Emergency Generator

[NickParker]

As far as I know, Emergency generator operating as a standalone unit, would be operated in Isochronous mode always.
Recently a chinese vendor came up with a proposal that it always it is operated in droop mode. Is it the case? because in emergency mode, when some motors are started, then speed will not automatically be corrected if operated in droop mode. but the vendor is saying the emergency generator operates in droop mode by default.

 

[goutam_freelance]

The fuel supply to the diesel engine needs to be adjusted based on load. For example, if a large motor is started, there will be a fall in speed of the EDG set, and the governor will increase the fuel flow as a function of the speed drop to bring the EDG to rated speed.

For this, the characteristic of the governor needs to have a droop, i.e., if the speed is less, the fuel flow will increase almost proportional to the drop in speed.

 

[waross]

Droop mode is the default for standby.
The no-load frequency is either 61.8 Hz or 51.5 Hz, dropping to 60 Hz or 50 Hz respectively at full load.
You must be looking at a set rated above 100 KVA.
I have installed dozens of standby sets and I cannot remember a set under 100 KVA that had a governor option to run in isochronous mode.
The speed of an induction motor varies about 2.8% from no load to full load.
I have never had a customer even notice the slight additional speed variation running on a generator in droop mode.
Droop mode is the Proportional of PID control. P--
Isochronous is Integral of PID control. PI-
And the first response of the governor is always a proportional response.
When the governor has responded in droop mode to a change in load, the Integral or Isochronous feature detects the error between the set point and the actual speed and slowly corrects back to the set point.
Compare the action:
Isochronous:, A motor is running at rated frequency/speed. The load on the set is increased and the motor frequency/speed drops below the rated frequency until the integral feature makes a correction.
Droop: A motor is running at slightly above rated frequency/speed. The load on the set is increased and the motor frequency/speed drops, but remains above rated frequency/speed.

Note #1. Both modes initially respond in droop mode and the frequency will typically drop below nominal rated frequency as the droop responds.
Note #2. When the application of a relatively large load pulls the frequency down below the nominal frequency by more than 3 Hz, ie: below 47 Hz or 57 Hz, the UFRO feature of the AVR becomes active to avoid the motors becoming magnetically saturated.
This is a feature of the AVR and applies for both droop and isochronous.
In control jargon, 3% droop would be described as 3% Proportional plus 3% offset.

 

[waross]

Typical droop response. Integral will correct the error back to the set-point.
Where to use integral? Typically on systems with small expected load changes.
On a utility grid, load changes tend to be very small in relation to the capacity of the grid.
One set, the swing set, of all the sets in a grid will be in Isochronous. All other sets will be in droop mode.
Often the effect of increasing or decreasing load will be seen as single cycles, gained or lost.
As the swing set approaches full load or no load, the load dispatch center will call up additional capacity or drop capacity off-line.
This is a historical view, from a time when most electric clocks depended on small synchronous motors and a very accurate grid frequency was a must.
There may be changes in present day operating procedures, but this is a description of Droop versus Isochronous.
Historical grid control is for illustration only.

 

[NickParker]

Droop mode is the default for standby.
The no-load frequency is either 61.8 Hz or 51.5 Hz, dropping to 60 Hz or 50 Hz respectively at full load.
You must be looking at a set rated above 100 KVA.
I have installed dozens of standby sets and I cannot remember a set under 100 KVA that had a governor option to run in isochronous mode.
The speed of an induction motor varies about 2.8% from no load to full load.
I have never had a customer even notice the slight additional speed variation running on a generator in droop mode.
Droop mode is the Proportional of PID control. P--
Isochronous is Integral of PID control. PI-
And the first response of the governor is always a proportional response.
When the governor has responded in droop mode to a change in load, the Integral or Isochronous feature detects the error between the set point and the actual speed and slowly corrects back to the set point.
Compare the action:
Isochronous:, A motor is running at rated frequency/speed. The load on the set is increased and the motor frequency/speed drops below the rated frequency until the integral feature makes a correction.
Droop: A motor is running at slightly above rated frequency/speed. The load on the set is increased and the motor frequency/speed drops, but remains above rated frequency/speed.

Note #1. Both modes initially respond in droop mode and the frequency will typically drop below nominal rated frequency as the droop responds.
Note #2. When the application of a relatively large load pulls the frequency down below the nominal frequency by more than 3 Hz, ie: below 47 Hz or 57 Hz, the UFRO feature of the AVR becomes active to avoid the motors becoming magnetically saturated.
This is a feature of the AVR and applies for both droop and isochronous.
In control jargon, 3% droop would be described as 3% Proportional plus 3% offset.

