Reactive Droop Compensation vs PF & AVR 1

[ffllyyeerr]

I am working on a grid tied, 200 KVA, 2400 V synchronous generator hydro powered (Pelton Wheel) currently inop with symptoms pointing to a failed static exciter which is to be verified shortly.

The generator is old without documentation. The field resistance measures 3.6 ohms through the brushes and slip rings. This has prompted the exciter manufacturer to indicate the present 100 Amp unit is oversized for the generator and has recommended replacing it with their now available 50 Amp unit. This makes sense to me as there has been ongoing brush/slip-ring arcing and other problems possibly due to overexcitation as there is no excitation limiter in the system. When online, excitation is controlled via a VAR/PF controller and the unit operates normally PF so as to maximize KW output.

The newer proposed 50 Amp replacement exciter comes with a different regulator that operates on reactive droop compensation.

Not sure about how this might work. One paper suggests droop compensation may not be appropriate for a small generator due to over/under excitation when it attempts to follow abnormal grid voltage variations. Changing to droop compensation also involves adding another special built CT due to the voltage level and the 30 VA burden. From a limited understanding of droop compensation it appears to follow a straight line relationship between terminal voltage and reactive power - difficult to see how this is going to maximize KW output.

Looks like a fork in the road here - which way to go? Any help would be appreciated.

 

[waross]

Will the grid operator allow the change.

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

 

[ffllyyeerr]

That's a good question but I suspect so as it's a rural electric co-op and their interconnect requirements are minimal to none.

 

[ScottyUK]

See if you can get the machine's Vee curves to help determine whether the smaller excitation rectifier is going to be big enough to maintain full machine capability. I don't see much sense in crippling a generator's capability for the cost difference between a 50A rectifier and a 100A rectifier. If you are confident that the machine operating conditions will never change then it might be OK, but things do usually change if left long enough.

(edit - fixed typo )

 

[catserveng]

Basler wrote a very good paper on this issue several years ago, a copy is attached.

You should be able to apply a VAR/PF controller, like a Basler SCP-250 to your new regulator if you want to operate in Power factor control mode.

Hope that helps, MikeL.

 

[MarcMarc22]

Hello ffllyyeerr

Maybe you already know, but just to add to your knowledge, a reactive droop compensation behave in a way that will prevent / reduce MVars sharing between units. The typical case where this will be used is where you have 2 units sharing the same step-up transformer. (2 or more, I saw 5 units at one place) Therefore, you will end up with 2 generators controlled by 2 different excitation systems trying to control the same output voltage (since there is no transformer impedance between the units, both units see the same voltage at the generator PTs). it would be about the same as having 2 guys trying to drive the same car at the same time. Here is what would happen if you do not have any mean to prevent it : (either reactive droop compendation or Cross current compensation or any other idea that I do not know). It's the opposite of what we call Line droop compensation where you would actually boost a unit support to the grid by compensating for the stepup transformer impedance (in the example below, just swap the droop sign and forget one unit but the idea is the same).

In this example, lets assume that Unit 1 is reading a slightly higher generator voltage than the setpoint voltage (assuming the exciter is in Automatic voltage regulator control and that the setpoint is 1pu ( 100% of the terminal rated voltage) and that Unit 2 is reading slightly lower voltage than the exciter setpoint in the same operating condition. (Auto is important here since this behavior does not happen in Manual or FCR control).

Unit 1 exciter will try to reduce the voltage, since U1 and U2 see the same voltage, the second unit will try to counteract this by increasing the voltage. In a matter of seconds (closer to one than 10 hehe) one unit will seat on the under excitation limiter (if you have any) and the second will seat in the over excitation limiter (OEL if you have any). Otherwise both units will trip.

One way to get around this is to use a reactive droop compensation. The way it works depends on the vendor. Typical value for situation like this will be 5% or 0.05pu. As per IEEE a reactive compensation droop will have a positive sign, however, some vendors (ie ABB) will swap the sign to facilitate the implementation. Basically, they will multiply the detla MVars x droop and it will be added to the setpoint. So in the example above, the unit that was absorbing MVars will have : -Q X -Droop = POSITIVE VALUE added to the summing point.( Example : 1.00 + (-0.2q * -0.05droop) = 1.01) It will therefore try to increase its setpoint until it stop pulling MVars. The second unit will have +Q X -Droop = NEGATIVE Value (1.00 + (+0.2 * -0.05) = 0.99 and it will therefore reduce its setpoint until it stop pushing MVars out. After few wiggle, it will stabilize to a new operating point stable where both units stop fighting.

What is the downside of it? The unit does not know if the MVars it's pushing are going to the grid OR to the other unit and therefore in the event where there is a fault on the grid and that your units needs to react to support the grid voltage, it will actually go the way around and not support the grid or at least not as much as it should.

Solution : Few vendors came with a solution where both exciters talk together and knows what the other unit is PUSHING / Pulling and therefore it will share the grid support with the second unit. For instance ABB is using a solution called CCC or cross current compensation.

About the PF/MVars controls, this is typically not allowed by the grid operators everywhere in Canada / USA since you will provide close to nothing in terms of voltage support if the grid needs it. The exciter will try to keep a constant PF or Amount of MVars out instead of Pushing / Absorbing the requested MVars to keep the voltage constant. Since this unit is incredibly small, you might have special permission that I'm not aware of.

About reducing the size of the exciter, having a big exciter guarantee you what we call the ceiling capability (the ability of a unit to exceed is nominal field current for a short period of time to help the grid to ride through a disturbance). So I'm not sure why reducing the system will help with the problem you have, at the end of the day, you decide what is the nominal at the exciter, that the bridge can do way more is not a problem, you just pay for something that you might not need. If your Ifg rated is 40A, having a 100A system is no concern, unless there is something that I'm not aware of.

Hope it helps a little.

 

[edison123]

Nicely explained, Marc.

To OP, having a bigger exciter/excitation system doesn't necessarily mean over-excitation. The excitation is controlled either by the AVR or the operator.

Muthu

http://www.edison.co.in

 

[crshears]

What Marc and Muthu said; in my view using excitation limiters with an inverse current characteristic is the way to go. Under-sizing the exciter is NOT the answer.

CR

"As iron sharpens iron, so one person sharpens another." [Proverbs 27:17, NIV]