Paralleling generators of different size

[omegamanltj]

Good Day,
I wish to describe an unusual situation I have come across and see if anyone has any opinions on which is the best way to set up the plant.
I have recently been relieving as a casual on two offshore facilities- floating production , storage and offtake facilities, (FPSO's) The two vessels FPSO’s which are almost sister vessels but not quite. The power distribution configuration is that there are three gas turbine generators on deck which are about 10 MW rating at 6.6 kv and there are three diesel generators below in the “ships engine room” which are about 900 kw at 415 V. These are known as the “essential generators “ , as they can supply the “essential loads” , that is to say everything other than the oil production equipment. This includes the ships services if the vessel has to disconnect from the mooring in case of cyclones, which has happened twice time since I have been here.

As with most ships the star points of the generators are not connected to earth. I believe this is whats called an I-T network where there is no neutral as such, or you could consider it to "float" above earth. This is typical marine practice.
Normally while on the oil field, all consumers, both topsides and essential are supplied by the GTG’s.- see the diagram in the attachment. the breakers circled in blue are normally opened when the vessel is connected to the oil field- i.e the "normal situation". The breakers circled in red are opened in the situation where a GTG and diesel are paralleled prior to load transfer and disconnect from oil field.
The gas turbines are usually powered by the oil field gas but can run on diesel, but they are very thirsty. The HV consumers are virtually all related to the oil production equipment and thus are not needed for sail away. For the sake of this discussion I will call the two vessels A and B.

When we disconnected from the oil field on vessel A, I noticed that it was possible to run for some time , hours in fact, with a diesel generator and a GT paralleled, before we eventually transferred all load to the diesels and sailed off. There was however a switch that you had to activate called “feedback” . I cant recall what that switch actually did though.
However on vessel B it is only possible to parallel them for about 25 seconds while transferring the load . A relay supposedly trips the Diesel gen if they are paralleled longer than that. One of the engineers on vessel B told me this was because of “circulating currents” , but I dont see how that could be because there is no earth connection. The senior electrician told me it was in case of fault current on the GTG you could possibly overload and destroy the diesel alternator as it was rated much smaller than the GTG, and when I told him of the practice on vessel A he seemed to think it was very dangerous.
However I cant see how fault currents could feed back into the diesel as no part of the circuit is earthed.
I noticed by reading some documentation that vessel A was originally fitted with a timer to prevent long time paralleling but it was removed ( no mention why) but it was stated that when running in parallel that the switchboard should not be left unattended.
I suspect the reason is simply because the load sharing is basically manual and there could be a risk that if the diesel load dropped off too much the GTG could start running it like a motor- although there should be a reverse power trip if that were to occur. i.e its more of a mechanical reason than an electrical one.
Does anyone have an opinion about this set up - can you see any potential problems electrically wise , i.e with fault currents or circulating currents when running a 10 MW GTG in parallel with a 0.9 MW diesel generator.

 

[waross]

Think of a 100 KVA co-generator running in parallel with a grid with a few Giga Watts capacity.
Load sharing is a governor issue. Running in "droop" mode makes control simple.
Over loads and over currents are a protection issue.

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

 

[omegamanltj]

Thanks Waross,
That's what I thought. As I said I did find some documentation that said the paralleling timer had been removed. The document also said it had been "tested and proved" to be okay. I assume that means someone qualified checked out the protection and found it was adequate. I am also assuming the timer had been removed because the load sharing and transfer is all done by hand and possibly they found that when people were under time pressure they were more likely to make a mistake than when they could take their time.
Just to confirm,..... in an I-T system, It is not really possible to get circulating currents unless you get two simultaneous earth faults on two different phases and then widely spaced apart physically?

 

[crshears]

Hi Bill,

A friend who did considerable work for Black & Mac in the Eastern Canada offshore oil patch once related a situation where aboard a drilling platform the turbine-driven units and their loads simply refused to play with the Diesel-driven units and their loads; when parallelling them was attempted, one or the other would immediately trip back out. Both systems were reported to be three-phase wye with ungrounded neutral, similar to what was reported in the OP.

My friend believed that the problem was incompatibility between the pulsating nature of the Diesel-generated electricity versus the smoother output of the turbines; I wondered whether there might be too much dissimilarity between the output waveforms causing a resultant "circulating current" perceived as an electrical fault...I never did find out exactly what the issue was.

CR

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

 

[Hoxton]

I wonder what the original design intent was, and was it the same on both vessels?

It is always a concern when you hear "there was a timer.... But it was taken out...(no mention why)"

This is what I call "design drift" - Operations / maintenance have some irritating feature, so they remove it. They think that they know what it's function was, but they do not actually know what the designer wanted to achieve.

The neutral earthing, protection may be a separate issue, but let us consider the governing.

