1MW Generator Field Collapse on Load Transfer 2

[rhoudyj]

Yesterday we had a power outage and our emergency diesel generator had to take over for a while. It had a mechanical failure a few hours in and we had to repair it. The generator in question is a bit older, about 25 year old, and is rated for 1MW. It supplies power to some computer HVAC equipment and approx 400kW of UPS load. It has about 650kW of load when all of the systems are operational. So anyway, we ran a 2 hour load bank test at 50% load and another 2 hour test at 100% in addition to a 100% step test with the same 1000kW bank. The generator passed these tests but when we simulated a power failure while connected to the original loads the field collapsed. I'm not quite sure what could be causing this. Do you all think it could perhaps be due to the VSR or an excitation issue? Did not get a chance to fix this before leaving. Still trying to think about what tests I'll want to do on the system later to try to diagnose this problem and any suggestions would be greatly appreciated.

Thanks,
Ronald Houdy, Jr.

 

[waross]

It may be that the loads are causing too much inrush current. That may cause field collapse.
There are several solutions.
1> Permanent Magnet Generator (PMG) to supply power to the Automatic Voltage Regulator (AVG).
2> Use an Uninterruptible Power Supply (UPS) to supply power to the AVR.
2> Old school, a CT measuring the load current and dropping the secondary voltage across a resistor may be used to supply a boost to the AVR under high current conditions and avoid voltage collapse.
The CT boost was often used to maintain the voltage and the current high enough to ensure tripping of overcurrent devices.

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

 

[rhoudyj]

Thanks for the input waross. I definitely have a feeling it's from the sudden current draw on the system. I did not think about supplementing the AVG with a power supply. I am somewhat concerned though that the AVG may be defective as I realized we did not do a reactive load bank test on the system. I read some literature that this cannot properly assess AVG condition. Do you have knowledge on this? I was also curious if you have any additional literature you can suggest.

 

[edison123]

My guess is the UPS harmonics are messing with the old style AVR.

Muthu

http://www.edison.co.in

 

[rhoudyj]

Thanks for the reply edison123. Perhaps? It is a bit odd that this has not been an issue before now if that is the case, though. How would you suggest testing for that?

Ronald Houdy, Jr.

"We can't help everyone, but everyone can help someone."
- Ronald Reagan

 

[waross]

Check your peak inrush current.
I suspect that the charging current of the UPS capacitors.
The longer that your transfer takes, the greater will be the capacitor inrush.
I believe that it is a reasonable assumption that the inrush of the capacitors of 400 kW of UPS may exceed the 100 kW of your block loading test.

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

 

[rhoudyj]

Interesting. I did not even think to check that. Do you think that there is still a possibility if that checks out that the AVG could be found to be defective if a reactive LBT is performed? I'm also wondering if it is worth performing a test on the excitation to make sure there are no faults present there.

Ronald Houdy, Jr.

"We can't help everyone, but everyone can help someone."
- Ronald Reagan

 

[edison123]

Is the generator is working on no-load after that field collapse? Then, I would suspect the AVR. If it is not working, then I would suspect the diodes.

Muthu

http://www.edison.co.in

 

[rhoudyj]

Haven't had the chance to test again since the field collapse, edison123. waross, I forgot to ask but what would you propose as a solution if I can indeed find the capacitor inrush from the UPS to be the problem? Retrofit a new AVG?

Ronald Houdy, Jr.

"We can't help everyone, but everyone can help someone."
- Ronald Reagan

 

[waross]

So anyway, we ran a 2 hour load bank test at 50% load and another 2 hour test at 100% in addition to a 100% step test with the same 1000kW bank.

Based on this I would move the AVR to the bottom of the list of possible causes.
Field collapse/voltage collapse has been with us since the beginning.
When the application of an overload or fault causes the terminal voltage and the AVR supply voltage to drop below the voltage level needed by the AVR to maintain the field we have field/voltage collapse.
With 400 kW of UPS, the inrush of the capacitors may push the generator into field/voltage collapse.
This is the most common and the most likely.
The HVAC load may be adding quite a bit of motor starting surge.
At 25 years old this AVR may not have UFRO (Under Frequency Roll Off). That will also contribute to the problem.
Look for a link for operation at 50 Hz or 60 Hz. If that is not present on the AVR then UFRO is not present. Throw the AVR away and replace it with one that does have UFRO. UFRO is almost universal on AVRs now.
Trouble shooting;
Look for high energization inrush.
Try a start with the HVAC delayed for a few seconds.
Solutions.
1. A few seconds delay on motor starting.
2. A UPS to supply the AVR.
Check that the AVR has UFRO and that it is set to the correct frequency.
Check your speed/frequency.
Frequency should be set at 61.8 Hz for 60 Hz systems and at 51.5 Hz for 50 Hz systems.



