Connected Load and Demand Load - Try searching already

[Soulfield]

New member here. Searched and although there is several posts on the subject I can not find direction on my question, so dont burn me!

All the posts Ive seen confirm what my logical thought that a "Connected Load" would represent everything in the system running wide open continuously. In comparison a "Demand Load" would be the "Peak" load of components in the system that you expect to run at one particular time. What I'm saying is a tenant space (lets say a salon) with hair dryers, electric water heating ect. may run heat in the winter and no AC. The "connected load accounts for both of these, yet the "Demand Load" I see stated on EE's drawings are always higher than the "Connected Load."

Example would be a 1,100 SF Salon Space
200A | 3 Phase | 4W
Electric Water Heating
Split System HP and FCU

Am I missing Something Here?

Shouldn't the "Demand" be Lower than the "Connected":?

 

[charlie b]

It is an interesting question. The trouble you are having is with the meaning of ?demand load.? Although what you described is a reasonable interpretation of the phrase, you aren?t using the phrase in a technically correct manner. It is natural to expect that not all loads in a building will be operating at full capacity at the same time. We are allowed to make allowances for that fact, when we calculate the total amount of power needed to serve the building. So one would expect that the ?demand load? will be lower than the ?connected load.?

But there are certain types of equipment that we have to account for at more than their rated capacity. Motors are one example. We have to take the largest motor in a building, and add 25% to its nameplate load. We have to do the same for storage-tank style water heaters. We have to do the same for any load that is likely to be operating for more than three hours at a time. There are other examples. When we take all these into account, it is not impossible that our ?calculated load,? also known as the ?calculated load after all demand factors have been applied,? or simply known as the ?demand load? might exceed the ?connected load.? I have not seen it happen often, but it can happen.

 

[Soulfield]

Charlie,

Thanks for the clear response. As I stated I'm certainly no EE or PE rather, my electrical know how is limited to 120V stuff in my own house and low voltage for A/V applications.

I figured motors or those parts of the system that draw the most power would be the culprit, and I believe that electric motors probably use more on "start up" than once running. Maybe Im wrong.

It just seems that nearly every panel schedule I see has the higher "demand" load as compared to "Connected" load as stated at the bottom of the panel schedules. I see a clear relationship between say number of light fixtures, their wattage, and the voltage they operate at + a power factor giving a KVA number, but motors I don't get. How does one determine a wattage or amperage number for a motor when only the HP and voltage are known? Is their a formula or are there more variables?

Anyhow thanks again for your direction. The more I know the better my understanding of the "big picture" in building design and system. The help is greatly appreciated!

 

[charlie b]

How does one determine a wattage or amperage number for a motor when only the HP and voltage are known? Is their a formula or are there more variables?

You also need to know if the motor is single phase or three phase. Then all you need to do is look up the amperage in a table. If you have access to a copy of the NEC, you will find the single phase motor amps in Table 430.248, and the three phase motor amps in Table 430.250.

There is a formula that relates these quantities. Essentially, you get power by multiplying voltage times amperage. So if you know two of the variables (e.g., horsepower and voltage), you can get the third (i.e., amps). But there are some complications to the formula. Motor efficiency comes into play, so it is not a simple unit conversion from horsepower to watts. Also, if you are dealing with a three phase motor, then the factor of "the square root of three" will appear somewhere within the formula.

 

[charlie b]

It just seems that nearly every panel schedule I see has the higher "demand" load as compared to "Connected" load as stated at the bottom of the panel schedules.

I just took a quick look at a set of panel schedules for a project we issued last Friday. Out of the first thirty schedules, 12 had a higher connected load, 12 had a higher demand load, and the rest had the same value for connected and demand load. It is not something I generally pay attention to. My only concern is that the demand load has to be less than, or equal to, the rating of the panel. I don't bother comparing the connected load to the panel rating, for that has no meaning to me.

