We have installed at our facility a 13.8 kV, 16.8 MW synchronous turbine-driven generator. We are currently experiencing an issue when placing the generator in service which I will attempt to summarize as best as possible.
Initially, the total power consumption by the facility is approximately 10 MW, which is supplied by the utility, at 13.8 kV with a power factor of 0.83 (or thereabout). When placing the generator in service, the excitation is turned on, after which the generator's supply is automatically synchronized to the utility's supply. Once the three breakers (generator, utility and tie) are closed, the generator is then loaded in Droop.
A power factor switch is turned on in order to toggle the excitation's control system (DECS 200) mode from AVR to PF control. Once this is completed, despite the power factor setpoint of the DECS being set at 0.8, the actual power factor drops immediately to approximately 0.2 - 0.3. The amperage drawn by the generator increases instantaneously from 250A to 800A. As the generator is loaded to match the demand from the users on its own loads, the power factor indicated for the utility drops from 0.83 to 0.4 or thereabout, with an additional current draw as well.
Once the generator is loaded to the required 5.5 MW, the tie breaker is opened and the power factor switch immediately turned off to return the mode of the DECS to AVR (with a pre-position setpoint of 13.8 kV). At this instant, there is a voltage spike from the generator, sometimes reaching between 14.8 - 15 kV (max range of the measurement device), which oftentimes causes upsets to lower-voltage equipment. The system then normalizes to the required 13.8 kV and 0.8 power factor while being operated in isochronous control. Based on this response, it seems that there is significant additional excitation which is immediately dissipated when the control is switched from power factor to AVR control.
Thus far, we have listed the DECS as the most likely cause of the issue, specifically when it is operated in power factor/droop control mode. The generator appears to be performing as required based on the output from the DECS as it functions without any issues in AVR/isochronous mode. Checks on the DECS were made to verify that the contacts were switching correctly based on the input signals received.
Although I do not believe that the generator's performance is to be questioned, it must be noted that this action was only observed after conducting some repairs to the unit approximately 1 year ago.
We do understand that operating the generator under these conditions means that we will be generating power inefficiently, as well as placing strain on the generator's windings based on the increased current draw/heating of the windings.
Based on the scenario provided, I would like to request the assistance of anyone who can provide some further insight/guidance into this issue, as it pertains to the reason for the behaviour of the generator in this condition, as well as any possible corrective measures that can be implemented (beside replacing the DECS) in order to resolve the issue or other checks that can be done on the system to gain more insight into the issue.
I do look forward to any feedback and questions.
Thanks and Best Regards.
You have most likely answered your own question:
>...., it must be noted that this action was
>only observed after conducting some repairs to the unit
>approximately 1 year ago.
What, exactly was the work done approximately 1 year ago? Why were the repairs necessary?
It would appear that everything was working correctly PRIOR to the repair work, so if you identify what was done then you can redo the repair works to correct the issue(s).
Based on the information provided, It would seem something is amiss with either the PT inputs and/or the CT inputs, and it's even possible the DECS 200 configuration has been improperly changed.
Again, if it worked properly before the repair works, something done during the repair works would seem to have been the problem.
Please write back to let us know what you discover!
As a Generator TFA and based on my experience, I would look at the current transformers, in particular the polarity.
When you are in automatic mode the regulator is only reading the PTs and hence do not care about the CTs. In that configuration, it seems that you are not experiencing any problem. But switching in PF mode will cause the AVR to read the CTs and the VTs to determine what is the actual reactive power and try to match the current value with the set point. If the AVR is reading a false reactive power, the regulator may act in te opposite way as it should. Send negative excitation order to increase the MVArs etc.
What you can do is to cross the MVAr and the MW values between the AVR and the protection relays for example. Obvioulsy you need to be sure of the relay you are reading.
Second, when you operate the generator in automatic mode, try to send +Exc order from the local display and check that MVAr are increasing.
The last thing I can think of, is how the AVR is expecting to see the phase sequence. Some companies make mistakes like setting up the AVR settings in ABC while the real sequence is ACB.
You may also look at the PTs, because if two phases are inverted, the same problem as described will happen.
Many thanks for your inputs CSA and Saul!
In October 2015, we experienced a mechanical failure of the PMG, which is coupled on the rotor shaft of the generator. One of the magnets was released onto the neighbouring exciter windings. As such, the PMG was completely removed from the shaft (as there is an alternative UPS to the DECS) and the exciter re-wound. With no physical damage to the stator windings, they were only varnished and baked.
It is indeed correct to say that everything worked with no issues prior to the repairs. Prior to the re-installation of the generator, the settings on the DECS were compared to the "As Left" settings from the commissioning of the unit, with no differences noted. All electrical connections were reconnected using photos taken, together with the wiring drawings in order to double check ourselves that no wiring mix ups were implemented.
