"Dirty" voltage on bus.

[moonshineJ]

Here is scenario (this conversation resulted from the suspected ?dirty? 450 VAC bus voltage, perhaps affecting some electronics):

#2 generator output wave shape is ?fairly dirty with noticeable noise on the peaks?

#1 generator output wave shape is ?extremely dirty with really bad clipping onpeaks?

The most interesting thing: ?When you run them in parallel, a clean signalis provided, but as soon as you single up on one, it becomes dirty.?

If the opposite were true (good outputs from each individual generator, but ?dirtyvoltage? on bus, when they run in parallel) ? I would give a half dozenpossible reasons. But this one?

I recommended to those folks to ensure that the DC outputs of each voltage regulator look good on an O-scope. I rule out the failure of any one of six diodes of the brushless exciter?s rectifiers, as this would cause BIG problems. I also recommended to check if there are any loads on bus having SCRs, or similar devices, e.g., I saw some bad "dirt" coming from operating on-demand (tankless) waterheaters.

What else possibly could cause this problem?

 

[Besoeker]

Here is scenario (this conversation resulted from the suspected ?dirty? 450 VAC bus voltage, perhaps affecting some electronics):

#2 generator output wave shape is ?fairly dirty with noticeable noise on the peaks?

#1 generator output wave shape is ?extremely dirty with really bad clipping onpeaks?

The most interesting thing: ?When you run them in parallel, a clean signalis provided, but as soon as you single up on one, it becomes dirty.?

If the opposite were true (good outputs from each individual generator, but ?dirtyvoltage? on bus, when they run in parallel) ? I would give a half dozenpossible reasons. But this one?

I recommended to those folks to ensure that the DC outputs of each voltage regulator look good on an O-scope. I rule out the failure of any one of six diodes of the brushless exciter?s rectifiers, as this would cause BIG problems. I also recommended to check if there are any loads on bus having SCRs, or similar devices, e.g., I saw some bad "dirt" coming from operating on-demand (tankless) waterheaters.

What else possibly could cause this problem?

Non linear loads.

 

[Jraef]

Non linear loads.

And by running two generators in parallel, you are increasing the capacity of the system, which decreases the net effects of the same non-linear loads with regard to the distortion they create. Voltage distortion is caused by current distortion, current distortion is always as a percent of the demand on the supply. More supply, same demand, lower current distortion, less voltage distortion.

 

[LMAO]

And by running two generators in parallel, you are increasing the capacity of the system, which decreases the net effects of the same non-linear loads with regard to the distortion they create. Voltage distortion is caused by current distortion, current distortion is always as a percent of the demand on the supply. More supply, same demand, lower current distortion, less voltage distortion.

Correct. Other way to explain: Two gens in parallel have half the impedance and thus less voltage distortion for the same current.

What are the loads? Different non-linear loads have different types of distortions. With diode you get a lot of harmonics at 5th, 7th, 11th, 13th, etc.. With IGBTs (active front end) you get high frequency distortion at carrier frequency, SCR are also different.

Depending on distortion source, different solutions are available.

 

[Besoeker]

Correct. Other way to explain: Two gens in parallel have half the impedance and thus less voltage distortion for the same current.

What are the loads? Different non-linear loads have different types of distortions. With diode you get a lot of harmonics at 5th, 7th, 11th, 13th, etc.. With IGBTs (active front end) you get high frequency distortion at carrier frequency, SCR are also different.

Depending on distortion source, different solutions are available.

Yes, the lower the supply impedance the lower the distortion. If the supply was infinitely stiff there would be no distortion. But real life kicks in.

I'll add something to your comments about harmonics though.
Your 5th, 7th, 11th, 13th, etc. relates to three phase six pulse equipment; 6n?1. In other words, industrial.

I've been dealing with power electronics, mainly industrial variable speed drives for a long time. In the early days of my career these were SCR DC drives for the most part and we were aware of the harmonic content and the effects long before it was more widely considered to be an issue. However, customers became more savvy about the potentially harmful effects. Contracts started to require a before and after survey.

One thing that surprised me was the 3rd harmonic (and multiples) content. I saw this on the 11kV supply. I pondered that. They emanate from single phase non-linear loads. The thing is that there are LOTS of them.
The is the voltage at our house.

You can see the flattened tops by eye. That is caused by 3rd harmonic.

 

[moonshineJ]

Yes, the lower the supply impedance the lower the distortion. If the supply was infinitely stiff there would be no distortion. But real life kicks in.

I'll add something to your comments about harmonics though.
Your 5th, 7th, 11th, 13th, etc. relates to three phase six pulse equipment; 6n?1. In other words, industrial.

I've been dealing with power electronics, mainly industrial variable speed drives for a long time. In the early days of my career these were SCR DC drives for the most part and we were aware of the harmonic content and the effects long before it was more widely considered to be an issue. However, customers became more savvy about the potentially harmful effects. Contracts started to require a before and after survey.

One thing that surprised me was the 3rd harmonic (and multiples) content. I saw this on the 11kV supply. I pondered that. They emanate from single phase non-linear loads. The thing is that there are LOTS of them.
The is the voltage at our house.

You can see the flattened tops by eye. That is caused by 3rd harmonic.

Thank you for good explanation. I recommended to them to strip the swbd of any "questionable" loads and see what happens with jam, as well as the flattened peaks.