 

[kwired]

It is an interesting question. The trouble you are having is with the meaning of ?demand load.? Although what you described is a reasonable interpretation of the phrase, you aren?t using the phrase in a technically correct manner. It is natural to expect that not all loads in a building will be operating at full capacity at the same time. We are allowed to make allowances for that fact, when we calculate the total amount of power needed to serve the building. So one would expect that the ?demand load? will be lower than the ?connected load.?

But there are certain types of equipment that we have to account for at more than their rated capacity. Motors are one example. We have to take the largest motor in a building, and add 25% to its nameplate load. We have to do the same for storage-tank style water heaters. We have to do the same for any load that is likely to be operating for more than three hours at a time. There are other examples. When we take all these into account, it is not impossible that our ?calculated load,? also known as the ?calculated load after all demand factors have been applied,? or simply known as the ?demand load? might exceed the ?connected load.? I have not seen it happen often, but it can happen.

I guess that would make sense if the demand load uses the 25% adder for largest motor, or water heater and the connected load does not include it, otherwise they are both increased proportionally. What about a 100 HP motor that only is driving 85 Hp of load? I know NEC says we size motor circuit to tables in 430, but how do PE's add this load to the service or feeder? Is it added as 85 Hp or 100 Hp plus power factor, efficiency and demand considerations?

I would still think the connected load would be bases on 100 Hp but the demand load may be based on 85HP still doesn't make sense how you can get higher demand than connected though.

 

[charlie b]

What about a 100 HP motor that only is driving 85 Hp of load? I know NEC says we size motor circuit to tables in 430, but how do PE's add this load to the service or feeder? Is it added as 85 Hp or 100 Hp plus power factor, efficiency and demand considerations?

Presuming this is a 480 volt system, I would look up 100 HP in Table 430.250, see the value of 124 amps, multiply that by 831 (i.e., 480 volts times the square root of 3), get 103 kVA, and show (on the panel schedule) a load of 34.3 KVA on each phase. The panel schedule excel file I use would automatically take note of whether this was the largest motor on the panel, and if so it would add 25% (8.6 KVA) to each phase at the bottom of the page. I would ignore the fact that the mechanical load was only 85 HP, unless I really needed to reduce the calculated load (e.g., if the total was too close to my intended panel rating and I didn't want to get a bigger panel).

 

[JoeStillman]

How I calculate Demand

How I calculate Demand

If you see a panel schedule with a higher demand than connected load, it just means nobody bothered to update the panel schedule. In a large A/E firm, that task is often delegated to an inexperienced designer or CAD guy. (Not that the Engineer of Record isn't still ultimately responsible.)

For residential panels, the demand is calculated according to article 220 and I put the calculation right on the drawing. For commercial, I'll use 10 kVA of receps @100%, remainder @ 50% and any other demand factors like kitchen equipment and non-coincident loads. Lighting is always @100%. In a pinch, I might put a diversity factor on a large group of HVAC loads, but there has to be a sound basis for it like units on opposite sides (i.e. sunny/shady) of a buiding or pump-and-a-spare type reserve.

When one of my guys actually put a 3-pole breaker in a single-phase panel, I got fed up with error filled panel schedules and I wrote an AutoLisp program that calculated connected load and demand automatically. Each load gets characterized as lighting, power, receptacle or non-coincident and the program does the rest. It also prevents panel schedules with more than 42 poles and branches bigger than the mains. (All silly mistakes that I've seen over the years.)

On the heat pump, you can't count the heat and cooling as non-coincident. They both work at the same time when it's in defrost mode.

 

[pfalcon]

I guess that would make sense if the demand load uses the 25% adder for largest motor, or water heater and the connected load does not include it, otherwise they are both increased proportionally. What about a 100 HP motor that only is driving 85 Hp of load? I know NEC says we size motor circuit to tables in 430, but how do PE's add this load to the service or feeder? Is it added as 85 Hp or 100 Hp plus power factor, efficiency and demand considerations?

I would still think the connected load would be bases on 100 Hp but the demand load may be based on 85HP still doesn't make sense how you can get higher demand than connected though.