The two avenues which are being considered are the functionality of the DECS when in PF mode and the phase differences between the two power sources (generator and utility). We will definitely check also the CTs as they play a vital role in the calculation of the PF when in PF mode.
Apart from verifying the CTs as an additional check, we did not have to remove the measuring CTs or disconnect wiring for the CTs which are used as an input to the DECS for PF. The setpoint of the DECS remains at 0.8 while the actual PF drops to the 0.2 - 0.3 range. One of the checks I will complete is to compare the DECS kVARs value to that of the generator's breaker independent metering system which also provides kVARs reading. This can confirm if the value being fed to the DECS during PF mode is accurate as this will not be a intermittent faulty reading. Can you please elaborate on "cross the MVAr and the MW values between the AVR and the protection relays for example?"
When the generator is operated in automatic voltage regulation (AVR), adjustments to the setpoint can be made manually using a Raise/Lower switch. We have adjusted this previously and the voltage and field current is increased. When in PF control, adjustments to the Raise/Lower switch have no effect on the actual PF, as the setpoint remains at 0.8 even as the PF decreases. However, with these adjustments, I did not observe if the MVARs was being adjusted as well. I will confirm the phase sequence as well. I will add the PTs to the list of things to check at the next opportunity.
Many thanks again for your inputs!
Whether you are in automatic or PF mode the DECS do exactly the same thing. It regulates the excitation field current flowing through the rotor.
The difference is the nature of the setpoint. In one case it regulates the excitation field current to match the voltage setpoint and in the other case it regulates the excitation field current to match the MVAr setpoint.
Let's say that the DECS is reading the right stator voltage and the right stator current magnitude, but not the right angle between them. Then you might have a flawed active and/or reactive power.
If the DECS is reading a false reactive power, for example if it reads a negative reactive power instead of positive one then it will never be able to reach the setpoint. By trying to reduce the gap it will actually increase it to the maximum untill it reach the limitations.
That is why you need to be 100% sure of the MW and MVAr read by the DECS.
Did you replace the PMG? Did you do a new magnetization of the magnets?
Thanks for the elaboration Saul!
Agreed on what the DECS does for the regulation of field current based on the mode and subsequent setpoint. I now understand what you meant by cross between the MW and MVArs. Since we made no changes to the DECS (wiring or software configuration), I believe this should be accurate. But it would definitely not hurt to double check.
We did not replace the PMG but removed it completely from the loop. Instead, we utilise a UPS to the DECS. This was actually done a few months prior to the mechanical failure of the PMG.
One new part of information we obtained is that the supply by the utility was switched from one distribution center to another at the end of 2015. With our generator out of service, it may also be worth checking the phases of this supply, just to ensure everything is as before. I will definitely schedule to check on the MW/MVAR readings obtained.
Thanks for the input Saul!!
You are welcome.
You should also check that the power of the UPS supply is strongh enough to support the excitation needs. Specially is the source has been swapped.
It sure seems, from the information provided, that when the DECS is transferred to PF control that it is not seeing the reactive current change--the signal from the CTs, usually. AND, we know the PF is going very low--but we DON'T KNOW if it's going in the LEADING or LAGGING direction. (I presume it's going in the LAGGING direction, because when the unit goes back to AVR mode the voltage is very high.)
I'm not extremely familiar with the Basler products, but one would think there would be some kind of "bumpless transfer" or smoothing of the jump in reference when transferring between PF (or VAr) control and voltage control ("AVR mode").
To my way of thinking--the DECS should not be transferred out of AVR mode to PF mode during grid-tie operation. It should be left in AVR mode, and the operators should be manually controlling the PF by adjusting the AVR setpoint both during grid-tie and isochronous operation. At least until the cause for the issue is understood and resolved. If transferring to PF mode is causing upsets and potential generator winding harm--and problems with other components when transferring back to AVR mode when switching from grid-tie to isochronous operation--then simply DON'T use PF mode until the cause is understood and resolved. That avoids the inefficiency issue, the potential generator winding damage, AND the component damage. Just don't use PF mode until the cause of the problem is resolved and understood.
Hope this helps!
Agree with CSA.
You should totally avoid operating the generator in PF mode until this problem is fixed.
Actually, I never seen any generator operating on the grid with a power factor of 0,2 or 0,3.
Did you notice any alarms/limitations while in PF mode?
Some good news is that we got the opportunity to load the generator, and it was not switched to power factor mode. This worked really good as there was no significant change as when it was switched. Now we will proceed to center our checks around the phase measurements for the generator as this may be the most likely cause of the problem. Just btw there were no alarms noted in previous synchronization attempts. Many thanks for your contributions.