peak utility readings vs actual loads

Status
Not open for further replies.

malachi constant

Senior Member
Location
Minneapolis
Question. When we ask for existing utility data, the utility in our area (Xcel Energy) reports max KW on a monthly basis. For instance on one facility that opened in 2011 with a 1600A 208V service we have a max reading in July 2012 of 138KW.

I am sizing a generator to back up this service. As it is has only been in operation through one summer I am being very conservative and sizing the generator at 200kW. One engineering colleague of mine says that the 138kW is only the max AVERAGE taken over a fifteen minute span, and that the actual instantaneous loads on the system during that fifteen minutes are almost certainly larger, maybe sustained for 5-10 minutes a time. He recommends that I size the generator even larger than I have it, or at least tie the air conditioning controls to the transfer switch such that the cooling is kicked off whenever the generator comes on.

I had never considered that peak demand was not actually peak. As it relates to electrical equipment (switchboards, feeders, etc) I am not concerned - code clearly allows you to take 125% of max meter readings and use that as a starting point on existing installations. But what about generators? How much larger could the real peak be than the metered peak?

I know, I know, I should get out the generator sizing program and plug in all my motors and lights and equipment. The problem with that is you don't know what the diversity factors are - how often the lights are on, how often the motors start, how long they run, etc - and there are a lot of loads to consider. When available, the peak load is, in my opinion, very strongly preferred.

All that to say...any thoughts/experience as metered peak load relates to actual peak load? Is it something I can ignore and not lose any sleep?

Thanks!!
 

Shoe

Senior Member
Location
USA
I'm curious to see other replys, but I would have done something similar to what you have done. As far as generator sizing, I'd be most concerned about the largest loads (chillers, etc.) as far as starting capacity is concerned. If they are all small loads, then you're probably OK.
 
Last edited:

shamsdebout

Senior Member
Location
Macon,GA
I was always of the understanding that the peak demand for the entire month was the highest reading for a range of time. I have heard of this being 15 minutes. I think that peak is the highest reading recorded by the meter so that would be the peak demand for that month.

I think your approach is one that I too would follow. I guess the only other thing to consider is if the user will need additional capacity.
 

mivey

Senior Member
One engineering colleague of mine says that the 138kW is only the max AVERAGE taken over a fifteen minute span, and that the actual instantaneous loads on the system during that fifteen minutes are almost certainly larger, maybe sustained for 5-10 minutes a time. He recommends that I size the generator even larger than I have it, or at least tie the air conditioning controls to the transfer switch such that the cooling is kicked off whenever the generator comes on.
He is right.

I had never considered that peak demand was not actually peak. As it relates to electrical equipment (switchboards, feeders, etc) I am not concerned - code clearly allows you to take 125% of max meter readings and use that as a starting point on existing installations. But what about generators? How much larger could the real peak be than the metered peak?
It would not suprise me if it was twice that. No way to tell without actually doing some real looking.
 

kbsparky

Senior Member
Location
Delmarva, USA
It would be easy to install some load-shedding means to limit the HVAC loads during periods of high peak usage.

No need to oversize beyond reasonable load requirements.
 

mivey

Senior Member
I was always of the understanding that the peak demand for the entire month was the highest reading for a range of time. I have heard of this being 15 minutes. I think that peak is the highest reading recorded by the meter so that would be the peak demand for that month.
It is the highest block of energy used in a 15-minute interval. Another way to say it is that it is the highest average of instantaneous demands in a 15 minute period.

It is by no means the highest demand encountered. Motor starting capacity is what the generator needs to be able to handle for a few seconds (generally speaking). The generator peak capacity may be what is required for periods much less than 15 minutes (the shorter period will always be equal to or greater than the 15 minute average).

I think your approach is one that I too would follow.
That would be VERY dependent on the load type.
 

beanland

Senior Member
Location
Vancouver, WA
Integrated Peak

Integrated Peak

Utility meters always measure energy. As a result, utility meters call the highest energy use in a set block of time the demand. Utilities use 1 hour, 30 minute, 15 minute (very common), and 5 minute intervals. As a result, the demand from the utility is never close to the instantaneous demand of the actual loads, it is the average load over the interval. I would expect the actual peak instantaneous demand to be 4-5x the average, maybe more. Without actual measurements, the HVAC interlock is an excellent idea. Remember to consider all motor starting loads including elevators which can have huge starting current demands and don't add up to much in the utility peak demand because they operate for such short periods of time.
 

malachi constant

Senior Member
Location
Minneapolis
Thanks for corroborating what I had heard. This goes under the "learn something new every day" category.

