Why is there no existing standard for backup generator sizing?

[chao_david]

I'm quite shocked to find out that there are no standards that show the formula for generator sizing. Even in our local electrical code. I'm not sure why.

I'm trying to set a standard for our company in sizing genset. Based on several white papers I've read and generator rentals websites, I've come up with this:

Generator kW = (Total running kW of all loads) * Demand Factor - Running kW of largest motor + Starting kW of largest motor

This is loosely based on Siemens Generator Sizing guide except I've included multiplying demand factor.

In estimating the starting kW of motors, is it valid to multiply the motor kW by 600% for direct online connection? Eaton did this in their video (except unit in kVA). However, some articles say that the 600% is for starting current and not kW. I kind of agree since we will multiply the starting current with the starting voltage which is expected to also drop to 70% voltage during starting. In terms of kW, this will result in around 4x the normal kW.

In summary:

1. Is my formula good enough?
2. For direct online, do you agree that the starting kW is 600% or is it 4x the running kW?

 

[waross]

Caterpillar and Onan have software that will size a generator based on voltage drop when the largest motor starts.
I have had excellent results sizing the largest motor at 3 times the KVA.
I have over a dozen residential installs with lots of A/C that all perform well.
Some customers opt to forgo the use of their largest A/C and use a smaller generator.
In such cases I have been able to determine that a factor of 2.5 times the motor KVA will start the motor but the voltage drop is unacceptable.
And by the way, I have over a dozen acceptable installs that the Caterpillar software default settings deems unacceptable.
But Cat is in the business of selling gen-sets, the bigger the better.

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

Starting KVA.
Motor starting is highly reactive.
A KW load tends to slow or stall the prime mover.
300% is adequate to support the KW demands.
KVA contributes to current and as a result, the voltage drop.
As long as the set is running at speed, the AVR compensates for voltage drop.
Permanent Magnet Generator excitation helps a lot. (PMG)
That said, the starting KW often slows the prime mver until the governor resonds to the extra load.
Here Under Frequency Roll Off is your Friend, (UFRO)
UFRO allows a few percent drop in frequency and then reduces the voltage proportionately to further drops in frequency.
The drop in voltage maintains the Volts per Hertz ratio to the motor and avoids magnetic saturation of the motor windings.
The drop in voltage also reduces the KW load on the generator.
UFRO greatly aids in motor starting and voltage and frequency recovery.

Beware of dependency on demand factors based on actual loads.
Residential service sizing is grossly oversized even after the application of demand factors.
Using demand factors based on code may result in an oversized gen-set, even with demand factors.
Applying demand factors to actual loads may result in an undersized gen-set.
Why?
Breakers are not typically loaded beyond 80% of their rating.
Transformer will accept short term overloads and the damage due to long term overloads may take years to manifest.
A typical standby generator has no such safety margin.
If a grid supplied installation may at times exceed the calculated demand load, it may not be cause for concern and it may go un-noticed.
Not so a standby generator.
If the generator capability is matched to demand calculations, then when (when not if, it will happen) you will experience low voltage and low frequency.

Bottom line.
300% largest motor KVA and very cautious with demand factor.
Or use Cat software and buy a larger set.
The extra cost is defensible in our Cover Your ASSets World.
The great majority of customers will be happy with my sizingand the cost.
The odd very demanding customer will need the larger set sized as per Cat.
I hope that this helps.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[waross]

This is for standby sizing.
Prime rated sets allow a 10% overload for a designated time.
You will have more wiggle room with a prime rated set.
Example
Often a 100 KVA prime set may be rated as a 110 KVA standby set.
The prime set may have accessories such as an oil cooler and a larger oil sump, or it may be the same set with a different nameplate.
Another advantage of the higher capability of a prime set is more years of service before an overhaul.
My comments and advice are based on standby sets.
Are you concerned with standby sets or with prime power sets.
I have seen engineers oversize the prime mover on a prime power set by another 25% for a government installation where they knew that an allocation for engine overhaul would be almost impossible to obtain when the engine started to lose power.
That extra 25% was intended to gain extra years before the engine could no longer support the load.
That was standby rating plus 10% for the generator end and
Standby rating plus 10%, plus 25% for the diesel engine.

And sets are rated in KVA, not KW.
KW is typically 80% of KVA for a power factor rating of 0.8
There are very few exceptions, mostly for small hobby sets.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[cranky108]

Why do you need a standard? A good engineer should be able to figure this out without a standard. That's why we went to school for all those years.

 

[Gr8blu]

Part of the reason there is no standard is that the TYPE of loads (resistive and/or reactive) can have an impact on the required generator output. Something that is purely resistive (which is really hard to find these days) needs KW - something that is primarily reactive needs more KVA (because power factor can be all over the map). Both of these are different from the demands of an intermittent load.

Modern reactive loads include: fluorescent and LED lighting, rotating machines, power electronics, smaller "personal" electronics, UPS systems, transformers, and power factor correction banks.

Intermittent load examples include AC and/or refrigeration.

Converting energy to motion for more than half a century

 

[PowerQualityDoctor]

You should also consider the level of harmonics, as this affects the generator loading.

See for example:

https://www.mirusinternational.com/downloads/Mirus case study - Plains All-American Pipeline.pdf

 

[catserveng]

You're asking for a "standard" for sizing gensets, but since most modern systems have varying types of loads, with a wide variation in tolerance for voltage and frequency drop, there really is a lot of legwork involved if you want to do it right.

