Generator EF question

[veritas]

In the diagram gen is high impedance earthed with the EF current limited to 10A (sorry, not shown on SLD).

I have a concern regarding EF protection on a steam turbine generator system as per the attached. For an EF at F1, the generator EF relay trips in 0.5s. This leads to a turbine trip, which in turn leads to a reverse power condition which in turn leads to a GCB and field CB trip. Thus it could be several seconds before the GCB trips. The fault is now on an ungrounded system. There is a neutral overvoltage relay fed from a broken delta VT secondary as shown. Trip time is 3s.

Question – is it necessary to have a 3s delay? There is nothing else to grade with. Gen trfr is sometimes back energised via the Aux trfrs. Would that be an issue?
I am concerned that the 23kV bus, two auxiliary trfrs and the gen trfrs are subjected to the voltage rise on the healthy phases longer than what is necessary. Danger is that fault could lead to another flashover which will be much more severe. Another concern is arcing faults with unearthed systems and the transient overvoltages.

Are my concerns valid or is 3s okay? Any other dangers? I was thinking of bringing it down to 0.1s.

 

[mgtrp]

What you have seems a very typical system. Since it is on a high-impedance earthed system, your equipment should be rated for the full phase-to-phase voltage to earth, and so should have no problems withstanding 3 seconds at the increased voltage during an earth fault. If you're especially concerned, you could check the equipment ratings.

This means that the consequences of delayed clearing of the earth fault are relatively low, since the equipment is rated for the overvoltages and the fault current is negligible--it is limited to 10A initially by the resistor, and then something similar or slightly lower once the generator breaker trips, the exact current being dependent on the bus and transformer capacitance. This shouldn't cause too much heating/damage for the few seconds that it hangs on for.

Since this consequences of delay are low, while the consequences of false trips are likely higher, some added delay to the earth fault proteciton is generally accepted to give added security. For comparison, I typically coordinate my protection so that the longest neutral overvoltage element is 5 seconds, and use 0.5 to 1 second coordinating time interval between different protection devices.

Please note also that the bus 59N element will pick up as soon as the earth fault occurs, even though the system is earthed via the generator. The 10A earthing resistor is not enough to significantly affect the neutral voltage during a fault as compared to a completely unearthed system. Thus 3 seconds is the total clearing time: the bus 59N won't wait until the generator circuit breaker has tripped.


One other point of interest in a more general sense is with generator 59N elements, which are typically set quite sensitively so that they can detect faults nearer the neutral end of the windings. These may actually pick up during earth faults on the transmission system (I believe this is due to capacitance between the HV and LV windings in the transformer causing a shift in the neutral point of the ineffectively earthed system). As a result, the generator 59N protection must be delayed to coordinate with the transmission line protection. The bus 59N element is likely set less sensitively since any earth fault in its protected zone should cause a larger neutral voltage--hopefully it won't pick up for transmission faults.

 

[veritas]

Hi mgtrp

Thanks for the very thought provoking post. Actually, the 59N element only starts timing once the GCB has opened. Thus the total fault duration is quite a few seconds more than 3s. Also, the installation is around 30 years old. Does this change anything for you?

 

[veritas]

Hi mgtrp

The bus 59N setting is 10V which equates to 5% of the maximum 3V0 output (190V).

Another issue that concerns me (maybe it's a non-issue) is with initial fault inception one has a very low fault current of around 10A. If it is a bus fault, what happens at the fault point once the gcb opens? Does t not become an arcing fault now that there is no earth return path? If this is not the case then yes, I'd be happy to leave the time delay at 3s. Arcing faults are notorious for their transient overvoltages, etc. - and this is what I am trying to minimise.