I spoke with the Owner on this project, they are comfortable proceeding with the modestly-sized generator in question, and if it has trouble they have no problem doing manual load shedding, or adding in automatic load shedding after the fact. A reasonable gamble considering we're putting a 200kW generator on a service that has only seen a metered peak of 138kW in 15 months of use.

Thanks again!
 

GoldDigger

Moderator
Staff member
Location
Placerville, CA, USA
Occupation
Retired PV System Designer
A reasonable gamble considering we're putting a 200kW generator on a service that has only seen a metered peak of 138kW in 15 months of use.

...The problem with that is you don't know what the diversity factors are - how often the lights are on, how often the motors start, how long they run, etc - and there are a lot of loads to consider. When available, the peak load is, in my opinion, very strongly preferred....

Diversity is your friend, but if there is a period between the grid going down and the generator picking up the load, you really have to allow for everything trying to start at once (unless there is something designed to prevent that), including AC compressor motors that have not had time to unload system pressure and so may keep drawing high current until their overload trips.
Better to take at least a cursory look at those possibilities first instead of waiting to see if there is problem when you first need the generator.
 
Last edited:
Question. When we ask for existing utility data, the utility in our area (Xcel Energy) reports max KW on a monthly basis. For instance on one facility that opened in 2011 with a 1600A 208V service we have a max reading in July 2012 of 138KW.

I am sizing a generator to back up this service. As it is has only been in operation through one summer I am being very conservative and sizing the generator at 200kW. One engineering colleague of mine says that the 138kW is only the max AVERAGE taken over a fifteen minute span, and that the actual instantaneous loads on the system during that fifteen minutes are almost certainly larger, maybe sustained for 5-10 minutes a time. He recommends that I size the generator even larger than I have it, or at least tie the air conditioning controls to the transfer switch such that the cooling is kicked off whenever the generator comes on.

I had never considered that peak demand was not actually peak. As it relates to electrical equipment (switchboards, feeders, etc) I am not concerned - code clearly allows you to take 125% of max meter readings and use that as a starting point on existing installations. But what about generators? How much larger could the real peak be than the metered peak?

I know, I know, I should get out the generator sizing program and plug in all my motors and lights and equipment. The problem with that is you don't know what the diversity factors are - how often the lights are on, how often the motors start, how long they run, etc - and there are a lot of loads to consider. When available, the peak load is, in my opinion, very strongly preferred.

All that to say...any thoughts/experience as metered peak load relates to actual peak load? Is it something I can ignore and not lose any sleep?

Thanks!!

Your thinking is fine and even more conservative that I may have choosen. Remember that the generators are also capable to handel overload for short periods and I think that even some wacky cycling load is still within the capability of a 200kW unit.

Still, run the gen sizing for starting loads to make sure you haven't got a motor with a nasty inrush.
 

mivey

Senior Member
Remember that the generators are also capable to handle overload for short periods
What is that range? Do you have the overload capabilities handy?

Since we are probably talking about emergency or standby rating, I don't think there is an allowance for operating above the standby rating. A prime unit can have a 10% overload for 1 out of 12 hours for a limited number of hours per year.
 
What is that range? Do you have the overload capabilities handy?

Since we are probably talking about emergency or standby rating, I don't think there is an allowance for operating above the standby rating. A prime unit can have a 10% overload for 1 out of 12 hours for a limited number of hours per year.

We are talking about part of a 15 minute demand period and not a sustained load at that either.
 

mivey

Senior Member
We are talking about part of a 15 minute demand period and not a sustained load at that either.
Sorry, but it doesn't have to be all that long. A generator can probably handle a 300% load for a few seconds, at least enough for a motor to start. But I checked and there is no rated overload capacity for a standby generator. Sustain a significant overload, even for less than 15 minutes, and you will trip the unit (hopefully trip, rather than damaging the unit as well as suffer other problems due to low frequency and low voltage).

