ISOCH & DROOP Generators Transient Behavior During Faults

[sykimk]

Hello,

I am working on transient stability study for my project and I have a question on ISOCH & DROOP generator’s behavior when all these are running in parallel during fault condition.

When you open the attached file, you can see two graphs i.e. generators’ Pe & Pm during a fault and post fault conditions.

This is an island power generation and there will be NO grid interconnection. All these machines are GTG and the same size and have the same parameters. Only one generator is set at ISOCH mode and the other two generators are set at DROOP mode with 4% setting.

Here are my questions.

1. After fault clear, I am wondering why its transient recovery patterns after 3PH fault clear are quite different between ISOCH and DROOP mode generators (page 1).
2. In generator mechanical power graph (page 2), I am wondering why generator mechanical power of ISOCH generator drops so faster than that of DROOP generators during fault condition. It would result in operating of reverse power protection leading to ISOCH generator trip.

Can anyone help me to understand why ISOCH & DROOP generators in parallel show a quite different transient behaviors during a fault?

I think I should recommend either ISOCH or DROOP mode for all generators but I am quite bias toward DROOP mode for all generators since ISOCH mode in all generators would not properly set to the same frequency (maybe impossible..) and it may result in operational problems. Did I understand correctly?

 

[sykimk]

Oops!!! graphs are not properly attached.
Please refer to the following links.

 

[ScottyUK]

A generator with a fault on its terminals tries to speed up, causing a response from the governor. The isoch machine will have a more aggressively tuned governor in order to regulate speed, and if the speed rises above nominal the isoch machine will tend to close its governor to minimum fire more rapidly than the droop machine will. That's something of a guess because I haven't seen the governor control loop, but I think it's a reasonably likely explanation.

Also... how good is your AVR model? AVR behavior has a massive influence on how the machine behaves during the fault.

The big utility-class machines which I'm familiar with should have the major governor loops (typically for control of shaft speed, back end temperature, and generator power) tuned well enough and with appropriate limiters to prevent a trip during a fault and during recovery from a fault. It's not exactly normal for them to operate in isoch mode but isoch capability is required for black start stations and it is rigorously tested by the transmission system operator. The other modes are present in some guise on all grid-connected gas turbine sets.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[nawao]

There is not problem in an island power generation system to operate all the generators is isoch mode if adequate governor contol is provided. In a fault condition, if adequate selectivity coordination is made, all generators will react the same way and will still in line until the fault is cleared by the upstream protections.

 

[nawao]

Sorry, I mean downstream protections

 

[waross]

You can use an Iso load control panel that controls the throttles of all the generators, but good luck running individual governors in iso mode together.
The problem at hand;
Issue one: As Scotty says, a terminal fault will reduce the load on the generators.
Issue two: The first response of a classic iso governor is a droop response. But depending on the loading, the droop related set point may not be the same as the other generators. After the first droop response, the iso governor detects the frequency error and makes a correction by resetting the droop set point to a value that will result in the proper frequency.
Said another way; The settings of the droop governors determines the load that each picks up. The iso governor picks up the balance of the load and controls the frequency.
You may have machine # one set to pick up 80% of its capacity, machine # two set to pick up 80% of its capacity and the iso machine setting itself to pick up the balance of the load.
Now you have a fault that may increase or decrease the load depending on the location, and an iso governor responding in droop but the droop value may not be the same as the other machines. Then, the iso machine starts to reset itself.
One power graph may not fit all faults or systems.
I watched helpless for over a year as the dealer techs tried in vain to run two 1.2 MVA machines in parallel in iso mode. They threatened to void the warranties if anyone else touched the settings. They could not even parallel for less than a minute to transfer the load from one machine to the other.


Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[nawao]

Woodward Digital Governors 721 were used in Lagos Airport to parallel six MAK generators in isoch mode without problems. Load share lines were interconnected between all digital governors. All six gens reacted adequately to large package loads like the Airport chillers. No problem as well with faults cleared in time by the corresponding feeder.

 

[ScottyUK]

If you have load sharing then you are no longer running each machine as an isoch machine, the machines are running as an isoch group: there is a subtle difference. Six true isoch machines will not load share and will behave as Bill describes.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[waross]

It is nice to know that there are governors available that have provision to be interconnected as an iso group. Thanks for the information nawao and thanks for the support Scotty.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[esee135]

Question - What is your time constant/inertia value for your governor reaction? It appears that your mechanical power is able to drop and recover 14MW in less than 1 sec.

Also, we would need a plot of the frequency response, unless you have a power dominant governor, which is very different from a speed governor.

Transiently, you are dealing with inertia. In general isoch governing will experience more deviations than a drooped governor (high inherent gain factor), generally 1/Rp is a typical gain for a droop governor Higher droops will make the governing system lazier, or more willing to share. During a fault recovery, you should notice a nominal frequency lead of the isoch governor, and a drooped speed set-point of the drooped governors.

 

[eddiet1]

You ask good questions concerning the different responses from the Isoch and Droop governors. Rarely do I get to see a lot of a response during a fault.

As mentioned by esee135, can you get the frequency plot? Also, advise on the inertia's and types of governors. That can make a big difference.

Also, how are you measuring mechanical and electric power? I would expect them to be the same, yet the graph indicates they are different (throwing some suspicion on the instrumentation and data collection).

 

[waross]

Another factor is the pu load on each machine when a falt occurs.
Assuming that the droop settings are 3% on each machine:
If the load is 100% the governor setting and the no load frequency will be 61.8 Hz
If the load is 20% the no load frequency will be 60.36 Hz.
This would apply to either droop or isochronous until the integral function has time to react.
Load levels on the various machines may well be making more of a difference in the individual transient behaviors than droop or isoch response is making.

Bill
--------------------
"Why not the best?"
Jimmy Carter