It has to be demonstrated that the assembly cannot exceed 85HP. You could do that in say, where current limiting would prevent the unit from applying >85HP or it would trip. Load monitors could also be used. But without demonstrating that the design is power limiting you can't derate the motor.

 

[charlie b]

If you see a panel schedule with a higher demand than connected load, it just means nobody bothered to update the panel schedule.

Or it means that the calculated demand load really is higher than the actual connected load. Please see my post #5 above. In each case for which the demand was the higher value, that particular schedule had nothing but motors.

So let me ask you to do the following math: Put five motors on a panel. Count four of them at 100% and one of them at 125%. Which value will be higher: demand :happyyes: or connected?:happyno:

 

[kwired]

It has to be demonstrated that the assembly cannot exceed 85HP. You could do that in say, where current limiting would prevent the unit from applying >85HP or it would trip. Load monitors could also be used. But without demonstrating that the design is power limiting you can't derate the motor.

Derating the motor is not the same as using only the equivelant KVA of 85hp for the feeder or service calculation IMO, that is what I am asking about.

 

[malachi constant]

If you see a panel schedule with a higher demand than connected load, it just means nobody bothered to update the panel schedule... For commercial, I'll use 10 kVA of receps @100%, remainder @ 50% and any other demand factors like kitchen equipment and non-coincident loads. Lighting is always @100%. In a pinch, I might put a diversity factor on a large group of HVAC loads, but there has to be a sound basis for it like units on opposite sides (i.e. sunny/shady) of a buiding or pump-and-a-spare type reserve.

Joe, I disagree with you that demand is never higher than connected. You are missing a few things. Lighting should be at 125%. Any electric heat needs to be at 125%. Any equipment that is on for more than three hours needs to be at 125% - if you are really dotting the i's this could be as mundane as a big flat screen TV in a lobby (but in that example I don't think anyone would begrudge you just calling it a receptacle). You are also missing largest motor. If you had a panel with few receptacles - say 90% lighting, elec heat, and/or motors - I would not at all be surprised to find the demand factor higher than the connected. And as we often put all our lighting on one panel, or all the motors on one panel, this is not that unusual.

I also question the use of a diversity factor. This is allowed in the code for very specific instances (elevators, kitchen equipment). But where in the code are you allowed to apply this at your discretion? Say you have ten overhead doors. The chances of all of them opening at the same time are very small. But you can't prove they won't all go on at the same time. So you don't get to apply a diversity factor. Same with the sunny/shady example. If I were the signing engineer, or inspector, I would in no way allow that calculation, unless you can prove that there is something built into their controls that one is locked out when the other runs. Heating and cooling is allowed to cancel each other out (unless it is something such as cooling that would run in the winter, like for an IT closet or elevator machine room, or process heating needed in the summer). Even in the pump-and-spare scenario I am told by our plumbing engineers that although these are not intended to run at the same time, if there is a significant enough water event they both will run at the same time, and I should size my branch circuits & panel accordingly.

The only time I ever used diversity factors like you suggest above is when trying to estimate the size of a generator - and even then, with the code revisions to 700/701/702 I don't think you can use diversity factors anymore (the revisions I am referring to went into effect in either '08 or '11).

Don't take liberties with the code. Learn it and follow what it says, not what you think is a more reasonable practice. That's how I see it at least.

 

[kwired]

Joe, I disagree with you that demand is never higher than connected. You are missing a few things. Lighting should be at 125%. Any electric heat needs to be at 125%. Any equipment that is on for more than three hours needs to be at 125% - if you are really dotting the i's this could be as mundane as a big flat screen TV in a lobby (but in that example I don't think anyone would begrudge you just calling it a receptacle). You are also missing largest motor. If you had a panel with few receptacles - say 90% lighting, elec heat, and/or motors - I would not at all be surprised to find the demand factor higher than the connected. And as we often put all our lighting on one panel, or all the motors on one panel, this is not that unusual.