Even if we talk about an overload that lasts just a minute or two, sizing a generator too small so that it sees the non-start overload cycling on and off at regular intervals is not what the generator is supposed to handle.

Now there is a rough rule of thumb that says take the 12-month 15-minute peak and add 20-25% to get the generator size. The problem with rules of thumb is that they are just that. You still need to look at the load to determine if the rule of thumb probably applies (like by making a load count or gathering 1 minute or less interval data).

Consider the two problems of undersizing and oversizing.If you oversize, you can have problems like wet stacking. If you undersize then you can't handle the peak load. There are times when you need to add load banks or parallel smaller units to handle both the peak load and average load.

Overloading a unit not designed to handle overloads is just not the way to go.
 
Sorry, but it doesn't have to be all that long. A generator can probably handle a 300% load for a few seconds, at least enough for a motor to start. But I checked and there is no rated overload capacity for a standby generator. Sustain a significant overload, even for less than 15 minutes, and you will trip the unit (hopefully trip, rather than damaging the unit as well as suffer other problems due to low frequency and low voltage).

Even if we talk about an overload that lasts just a minute or two, sizing a generator too small so that it sees the non-start overload cycling on and off at regular intervals is not what the generator is supposed to handle.

Now there is a rough rule of thumb that says take the 12-month 15-minute peak and add 20-25% to get the generator size. The problem with rules of thumb is that they are just that. You still need to look at the load to determine if the rule of thumb probably applies (like by making a load count or gathering 1 minute or less interval data).

Consider the two problems of undersizing and oversizing.If you oversize, you can have problems like wet stacking. If you undersize then you can't handle the peak load. There are times when you need to add load banks or parallel smaller units to handle both the peak load and average load.

Overloading a unit not designed to handle overloads is just not the way to go.

I recommend that you familarize yourself with ISO 8528 and ISO 3046, which define the short, long and sustained loading profile capabilities of standby and prime mover generators.
 

mivey

Senior Member
I recommend that you familarize yourself with ISO 8528 and ISO 3046, which define the short, long and sustained loading profile capabilities of standby and prime mover generators.
If you are familiar with whatever is the applicable standard, then why don't you just answer the question from post #11? It would make things much easier if you were so inclined. I read the short-time rating info from various manufacturer sites but if you have better info, please share.
 

mivey

Senior Member
I recommend that you familarize yourself with ISO 8528 and ISO 3046, which define the short, long and sustained loading profile capabilities of standby and prime mover generators.

I read the short-time rating info from various manufacturer sites but if you have better info, please share.

For example:

Kohler said:
Standby rating applies to installations served by a reliable utility source. Standby power is applicable to varying loads for the duration of the utility power interruption. There is no overload capability for the standby rating; use of the generator set above the standby rating is prohibited. Ratings are in accordance with ISO 3046/1, BS 5514, AS 2789, and DIN 6271.
 
If you are familiar with whatever is the applicable standard, then why don't you just answer the question from post #11? It would make things much easier if you were so inclined. I read the short-time rating info from various manufacturer sites but if you have better info, please share.

Your post was stating generalities and you're criticizing mine for specifics?

http://www.cat.com/cda/files/4252018/7/Demystifying Generator Set Ratings LEXE0549-00.pdf

The sum of it is that the same engine can deliver 25% more at its Prime rating with time limitation than it's Continuous rating. It will be WELL within any NORMAL demand fluctuation that occurs during the 15 min period in any 43200min(1 average month). This is what my 40 years of experience tells me. Y'all are welcome to anal-yze the 'actual' load conditions and come to a different conclusion.
 

mivey

Senior Member
Your post was stating generalities and you're criticizing mine for specifics?
If you consider it criticizing then fine. You just alluded to some information you knew about that would show something different than what I have said. I was asking for that information so I might could learn from it. It's OK if you don't have any but don't consider my asking for it to be criticism.

http://www.cat.com/cda/files/4252018/7/Demystifying Generator Set Ratings LEXE0549-00.pdf

The sum of it is that the same engine can deliver 25% more at its Prime rating with time limitation than it's Continuous rating. It will be WELL within any NORMAL demand fluctuation that occurs during the 15 min period in any 43200min(1 average month).
All well and good but does not differ from what I have said. The standby rating does not have an overload capacity rating.