 

[moonshineJ]

Correct. Other way to explain: Two gens in parallel have half the impedance and thus less voltage distortion for the same current.

What are the loads? Different non-linear loads have different types of distortions. With diode you get a lot of harmonics at 5th, 7th, 11th, 13th, etc.. With IGBTs (active front end) you get high frequency distortion at carrier frequency, SCR are also different.

Depending on distortion source, different solutions are available.

This sounds logical. They still have to figure out what causes it, but the effect of running two gens in parallel to put down distortions makes sense. I forgot to add that when ship is on shore power, [apparently] they have no problems..the shore power would have ideally low impedance, I guess.

 

[GoldDigger]

Thank you for good explanation. I recommended to them to strip the swbd of any "questionable" loads and see what happens with jam, as well as the flattened peaks.

Since the voltage distortion is the result of an even more distorted load current waveform, a scope with a clamp-on current probe would be a straightforward way to identify the problem loads.

 

[LMAO]

Since the voltage distortion is the result of an even more distorted load current waveform, a scope with a clamp-on current probe would be a straightforward way to identify the problem loads.

Yep, I bet current looks like hell. Voltage distortions are difficult to see but current distortion will be obvious.

 

[moonshineJ]

Thank you!

 

[Besoeker]

Since the voltage distortion is the result of an even more distorted load current waveform, a scope with a clamp-on current probe would be a straightforward way to identify the problem loads.

This current waveform was also taken at home:

I don't imagine ours unique. Multiply that by the number of residences fed from the grid and you can get an idea of the scope of the problem.

 

[kwired]

This sounds logical. They still have to figure out what causes it, but the effect of running two gens in parallel to put down distortions makes sense. I forgot to add that when ship is on shore power, [apparently] they have no problems..the shore power would have ideally low impedance, I guess.

Shore power is typically a more stout supply, you have kVA limitations of the supply transformer, but it is still fed by a larger source then your local generator and will pass on overloading for short term easier then a local smaller supply can create extra output from overloading. You start a heavy load with local supply and voltage and frequency likely are effected because the supply has lower limitations for handling the sudden change, but the utility supply is so much larger that same load may see some voltage drop as a result of local impedance - but frequency is unchanged.

 

[moonshineJ]

This current waveform was also taken at home:

I don't imagine ours unique. Multiply that by the number of residences fed from the grid and you can get an idea of the scope of the problem.

While there is some high math involved to explain this, intuitively it resembles the waterline with the faucet being open/closed followed by hydraulic knocks, and all kind of sounds, depending on how wide the faucet is open. For electrical, the more stable is power supply the less it changes under different loads (current). A fresh battery will keep for some time its voltage, even under high load, while voltage on depleted battery drops a lot.
I'd assume that those irregularities come due to reactive components being a part of the load? I never thought about it this way.

 

[GoldDigger]

...
I'd assume that those irregularities come due to reactive components being a part of the load? I never thought about it this way.

A waveform like the one shown requires more than just reactive components. It also depends on non-linear loads.
Reactive components alone can cause time displacement, but if there are no higher harmonics in the input voltage waveform it they cannot produce them.

If the harmonic content of the input waveform is significant, a reactive load can change the relative amplitude of the different harmonics in the load current and produce a more interesting waveform.

 

[LMAO]

A waveform like the one shown requires more than just reactive components. It also depends on non-linear loads.
Reactive components alone can cause time displacement, but if there are no higher harmonics in the input voltage waveform it they cannot produce them.

If the harmonic content of the input waveform is significant, a reactive load can change the relative amplitude of the different harmonics in the load current and produce a more interesting waveform.

Yes. The distortions you see is a result of non-linear load, a rectifier bridge to be more specific. These are 5th and 7th harmonics.

To get really technical, I always hated calling semiconductor devices (like diodes and transistors) "non-linear" loads. A reactor can also be non-linear (swing choke) having variable inductance based on current.

They should call semiconductor devices "non-differentiable response devices" because during their change of state (open/close) derivative of output signal does not exist.

 

[Besoeker]

Yes. The distortions you see is a result of non-linear load, a rectifier bridge to be more specific. These are 5th and 7th harmonics.

Third actually and some resonance.
It's a single phase system.
I do a lot of calculations for my colleagues in the lighting division. Single phase dimmers. Third harmonics dominate.

 

[LMAO]

Third actually and some resonance.
It's a single phase system.
I do a lot of calculations for my colleagues in the lighting division. Single phase dimmers. Third harmonics dominate.

That's residential power? That's a lot of distortion for single phase residential. Doesn't UK follow IEEE 519?

 

[Besoeker]

That's residential power? That's a lot of distortion for single phase residential. Doesn't UK follow IEEE 519?

Bear in mind that's current, not voltage.

 

[moonshineJ]

A waveform like the one shown requires more than just reactive components. It also depends on non-linear loads.
Reactive components alone can cause time displacement, but if there are no higher harmonics in the input voltage waveform it they cannot produce them.

If the harmonic content of the input waveform is significant, a reactive load can change the relative amplitude of the different harmonics in the load current and produce a more interesting waveform.

Good point.

 

[LMAO]

Bear in mind that's current, not voltage.

I know but I imagine that much harmonics result in significant voltage distortion as well (depending on source impedance). Also, IEEE 519 has guidelines for current as well as voltage.