Hypothesis: The three GT units have an automatic load sharing system with "isochronous" speed control giving an essentially constant 50Hz frequency control. This may have the additional complexity of being able to shut down one or more GT at times of low system load. This is a relatively complex system, but is often what is required for a base load isolated system.

The three Smaller diesel sets are designed for essential supplies duty, perhaps for an extended period, but do not require a complex control system, and in fact may benefit from simplicity.

Thus these sets may be fitted with a standard governor system which will be a "speed droop" system, ie as the load increases, the speed drops. This is a stable system since the opposite (load increases, speed rises) would drag in more load and increase the speed until you trip on over speed.

The diesels would then load share, assuming all droops are equal, but start stop etc would probably be manual.

You cannot run the diesels in parallel with the GTs for extended periods without adjusting the diesel governors, because the GTs would always "win" because they are large in comparison to the GTs, and are on fixed speed control.

Ok, so you want to transfer to diesels and shut down the GTs?

Sequence:

Power down HV Topsides

Open HV Topsides breaker

Start diesels and Synchronize to bus

Increase power of diesels until GTs are on zero power.

Open GT breakers

Essential and Emergency consumers now on diesel power

Operate diesels to ensure each has adequate load (bad practice to run diesels on low load)

I guess that your actual sequence is a little different, but the principle will. Be the same.

HTR01 may be normally open, since it may be bad practice to have both the step down transformers in parallel.

The timer may be there to prevent running for extended periods with GTs on low load.

It is a lot simpler to shut the whole lot down and then start the diesels, but not if you have a cyclone coming in!

The earthing and protection issues are items to be considered, but only after the operation transfer is defined.

What do your operation procedures state about the transfer sequence?

Trust that this helps you

 

[omegamanltj]

Thanks, I am not on board at the moment , but form memory the transfer procedures are pretty much as you have described except we reduce load so it can be handled by one GT then synchcronize , then drop the GT load down to about 100kw and then open the breaker on the GT, this is to reduce the risk of it going into revere power.
You are correct that the GTs normally run in Isochronous mode. But I have always wondered why and indeed how it was even possible, given that there are normally two GTS loadsharing.
Having said that the GT load power management system is far more sophisticated than that of the diesels. The former being a computerised system provided by the maker ,Solar turbines and the latter being standard shipyard supply based on mechanical relays.

 

[waross]

Think of your GTs as a grid. The grid has load sharing and stable frequency under isochronous control. A relatively small set is to be paralleled with the grid (GTs). The small set is put in droop control and the speed set point advanced until the desired load is placed on the small set. The small set will now run at a fixed load. In the event that block loading pulls down the grid frequency slightly, the droop control will pick up more load and assist the frequency recovery of the grid.
Diesel generators are normally connected to grids with no issues with "Pulsating" power. A sometime issue is different winding pitches between paralleled generators. This may be addressed by derating the sets slightly.
In the event of a trip within a minute or so of paralleling I would suspect voltage errors. A voltage difference between sets will cause circulating reactive currents which may cause an inverse time overcurrent trip. If a trip occurs in about 25 seconds my first thought would be improper voltage settings and reactive currents. This may be caused by an inaccurate Volt meter in the system.
A second issue may be low load. If one set or system is on isochronous control and the load drops below the load setting of the set on droop control, there will be reverse power and a slight frequency rise. Generally the reverse power will cause a trip. The frqeuency rise will normally be within accepted limits and will not cause an over frequency trip. However to low an over frequency setting may allow an over frequency trip.
Anecdote time;
Some challenging paralleling situations;
A remote sawmill in the third world running on diesel generators. One set had an electronic governor and the other set had a hydraulic governor.
The sets had to be in parallel to start the two largest motors.
When the sets were started in the morning and the motor started, the electronic governor would open faster and hog the load. The generator breaker would trip. This would drop the entire motor starting load on the second set with the hydraulic governor and it would trip out on over current. The solution, run the set with the hydraulic governor for about 1/2 hour so that the oil was up to operating temperature. After about 1/2 hour the hydraulic governor would respond fast enough and track close enough with the electronic governor to avoid over current trips.
A very small utility running a mix of 600 kW and 350 kW sets experienced breaker trips when synchronizing. About once a year the coupling key on one of the 350 kW sets would shear off when paralleling was attempted. If the paralleling attempt was successful there woould be no further trouble. There was an error in the synchro-scope wiring. When the 'scope indicated synchronisim, there was actually a 30 degree phase angle error.
The wiring was corrected and the issues went away.
Given that these two plants ran for years I am skeptical of sets that are claimed to be unable to work in parallel. Apart from low load issues, the problem is generally inappropriate settings, inaccurate meters or improper control wiring.

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