Generator AVR Operation and Response

[ThePunisher]

We have a 5.5MVA, 4160V generator connected to a bus with some static loads and a couple or 4kV motors. On an island mode scenario, if we start the largest motor in place, the load flow shows a surprisingly 42% voltage drop on the 4160V MCC bus. The generator terminal voltage set on the study is 1.036P.U. nominal voltage. The load flow calculation does not dynamically calculate exciter/AVR response, but considering even a 1.10 P.U.terminal voltage will not help.

The generator datasheet indicates maximum allowed starting load of 10.2MVA at 10% P.F. with maximum voltage drop of 15%, voltage drop at sudden increase of rated load = 7% and voltage rise at sudden drop of rated load =9%

I would appreciate some experience and expertise on what the above generator specs means?

 

[waross]

Is this a measured or calculated 42% voltage drop.
What size are the motors.
What type of motors?
What type of excitation; Static, rotary exciter, rotary exciter with a Permanent Magnet Generator powering the AVR?

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

 

[JJHorak]

I am not a good resource, but I am curious how you calculated the voltage drop. I mean, what imedances (motor and generator) you are using?

 

[ThePunisher]

Hi waross, Timesabroom,

I will provide more details tomorrow since its 10PM here and the SLD diagram with load flow results were left in the office.

The generator is 5.5MVA, 4160V, 763A, 0.68 P.F., 900RPM, XD"(saturated)=4.8%, XD' (saturated)=8.8%, XD(unsaturated)=61%

The load flow was performed with base load running (all 4160-600V transformer feeders) all 4kV motors at stop and the largest motor is started.

I will provide the base load kW+jkVAr and the motor HP rating and impedance tomorrow.

The motor starting power factor is 10% with LRA=5.9 x FLA

 

[ThePunisher]

By the way waross,

The 42% VD is calculated using SKM.

The generator is a diesel genset, ABB, machine type: AMG 0900XV08 DSEA. Efficiency at P.F.=0,68 at 100% loading is 96.20%

 

[rbulsara]

I am not sure LF study is the way to find a generator response on motor starting. You need a transient analysis which would model the AVR response.

Also The maximum starting load on gen spec most likely applicable when there is no initial load on the gen. The VD will change if the gen is already loaded.

The other spec means that there will be momentary voltage dip not exceeding 7% when gen is block loaded with 100% rated load, but it would recover in due time provided all setup is correct. Likewise there will be a voltage overshoot of 9% at sudden loss of the 100% load, but it will recover. Voltage dips/overshoots are also accompanied with frequency dip and overshoots as well.

To estimate how fast the generator will recover, ask the mfr to provide the info- the voltage and frequency recovery curve at various load steps. Only way to verify them is to perform tests at various load steps.

ISO has some standards (not that they all are mandatory) for generator prime mover performances.


Rafiq Bulsara

http://www.srengineersct.com

 

[ThePunisher]

Hi Rafiq, I understand your point and thanks for clarifying.

For rafiq and waross,

I was testing voltage drop issues using load flow and static motor starting to see initial problems before going further. The MCC feeds as combination of 4160-600V transformers and 4-4kV motors (2500HP, 800HP, 300HP, 250HP). With no motors being started, the initial loading on the generator is 790kW+j394kVAR (882.88kVA)with 101.33P.U. voltage on the 4.16kV MCC. If the 2500HP motor is started, it causes a 40.49% voltage dip on the MCC, the generator loading is 1172.2kW + J4890.2kVAR (5028.75kVA).

The loading seems ok but the voltage drop isn't. The 2500HP motor has P.F.=90.1, Eff=95.5%. This will provide 1952.87kW + j940.28kVAR (2167.44kVA).

If I use the usual check (see Forum Faq):

Total kVA = Running + (3 x largest motor)
= 882.88kVA + (3 x 2167.44kVA)
= 7385.20kVA

It seems that the genset will not be able to carry impact loading due to 2500HP motor starting with transformer loads running.

Furthermore, the resulting voltage drop assuming that much starting kVA is drawn with other loads running is already out of range for the AVR and excitation to compensate. Any comments and guidance?

 

[ThePunisher]

The generator impedances are:

XD (U) = 61%
XQ (U) = 29%
X1 (U) = 3.7%
X0 (U) = 2.4%
XD' (S) = 8.8%
XD"(S)=4.8%
XQ" (S) = 6.6%
X2 (S) = 5.7%
XP (S) = 6.6%

S = Saturated, U= Unsaturated

The genset input data showed that XD"(S), XD'(S) was used.

The XD used is 2.75P.U. and XD(S)= 0.45P.U. (I am trying to figure out where these values came from since they are not in the datasheet...I am guessing that they were calculated using SKM equations).

 

[waross]

My rule of thumb is proven for much smaller installations, but with that ratio of motor size to generator capacity I would expect the motor to start but case voltage and frequency drops that may be unacceptable to other equipment.
I'm not sure how much voltage drop to expect, but what I do expect is the genset speed will be pulled down somewhat. Then, in addition to the the load voltage drop, the UFRO feature of the AVR will intentionally drop the voltage. (A common UFRO setting is -3Hz. The AVR allows a 3Hz drop in frequency and below that maintains the V/Hz ratio at V/(Hz - 3 For 60 Hz, the V/hz ratio will be V/(60-3)Hz)This makes for an easier start on the motor than DOL on the grid and allows the prime mover to recover speed but, as anticipated, often leads to low voltage issues with other equipment.

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

 

[ThePunisher]

Thanks waross.

I am investigating the generator inputs on the program. That 2.75P.U. used as Xd is high.

On load flow and motor starting, the Xd will be significant in calculating generator terminal voltage. However, is it going to be the saturated value or unsaturated value of reactance?

The SKM static motor starting behaves like load flow except that it takes the motor starting kVA (+ starting P.F.)and calculates the drops.

 

[jomamh]

A synchronous generator is, by its own nature, a rather high impedance device. Without field regulation, the terminal voltage depends strongly on the load, particularly on the reactive component.
In island mode, a quick acting AVR is always necessary in order to mantain the terminal voltage within acceptable limits. Its action takes place during the transient period of the generator, in the first tenths of second after the sub-transient period.
For the exact determination of the voltage vs. time curve in a motor starting process, is not sufficient a load flow computer program. Normally, the gen manufacturer also supplies the AVR, and can give values for the transient behaviour of this ensemble.

 

[ThePunisher]

Thanks jomamh. I contacted ABB and hopefully we will get more details.

I am suspecting though that there might be some errors on the generator inputs especially the direct axis transient and steady state reactances.

When the generator is running and pre-loaded, is it the steady state or transient reactance that will making up the generator internal voltage drop ((IaRa + Ia jXs)?