn from the power system to meet the excitation requirements of the generator. This is often a problem to the power system more than to the generator as it causes voltage dip in the system and consequential tripping of auxiliaries etc. The problem would be more severe if the hydro turbine is away from the grid.
Considering the Hydro machine has large rotor, it can stay on for many minutes without getting damaged.
The low time setting for trip on loss of excitation comes from the need to minimise the impact on the power system (due to large VAr drawal).

 

[wcaseyharman]

With no load it seems possible that the machine stayed in sync - with a salient pole machine the reluctance torque from the variable air gap plus any residual magnetism could possibly have kept the machine from continuing asynchronously since there was little load. Do you have speed probe data?
The MVA drawn during a loss of field event depends on the stiffness of the grid and the generator reactances and whether the machine is asynchronous or not. If the machine stayed in sync I would guess your Xd would be around 1 per unit or a little over.
I would expect higher VAR loading if it went asynchronous as the transient impedances, which come in to play during an out of step event, are much smaller.
What type of turbine? I’m assuming a pelton as a Francis turbine would have pulled considerable amount of watts (reverse power) churning the water in the draft tube.
My very uninformed opinion is the fact that the machine was immediately unloaded saved your generator.
The fact that your leading MVA capability is less that your machine’s armature limit makes me suspect your machine has an end iron heating limitation. If that is true I would inspect the ends of the stator core to make sure the core ends didn’t overheat and damage the stator insulation.

Just my $0.02 as a power plant electrical engineer.

 

[bacon4life]

Ah, that make sense that the immediate closing of the wicket gates would have made this much less severe than having the turbine continue operating will full water flow.

Casey- I'll ask whether there is enough speed data to determine if it stayed synchronous. I think it is a Francis turbine, but I am not very familiar with the mechanical side of things. The generator is higher than the tailrace, so I am unclear whether the scroll case will fill with air or if it will stay filled with water. Would 0.3 MW make sense for churning water in the draft tube? I only have real power measurements from remote transmission line terminals, so I am not sure if the squiggle in the transmission line flows accurately capture real power flow to the generator.

Raghunath-I had not heard of the concern about voltage dip, but it makes total sense for large units. Our largest unit would drag the voltage down to 0.7 PU. Do you have any references that discuss this reasoning?

Muthu-Thanks for pointing out IEEE 492. The rest of section 7.6.5 indicates that some hydro units remain online without excitation for long enough to manually restore the excitation system.

Some users utilize a loss-of-field relay to trip the generator breaker, thereby removing the unit from the system. This may also initiate closure of the turbine water inlet valves or gates. Some users provide alarm indication only. The manufacturer may require that the unit be tripped if the load is more than 15%. Time can often be saved by running back the turbine inlet water valves or gates to a speed no-load position and, following precautionary checks, restoring excitation and re-synchronizing the system.
Where neither loss-of-field tripping nor alarm indication is provided, the operator must recognize the condition and manually perform the functions described for the relay above. If the loss-of-Þeld condition has persisted for some considerable or unknown length of time, the rotor should be inspected before operating again.

 

[wcaseyharman]

Maybe - the onetime I am aware of where we motored our francis turbine unintentionally it pulled about 6 MW on a 50MW machine, so yeah, possibly.

We had a 50MW machine lose its field due to a failure of the exciter leads and the amount of VARs it pulled in was astounding - I think it was close to 100 MVAR - luckily our loss of field trip took the machine out within 3 seconds and the machine was fine - we performed an internal inspection and basic electrical testing and found no issues. I was happy that the machine next to it on the same bus rode through the whole thing and all my relay and exciter settings worked great.

 

[bacon4life]

The speed data was only available for the period after the generator breaker finally opened, so all I cannot tell if the unit stayed at synchronous speed or if it was operating with slip.

 

[edison123]

"The rest of section 7.6.5 indicates that some hydro units remain online without excitation for long enough to manually restore the excitation system."

Manually intervene is the key point here. If your machine ran for 15 minutes with no operator action, better to check the rotor health.

 

[RRaghunath]

Ah, that make sense that the immediate closing of the wicket gates would have made this much less severe than having the turbine continue operating will full water flow.

Casey- I'll ask whether there is enough speed data to determine if it stayed synchronous. I think it is a Francis turbine, but I am not very familiar with the mechanical side of things. The generator is higher than the tailrace, so I am unclear whether the scroll case will fill with air or if it will stay filled with water. Would 0.3 MW make sense for churning water in the draft tube? I only have real power measurements from remote transmission line terminals, so I am not sure if the squiggle in the transmission line flows accurately capture real power flow to the generator.

Raghunath-I had not heard of the concern about voltage dip, but it makes total sense for large units. Our largest unit would drag the voltage down to 0.7 PU. Do you have any references that discuss this reasoning?

Muthu-Thanks for pointing out IEEE 492. The rest of section 7.6.5 indicates that some hydro units remain online without excitation for long enough to manually restore the excitation system.

I failed to mention that the 0.2s trip may not be relevant for your generator considering the small size and the fact that it cannot impact the grid / power system. You may like to review the protection settings to prevent recurrence.
IEEE Std. C37.102 for Generator Protection has some discussion on the subject.