I also question the use of a diversity factor. This is allowed in the code for very specific instances (elevators, kitchen equipment). But where in the code are you allowed to apply this at your discretion? Say you have ten overhead doors. The chances of all of them opening at the same time are very small. But you can't prove they won't all go on at the same time. So you don't get to apply a diversity factor. Same with the sunny/shady example. If I were the signing engineer, or inspector, I would in no way allow that calculation, unless you can prove that there is something built into their controls that one is locked out when the other runs. Heating and cooling is allowed to cancel each other out (unless it is something such as cooling that would run in the winter, like for an IT closet or elevator machine room, or process heating needed in the summer). Even in the pump-and-spare scenario I am told by our plumbing engineers that although these are not intended to run at the same time, if there is a significant enough water event they both will run at the same time, and I should size my branch circuits & panel accordingly.

The only time I ever used diversity factors like you suggest above is when trying to estimate the size of a generator - and even then, with the code revisions to 700/701/702 I don't think you can use diversity factors anymore (the revisions I am referring to went into effect in either '08 or '11).

Don't take liberties with the code. Learn it and follow what it says, not what you think is a more reasonable practice. That's how I see it at least.

The 125% factor for continuous loads, motors, etc does not mean the loads are going to draw more, a 10kW heater still demands 10kW not 12.5kW. This factor is for minimum conductor sizing and overcurrent protection selection, not for uprating the actual load.

If you have 10 overhead door openers the chances are not that good they all run at same time. If they do they likely only run for 30 seconds or so and then shut off again. This will not result in any significant heating in the supply system and what extra heat was produced will dissipate rather quickly. Heating and cooling - only needs to not run simultaneously not that cooling can run in winter. If it is winter and heating and cooling are running at same time they are both counter productive of each other. The place you usually run into both running at same time is with heat pumps during defrost cycle, but this cycle is for a limited time and reality deserves some consideration although the code does not really address that.

I don't know how it is possible to demand more than what is connected without something being overloaded. A motor operating in its 1+ service factor maybe but that service factor is not supposed to be continuous either it is supposed to be for temporary conditions. Motor inrush during starting need to be able to be supported by the source but is not taken into consideration when sizing conductors or overload protection, it is considered when selecting short circuit and ground fault protection but only to prevent unnecessary tripping during starting.

 

[david luchini]

Joe, I disagree with you that demand is never higher than connected. You are missing a few things. Lighting should be at 125%. Any electric heat needs to be at 125%. Any equipment that is on for more than three hours needs to be at 125% - if you are really dotting the i's this could be as mundane as a big flat screen TV in a lobby (but in that example I don't think anyone would begrudge you just calling it a receptacle). You are also missing largest motor. If you had a panel with few receptacles - say 90% lighting, elec heat, and/or motors - I would not at all be surprised to find the demand factor higher than the connected. And as we often put all our lighting on one panel, or all the motors on one panel, this is not that unusual.

I think Joe is correct that demand load is never higher than connected load, though I disagree that the reason you see demand loads higher than connected is because "nobody bothered to update the panel schedules."

The demand factor for lighting is 100% (not 125%) for all occupancies except Dwelling Units, Hospitals, Hotels/Motels and Warehouses. These other occupancies will have a lighting demand factor less than 100% (See T220.42). The demand factor for Electric Heat is 100% (not 125%) of connected load (See 220.51) except where the AHJ grants permission to use a demand factor of less than 100%.

220.40 says that the calculated (demand) load of a feeder or service shall not be less than the sum of the loads on the branch circuits supplied (connected load), as determined by Part II of this article, after any demands factors permitted by Parts III or IV, or required by Part V have been applied. None of the demands factors permitted by Parts III or IV, or required by Part V are greater than 100%. Therefore, the demand load can never be higher than the connected load.

The reason we often see demand loads greater than connected loads, is that people are applying the 125% requirement for continuous vs. non-continuous loads for feeders (from Article 215) into what they call the "load calculation" even though they are two separate issues.