Besides, what does the prime rating have to do with the standby rating that I was talking about? Based on what the OP has said, this appears to be a backup unit, not a prime power unit. And the information you provided does not say you can exceed the standby rating.

However, I have provided information about short-time ratings which I understand to be for motor starts and such. The normal short-term capability seems to be 250-300% of the unit rating (and on the order of around 10 seconds).

Nothing you have provided suggests we can exceed the standby rating for times on the order of minutes. Especially with a load that might be twice the 15-minute average or more. I would say that if it were a winter peaking load then maybe we could exceed the standard rating a little but that would remain to be seen.

This is what my 40 years of experience tells me. Y'all are welcome to anal-yze the 'actual' load conditions and come to a different conclusion.
Probably will work fine for most loads. However, there are reasons they make load banks and parallel units and if one exercises due diligence then one should at least look at the loads and not just the peak demand data. JMO.
 
If you consider it criticizing then fine. You just alluded to some information you knew about that would show something different than what I have said. I was asking for that information so I might could learn from it. It's OK if you don't have any but don't consider my asking for it to be criticism.

All well and good but does not differ from what I have said. The standby rating does not have an overload capacity rating.

Besides, what does the prime rating have to do with the standby rating that I was talking about? Based on what the OP has said, this appears to be a backup unit, not a prime power unit. And the information you provided does not say you can exceed the standby rating.

However, I have provided information about short-time ratings which I understand to be for motor starts and such. The normal short-term capability seems to be 250-300% of the unit rating (and on the order of around 10 seconds).

Nothing you have provided suggests we can exceed the standby rating for times on the order of minutes. Especially with a load that might be twice the 15-minute average or more. I would say that if it were a winter peaking load then maybe we could exceed the standard rating a little but that would remain to be seen.

Probably will work fine for most loads. However, there are reasons they make load banks and parallel units and if one exercises due diligence then one should at least look at the loads and not just the peak demand data. JMO.

You make several assertion that is contrary to the document that I've supplied.

Try to actually read what that paper says. I'll not be your reading tutor.
 

mivey

Senior Member
You make several assertion that is contrary to the document that I've supplied.
Since you do not seem to want to explain your statements then I'll have a look for myself:

The generator peak capacity may be what is required for periods much less than 15 minutes.
A generator can probably handle a 300% load for a few seconds, at least enough for a motor to start.
...
there is no rated overload capacity for a standby generator.
...
Sustain a significant overload, even for less than 15 minutes, and you will trip the unit
...
Even if we talk about an overload that lasts just a minute or two, sizing a generator too small so that it sees the non-start overload cycling on and off at regular intervals is not what the generator is supposed to handle.
These items are not addressed in your paper but are addressed in sources like the paper I referenced discussing things like motor starts and other short-time loads and generator protection schemes.


Since we are probably talking about emergency or standby rating, I don't think there is an allowance for operating above the standby rating.
And we have:
weressl's reference said:
{Fig 3: no overload except for the prime rating and even it maxes out at the 2.5 MW genset standby rating. This is consistent with other manufacturing data, including the one from Kohler that I quoted earlier}


A prime unit can have a 10% overload for 1 out of 12 hours for a limited number of hours per year.
And we have:
weressl's reference said:
A Prime-rated generator set is capable of providing full nameplate rating for a period of time, but it must have an average load factor of no more than 70 percent of the Prime rating. Ten percent overload is allowed for emergencies at a maximum of one in 12 hours and no more than 25 hours per year.


I would say that if it were a winter peaking load then maybe we could exceed the standard rating a little but that would remain to be seen.
And we have:
weressl's reference said:
As these ambient changes move away from the maximum ambient design, they will affect engine performance in a positive manner.


However, there are reasons they make load banks and parallel units and if one exercises due diligence then one should at least look at the loads and not just the peak demand data.
I should also add load-shedding in addition to load banks and parallel generating units.

You make several assertion that is contrary to the document that I've supplied.

Try to actually read what that paper says. I'll not be your reading tutor.
If you make statements but can't or won't explain then we have communis interruptus. It's really up to you but if you don't want to communicate then what is your function?
 
Status
Not open for further replies.
Top