Power fluctuations when two generators are in parallel

[orense]

Hi,

This is a bit off topic with respect to PLCs, but since there are many experienced engineers on here, I'll give it a try.

We are commissioning a diesel electric platform supply vessel, and we are doing load testing on the main switchboard.

When we run one generator set solo, the load from the loadbank is very stable.
The output is very stable.

When we run two in parallel, the load is not stable on the generators, even if the loadbank has stable load.
We have been doing some measurements, and it seems that there are power fluctuations between the generators in parallel. It is not due to any engine governor issues, because the fluctuations are too fast, more than 5 hz.
We have used several measuring devices, so we know that the fluctuations are real. They are eapprox 5-6 % of the rated kw of the generator sets.

The engines are 1800 rpm, and they are approx. 1500 kw each.

It seems that there is some kind of natural phenomenon where power is fluctuating between the generators. They are considering if flexible coupling between engine and generator might have wrong specification.

Any ideas anybody ?

 

[Christoph]

Are these generators synced with respect to the 60Hz cycle??~~~~

 

[orense]

Hi,

Thanks, yes they are.
The vessel has 60hz, and that is the frequency of the generator output.

The engines are running with 3% speed droop.

 

[DBLD99]

What is the control scheme you are using when running in parallel. It sounds like you are using both engines at 3% droop. One engine needs to be an ISO engine, which in general responds to the upset quicker than the 2nd engine. If both governors are set for the same control you will never achieve perfect control of the freq.

 

[substrate]

You say it's not the governor, but I had an issue with load balancing some diesel generators because the engine computer did not have a fine enough output resolution to provide even load balancing ... the governor output was just a little too course, so the engines could not load share properly. We were seeing the same thing as you were, but at a lower frequency I believe. 5 Hz does seem quite high for it to be a governor issue, I agree. Anyway, just thought I'd give you my experience.

The gen sets were Electromotive diesels (EMD) with Detroit Diesel engine computers.

 

[orense]

Hi,

Under normal cirumstances the engines are controlled by the PMS. The PMS control the rpm and thereby the load by using pulses converted into voltage level. This is called active droop or compensated droop.

However, it does not make any difference, the issue is still there even of the active droop is enabled or not. If it turned off, so the PMS can not control the loadsharing they still have the same power fluctuation problem.

 

[orense]

Substrate, my previous post was for DBLD99's post.
I probably wrote it while you posted your comments.

Thanks for the comments.

I don't think it is the engine controls. The engine manufacturer has tried to tune their control system and it did not improve the oscillations.

The oscillations are fast, so we do not think it is related to engine control.
The engine rpm is very stable, but the power output is oscillating at high frequency

 

[V0N_hydro]

How can you be absolutely positive it is not related to the governors? Can you plot the throttle position vs. power oscillation and see the that throttle is totally constant? Can you lock the throttle of the machines to constant position to be sure they are receiving constant fuel supply?

What voltage harmonics are observed?

Are the generators solidly grounded or high resistance ground?

are the two generators identical make and model and manufacture? If they have different pitch windings the generated voltage waveform will have slightly different distortions and that difference would result in current flow.

 

[shooter]

If the settings are all the same the PID of the attenuator is reacting too fast. Set one loop on K and set the Integrator on zero for one set, so this will run on constant speed, one set should be set as master and the other as slave. If both are master they are trying to control the speed both, and as one is trying too slow down, the other is faster. Then the goal is reached but the I factor is still in the game.