 

[malachi constant]

The reason we often see demand loads greater than connected loads, is that people are applying the 125% requirement for continuous vs. non-continuous loads for feeders (from Article 215) into what they call the "load calculation" even though they are two separate issues.

(Talking commercial here...)

How do you separate those issues? If you are loading up your panel that you want to protect with a 100A breaker, and the only load on the panel is 35kVA (~95A) of connected lighting, don't you have to apply per 215.3 a 125% factor to this load in order to size the breaker correctly? So 95A x 125% = 119A. Now you can't use the 100A breaker.

Maybe you can separate them if you are doing something that does not require a single main OCPD? Like if you are using the six disconnect rule you have to apply the 125% continuous factor to each of the six main breakers, but when adding up the total loads on the service panel you can disregard the continuous factor. I could buy that. Never thought of it that way before. But if you have a single main disconnect, would you not have to apply the 125% factor for continuous loads?

I was under the impression lighting was considered a continuous load. I am not seeing anything that absolutely defines it as such, but if anything is going to be left on for three hours at a time it is a light or an electric heater. Even with the widespread use of occ sensors I would have a hard time considering a luminaire as a noncontinuous load.

 

[charlie b]

We are getting confused again over the use of the word, ?demand.? There are two common meanings for this word, and they are very different from each other.

?Demand? is used to describe the amount of load you actually see during day-to-day operations. That value is certainly less than the total connected load, for the simple and obvious reason that not everything will be running at full capacity at the same time. Mathematically, in this context of the word, ?demand,? it is physically impossible for the demand value to exceed the connected value.

?Demand? is also used to describe the results of a formal load calculation. That process sometimes requires us to add 25% to a load?s nameplate value. That can result in the calculated load being higher than the total connected load. The number that appears on a panel schedule under the phrase ?demand load? is this value, not the ?day-to-day actual load you really see? value.

 

[charlie b]

220.40 says that the calculated (demand) load of a feeder or service shall not be less than the sum of the loads on the branch circuits supplied (connected load), as determined by Part II of this article, after any demands factors permitted by Parts III or IV, or required by Part V have been applied. None of the demands factors permitted by Parts III or IV, or required by Part V are greater than 100%. Therefore, the demand load can never be higher than the connected load.

You are missing the fact that the calculation of the branch circuit loads is not just the connected load, but sometimes includes a 25% factor. Specifically, 220.14(C) tells us to calculate motors per 430.22 and 430.24, and that is where the 25% comes into play. As I mentioned earlier, I just issued a design that had some panels with nothing but motors, and their panel schedules correctly show demand as being higher than connected.

 

[david luchini]

(Talking commercial here...)

How do you separate those issues? If you are loading up your panel that you want to protect with a 100A breaker, and the only load on the panel is 35kVA (~95A) of connected lighting, don't you have to apply per 215.3 a 125% factor to this load in order to size the breaker correctly? So 95A x 125% = 119A. Now you can't use the 100A breaker.

Maybe you can separate them if you are doing something that does not require a single main OCPD? Like if you are using the six disconnect rule you have to apply the 125% continuous factor to each of the six main breakers, but when adding up the total loads on the service panel you can disregard the continuous factor. I could buy that. Never thought of it that way before. But if you have a single main disconnect, would you not have to apply the 125% factor for continuous loads?

I was under the impression lighting was considered a continuous load. I am not seeing anything that absolutely defines it as such, but if anything is going to be left on for three hours at a time it is a light or an electric heater. Even with the widespread use of occ sensors I would have a hard time considering a luminaire as a noncontinuous load.

You separate the two issues because they are in completely separate sections in the code. Article 220 tells you how to calculate the load for any specific feeder. Article 215 tells you how to size the feeder conductors based on the load on the feeder calculated by Article 220. These are two separate issues covered by two separate code sections.