The diesel generator capacity is around 1250kW, but why would emergency generator is operated in droop mode as it is not to be tripped in some cases say, in sailing condition on a ship.

 

[waross]

Just run it in 3% droop mode like everyone else, Nick. We ran a micro utility with about 5000 clients in droop and no-one ever knew.
We originally had a mix of 600 KVA and 350 KVA machines and then replaced a couple of 350 KVAs with two 1600 KVAs.
We considered using isochronous but in the end stayed in droop mode.
The 1600s were second hand machines originally used as standbys at an American Embassy where they ran in droop mode.
We used them as is with no changes to operating mode.
Set them up for a no-load frequency of 61.8 Hz or 51.5 Hz and 3% droop.
That is probably the settings that will be as delivered.

 

[crshears]

I figured Bill would describe it better than I ever could, and I was right.

The only codicil I might add would be to suggest having a zero load speed of 50.75 / 60.9 Hz on all units, such that the frequency will be at nominal when system load is 50% with all units loaded at 50%. That way, frequency will average out to nominal, rather than consistently being above nominal.

I say this as a person who still uses synchronous clocks .

 

[waross]

That is a good idea as far as it goes, CR, but it doesn't go far enough.
When you consider other factors such as induction motor V/Hz characteristics, interaction with the AVR, specifically UFRO, block loading and possibly the settings of the automatic transfer panel. you get back to the industry standard of 61.8 or 51.5 Hz.
As a technically aware individual owner operator of an "Own Use" set, there is nothing wrong with your settings, most of the time. You may never have a problem, but you may encounter a glitch that may be cured by going back to the default settings.
As a technically aware individual owner operator you will notice and correct any such issues.
For the majority of installs, best to use the default and never worry.
And by the way, I was called out to look at a set that was not performing well and found that an installer "Who knew better" had reset the no-load frequency to exactly 60 Hz.. (I forget the details, it was a long time ago, but the customer was having issues)

 

[waross]

Good point about the clocks CR. If I could not convince the customer to live with it, I would help him find new clocks.
And those old clocks. I remember checking and setting the frequency of an old set out in the jungle of the Moskito Coast with a synchronous clock and a wrist watch.

 

[Amp Dylan]

I’ve seen similar setups on marine vessels. They’re often in droop because they don’t parallel with shore/grid, so simpler control and less risk of governor hunting. Isoch is nice, but droop is “good enough” for shipboard stuff if the load steps aren’t insane. I’d just confirm the no-load freq is set around 61.8 Hz or 51.5 Hz like others said.

If you’re worried about motor starts causing undervoltage trips, maybe check AVR response and whether UFRO is biting too early. Could tweak the pickup or add a delay.

What kind of loads are you starting that make you question the droop config?

 

[waross]

They’re often in droop because they don’t parallel with shore/grid,

Parallel with the grid must ALWAYS be in DROOP. Albeit at 5% rather than 3%.

What kind of loads are you starting that make you question the droop config?

The sub-transient responses of droop and isoch are virtually identical when experiencing block loading. The sub-transient response describes the greatest frequency and voltage drop resulting from block loading, that is applying large loads.
Have Caterpillar, Onan and F.G. Wilson been wrong for all these decades?
Use droop and sleep well at night.

 

[Amp Dylan]

Parallel with the grid must ALWAYS be in DROOP. Albeit at 5% rather than 3%.

The sub-transient responses of droop and isoch are virtually identical when experiencing block loading. The sub-transient response describes the greatest frequency and voltage drop resulting from block loading, that is applying large loads.
Have Caterpillar, Onan and F.G. Wilson been wrong for all these decades?
Use droop and sleep well at night.

Yeah makes sense. Honestly I’m not anti-droop at all, I just had a couple cases where guys assumed isoch was needed for motor starting, but turns out AVR + UFRO tuning fixed it.
I agree, block load hit looks pretty much the same either way for a standalone gen. And yeah, Caterpillar and Wilson didn’t get where they are by guessing.
Appreciate the insight.

 

[waross]

If the UFRO jumper is set for 50 Hz on a 60 Hz machine you will definitely have issues with motor starting.

 

[crshears]

Amp Dylan: They’re often in droop because they don’t parallel with shore/grid

Most AsHJ don't want to see parallels with shore at all; they more or less ubiquitously stipulate break before make, in both directions. That being said, though, with a demonstrable need for make before break and appropriate measures and precautions, parallels with shore may be permitted, but only for brief periods.

Grid-scale private generators play by a different set of rules.

 

[TugboatEng]

Between class certification and emissions exemptions, you're going to have very limited choices of emergency generator sets.