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

 

[edison123]

Bill, why would caps charging would cause a voltage / field collapse?

Muthu

http://www.edison.co.in

 

[itsmoked]

On close they look like a dead short. That robs the AVR of all the power it uses to create field.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[waross]

What Keith said, Muthu.
There is also some HVAC load. That may be an inrush of couple of hundred HP of compressor motors.
Block loading of 100% load on an unloaded set may drop the voltage as much as 50% for an instant until the voltage regulator reacts and compensates. The combination of motor starting inrush and capacitor charging inrush may drop the voltage to the AVR enough to collapse the field.
More and more sets are being shipped with PMGs. A PMG equipped set will handle most energization surges.
25 years ago PMGs were not that common.


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

 

[waross]

My guess is the UPS harmonics are messing with the old style AVR.

Muthu has made a good suggestion here.
Rectifier power supplies draw a very peaky current. This can also give rise to harmonic currents.
These current spikes will be reflected as sub cycle voltage dips. Some AVRs don't like the resulting distorted waveforms.
This is a known issue when a large capacity of VFDs are connected to a gen set.
You may have a degraded filter capacitor in the AVR that is causing problems.
It may be well to check the stability setting on the AVR.
At no load, turn the stability setting until the voltage becomes unstable. Then turn the adjustment back until the voltage becomes stable again. Then turn it a little more.
If this solves the problem I would consider it a temporary fix and I would order a new AVR to change at my convenience.
I am still suspecting a surge problem.
Some old AVRs used the same leads for sense and power. You must be able to separate the sense leads or terminals from the power input in order to use a UPS.


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

 

[iop95]

At 400kW, UPS is online model and caps are at battery voltage level so small chance to have caps inrush current, even if UPS is an old version with uncontrolled rectifier.
I think it's an inrush from HVAC and all other load. It's a common mistake is datacenters ... there are not delay in reconecting loads at power failure.
Few seconds between big loads would be enough.
Maybe a new AVR with UFRO may solve this issue but some delay in reconnection it's more sure for genset and entire power grid/system.

 

[crshears]

"Staggered loading" is our terminology for this; one thing I was told was that starting a good slice of the motors first is a means of ensuring there is a nominal mass of rotating equipment to provide some system inertia for the energizing of UPSs etc. Whether or not someone was BS'ing me I'm still not sure...

CR

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

 

[dpc]

It's probably worth replacing a 25-year old voltage regulator regardless of this actual issue. Not that expensive. Buy two. edison123 is correct that harmonics can play havoc with older AVRs.

 

[ScottyUK]

crshears,

There's certainly truth in the rotating load acting as an energy store - if the induction motors become super-synchronous as the generator speed collapses then they will dump energy back into the system. Large block loads with near-unity p.f. can cause the engine to slow sufficiently for this to happen, although the engine then has to meet the block load while providing accelerating torque to the rotating loads so it's by no means easy on the engine. With motor loads it's usually the alternator and AVR which get into trouble trying to meet the large and highly reactive load during DOL starting - when motors start they don't take much active power from the engine.

 

[waross]

Seriously consider staggered starting of everything with a contactor. First the motors and then the other loads, if possible.
Definitely stagger the motors with respect to each other.
The induction motors will contribute energy to the system whenever the generator speed (frequency) drops faster than the motor speed.
When a motor is energized DOL there may be a current transient in the first 1/2ycle. This is related to residual magnetism and the point on wave when the motor is energized.
As the point on wave is a factor, the effect of multiple motors started simultaneously is additive.
This may be significantly greater than the 6X starting surge and may take out instantaneous overcurrent protection.
The advantages of starting the motors first are two fold.
1. Get the motors online while you still have lots of available generator capacity.
2. Induction motor impedance has a high reactive component. As the inductive reactance is inversely proportional to the frequency, the motors will help to dampen the higher order harmonics.

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

 

[crshears]

Thanks, Scotty & Bill; good to know I wasn't being snowed.

CR

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