Lets say I had two feeders serving two separate panels. One panel was going to supply all my lighting loads, and the other panel was going to supply all my receptacle loads in a commercial setting, lets say a 25000 sf "Club". The minimum lighting load per T220.12 is 25,000 x 2 = 50,000VA. There area 500 duplex receptacles at 180VA each (per 220.14(I)) = 90,000VA.

I then apply the demand factors from Part III of Article 220. For lighting, I have 50,000VA with a demand factor of 100% (T220.42) = 50,000VA. For the receptacles, I have 90,000VA with a demand factor of 100% for the 1st 10kVA, and 50% for the remainder (T220.44) = 10,000 + (0.5)*80,000 = 50,000VA. The calculated (demand) load for each panel is 50,000VA.

Next, I need to size the feeders for each of the panels. The lighting I would consider continuous load, and the receptacles I would consider non-continuous load. I would go to Article 215 for this.

The feeder for the receptacle panel must have an ampacity of at least 50,000VA/208V/1.732= 139Amps. The minimum OCPD size would be 150A. This panel could be fed with #1/0AWG and a 150A C/B.

The feeder for the lighting panel must have an ampacity of at least 50,000VA*1.25/208V/1.732= 173Amps. The minimum OCPD size would be 175A. This panel could be fed with #2/0AWG and a 175A C/B.

Both panels have the same calculated load after demand factors have applied (Article 220.) The lighting panel requires a larger feeder than the receptacle panel (Article 215.)

 

[david luchini]

You are missing the fact that the calculation of the branch circuit loads is not just the connected load, but sometimes includes a 25% factor. Specifically, 220.14(C) tells us to calculate motors per 430.22 and 430.24, and that is where the 25% comes into play. As I mentioned earlier, I just issued a design that had some panels with nothing but motors, and their panel schedules correctly show demand as being higher than connected.

I am not missing that fact at all. 220.14(C) tells us to calculate the outlet load for motors in accordance with 430.22 and 430.24. That means that the 25% factor is already part of the "connected" load. The same number is carried through to the "demand" load in Part III.

For instance, lets take a look at a panel with nothing but motors. Say a panel that had four branch circuits feeding four 5HP, 208V, 3ph motors. Per 220.14(C), each motor branch circuit has a load of 7,521VA (as calculated per 430.22.) The total connected load on the panel (the sum of all the branch circuit loads) is 4*7,521VA=30,084VA.

Now lets apply the demand factors for the panel feeder, from Article 220 Part III. 220.50 directs us to 430.24, which tells us that the load would be (16.7A*1.25 + 3*16.7A) * 208V * 1.732 = 25,570VA. The load after the demand factors of Article 220 Part III are applied is smaller than the load calculated for the individual branch circuits per Part II of Article 220.

 

[malachi constant]

David,
In your example, let's say there existed options of getting a panel rated 100A, 125A, 150A, 175A and 200A. (I know they are generally 125A and 225A but I'd rather not reinvent the calcs to ask the question.) Your receptacle panel would be minimum 150A, correct? And your lighting panel would be minimum 175A, correct? So even though, using your terms, the step 1 "connected load" for the receptacle panel is greater, and the step 2 "demand load" of the two panels is equal, but the step 3 "I-don't-what-to-call-it-anymore load" (my term!) of the feeders dictates the lighting panel be larger.

I think we may be in agreement as to the application of the code, but are differing in our vocabulary? I don't know that code speaks to what each of these steps should be called. I have always considered the second step (what you call "demand load") to just be a step on the way to a final load - step 3, which I always called demand load. Maybe that should be "calculated load", though that seems to be what many here are also applying to step 2.

I remember one engineer working under me always getting irritated that in doing multifamily calcs I called the common area loads "house loads". One day we were walking through the codes for multifamily and he was like "oh, you're actually calling it what the code calls it!" And he relaxed after that. I wonder if what we have here is a bunch of people bristling about what they think it should be called, but there is no clearly defined name for it in the code. Or maybe there is! And I am calling it wrongly. Someone enlighten me!