The attached paper by Cummins offers a lot of things you should look at, I have used it many times in training technicians and sales engineers. Most of the info you've been provided so far by others gives you a lot of info, but since you have provided no details about the types of power systems and their loads you are trying to support, you may not be getting the answers you're looking for. More info from you will likely get you better answers to meet your needs.

Hope that helps, MikeL

 

[RRaghunath]

I have seen DG set with 2.5MW rated Diesel Engine coupled to a 3.5MVA Generator. This was a requirement as the Generator is expected to start a 1000kW motor DOL.
So, it pays to be cautious not to oversize the Diesel engine unnecessarily.

R Raghunath

 

[waross]

An Existing standard is The Orange Book published by the IEEE
My copy is copyrighted in 1980.

​

Orange Book

​

IEEE Recommended Practice For

​

Emergency and Standby Power

​

Systems for Industrial

​

and Commercial Applications
I was going to scan the Table of Contents until I saw that it ran to 10 pages.
The book is 208 pages.
That is the good news.
The bad news is that you may find it a little light on actual sizing formulae.


--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[Gr8blu]

Just a note about the IEEE "Color Books" in general. These IEEE standards were originally written to cover various aspects of industrial and commercial power systems - ranging from equipment sizing to protection, maintenance and installation, safety, and so on. Some of the material was voltage-specific (low vs high), but most books had at least some overlap with the others (i.e. they all had a safety section).

Of the original 13 color books, none are currently in print. Instead, IEEE is working on amalgamating the material into a completely new series of standards (the 3000 "dot" series) meant to collect information according to usage (and hopefully eliminate some of the duplication between books/standards). They are also being brought up to date to reflect the current status of power systems - including changes to the nature of electrical loads and power generation. Not all of the new standards are written yet, unfortunately.

Converting energy to motion for more than half a century

 

[waross]

I took exception to the Orange book's formula for droop.
What seems to universally accepted is that 3% droop = 61.8 Hz or 51.5 Hz no load frequency
That is: full load frequency times (100% + Droop% ie: 103%) = no load frequency.

The Orange Book gives the droop formula as.
One minus the (full load frequency divided by the no load frequency) x 100.
Or 1 minus (60 Hz / 61.8 Hz = .971) = .0291 x 100 = 2.92% Droop.
I have probably looked at over 100 generator specs showing a no-load frequency as 51.5 Hz or 61.8 Hz, and reporting the droop as 3%.
Who is right?
Orange book or the rest of the world.
Sanity check:
The Gurus working in utilities;
When generators other than the swing machine are run at 5% droop, how is the actual frequency calculated?
For 60 Hz, 5% droop = 63 Hz no load frequency.
or
For 60 Hz, 5% droop = 63.152 Hz no load frequency?

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[crshears]

My former utility kept it simple and went with 63 Hz at 5% droop - - but the droop standard for the province of Ontario is 4%, and we always went with a no-load frequency of 62.4 Hz.

CR

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

 

[waross]

Yes, that is based on the formula that I am familiar with, not the Orange Book formula.
Thanks for the information.
Base speed + droop pecentage.
Not, no load speed minus droop percentage as per Orange Book.
Has anyone seen the Orange Book formula applied?

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[wcaseyharman]

The industry including NERC correlate %droop with nominal frequency.
So a unit with 5% droop synchronized at no load at 60Hz would reach full load when the frequency dropped to 57Hz.

Casey

 

[waross]

So when the grid is loaded, the frequency is actually 57 Hz and all the frequency meters showing 60 Hz are wrong?
A conspiracy worthy of MAGA.
Or is it 5% droop is 5% of nominal frequency and no load speed is nominal frequency PLUS 5% of nominal frequency or 63 Hz.

What actually happens is that the incoming set is synchronized at 60 Hz, no load. The throttle/governor is then advanced to a position corresponding to 63Hz at no load., but the grid is locking the set at 60 Hz.
Another way to present it, For a grid tied machine, the load is governed by the governor. 5% droop = 3 Hz.
A governor setting corresponding to 60 Hz = no load.
A governor setting corresponding to 61.5 Hz = 50% load.
A governor setting corresponding to 63 Hz = 100% load.
Despite the governor setting, the grid tie locks the frequency at 60 Hz.

The industry including NERC correlate %droop with nominal frequency.

I agree with this part of your post.

So a unit with 5% droop synchronized at no load at 60Hz would reach full load when the when the frequency dropped to 57Hz.

How about; "So a unit with 5% droop synchronized at no load at 60Hz would reach full load when the governor is advanced to a setting corresponding to 63Hz no load speed."
I was originally commenting on the old Orange book formula that based droop on no load speed rather than nominal frequency.



--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[wcaseyharman]

From the perspective of the governor it’s basically the same thing - the error between the setpoint and the grid frequency in both cases is 5%, and a machine with 5% droop will be at max load in both cases.
And no, I have not seen that orange book formula used anywhere including in the governor control logic on the gas turbine and hydro turbines I’ve worked on/tested.

 

[waross]

Thanks for the confirmation, Casey.
It looks as if we are all on the same page and that page is not in the Orange Book. grin

--------------------
Ohm's law
Not just a good idea;
It's the LAW!

 

[crshears]

And the difference between the formula in the Orange Book and the formula generally used is trivial, relatively speaking; the effect of having some machines using each of the two formulae connected to the same system would not IMHO be profoundly significant.

Bottom line: no need to engage in page-splitting activities!

CR

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