How would you estimate peak water demand? 1

[CivilEngineer401]

I am working on a project for a 50 unit apartment building and trying to size the water service. The service is 750-ft long, so due to the length I am concerned about dynamic losses. Since I am not designing a fire flow, I feel that there is no need to be overly conservative and do not want to un-necessarily oversize the pipe. To estimate the peak flow, I started the same way I would estimate peak sewer flows:

50 units x 100 gpd = 5,000 gpd (or 3.5 gpm) multiplied by a peaking factor of 4 = 14 gpm total peak flow for the entire building

I quickly realized that this answer does not make sense since a few showers/sinks on at the same time would surely exceed 14 gpm. I searched online and found the “fixture units method” but the examples online had the peak demand of roughly 20 gpm when you are only looking at 1 house. For one house, I cannot imagine more than 2 sinks and an appliance operating at the same time one which would be roughly 5 gpm.

As an engineer, I started to get quite curios about how to estimate peak water demand. How would you estimate the peak instantaneous water demand be for 5 units, 50 units, 500 units, and 5,000 units?

 

[IRstuff]

Have you seen:

http://rpitt.eng.ua.edu/Class/Senior Design/Water demand and distribution.pdf

TTFN
faq731-376

Need help writing a question or understanding a reply? forum1529

Of course I can. I can do anything. I can do absolutely anything. I'm an expert!

 

[bimr]

You have an error. The flow from each residential unit is usually 100 gpd per person times 3.5 persons per unit, or a total of 350 gpd per residence.

Anyway, building services are usually estimated from the number of plumbing fixtures.

 

[fel3]

Depending on the size of the apartments, there could be fewer people per unit on average. Also, to the OP, I think a peaking factor of 4.0 is too small for a 50-unit complex. For example, about ten years ago, I did the civil engineering for a small (45,000 sf) medical office building that was later recorded to have a peaking factor of about 15.

I'm out of the office this week and don't have all my data at hand, but I do have my fixture unit spreadsheet on this computer. If we assume 50 apartments with one bathroom and one kitchen plus laundry, I get around 600 water service fixture units as follows. For a central laundry facility and/or more bathrooms per unit, adjust accordingly.

50 shower/bath combos * 3.5 = 175 WSFUs
50 clothes washers * 2.5 = 125 WSFUs
50 dishwshers * 1.0 = 50 WSFUs
50 domestic kitchen sinks * 1.0 = 50 WSFUs
50 lavatories * 0.5 = 25 WSFUs
50 water closets (1.6 gpf gravity tanks) * 2.5 = 125 WSFUs
50 hose bibbs * 1.0 = 50 WSFUs (I ignored the 2.5 for the first HB)

Total = 600 WSFUs. This translates to about 157 gpm peak flow. (This result is per the 2009 California Plumbing Code and probably about the same per other plumbing codes.)

If average demand is (50 u)(3.5 ppu)(100 gpd/p) = 17,500 gpd = 12.2 pgm, then the peaking factor would be 12.9.

==========
"Is it the only lesson of history that mankind is unteachable?"
--Winston S. Churchill

 

[Xalii]

I think we have to be careful when we are speaking about peaking factor. There is two different things:
1) the peak flow of the daily average demand of a large population that can range from 150% to 400%. This figure is used to design main pipes and tanks serving at least 200 to 500 households.
2) but a pipe for a single tap has to be design for the required flow at this tap, and this can be hundreds times more than the daily average times the peak factor for this tap as it is only working a few minutes. Therefore you have to design the pipe for 1 tap at a flow of 3 GPM, and for two taps at a flow of 6 GPM as you assume that both of them can be open at the same time, but for the pipe supplying 10 taps, you will not assume that all of them are open thus you don't need to design it for 30 GPM, but have to apply a simultaneous factor (Hunter's curve) according to IPC or other standards reducing the flow, using the WSFU as a unit for a "simple tap". Thus roughly if you have a pipe providing 10 WSFU you only use ~ 15 GPM and for the pipe supplying 20 WSFU you have only 20 GPM and so on.

As explained by fel3 for 50 units you would have a peak factor of 12.9 but for 100 unit, taking 1200 WSFU would give about 230 GPM, thus only a peak factor of 9.4.

You can have a look at the following table allowing you to design a simple simultaneous curve under chapter 6 : Distribution networks
:

http://designwss.weebly.com/2-concept-design.html

 

[CivilEngineer401]

IRstuff: Thank you, that pdf was very helpful.

bimr: I am basing 100 gpd based on what I have seen from township billings. I have looked at water billings for 3 townships that my company represents (total residential use divided by the number of customers, not the individual billings themselves) and based on what I have seen, the average home uses roughly 140 gpd. I looked at my own bill for my family of 5 (including a wife that loves to take long showers) and I used an average of 132 gpd last year. I assumed an apartment would have no lawn to water, less than 5 people per unit, and water efficient appliances so I thought 100 gpd seemed like a reasonable number.

Xalii: I agree with what you are saying about Peaking Factors. Although they are heavily used when analyzing entire systems, I guess they are not the right tool to use when analyzing a small system with individual fixtures.

 

[bimr]

Interesting approach that you have taken.

Most regulatory agencies have design standards that one is expected to follow. If you disagree with the agency design standards (ie 100 gpd per person), you should take it up with the agency rather than making your own methods and assumptions.

From your perspective, I see little reward for you to develop your own flow estimates unless maybe you own the water company. You are risking a lawsuit should you have guess incorrectly and you will not have a leg to stand on.

You are not actually saving any money anyway because the water lines are sized for fire flows anyway. Upsizing water pipes by 1 pipe diameter is relatively inexpensive as well.

If you reread my post above, I actually recommended that you use the plumbing fixture method to design the water capacity. Fel3 has provided an outline of that method.

 

[PEDARRIN2]

Since you are sizing the pipe, why would you worry about the daily water use. It is irrelevant. The fixture unit method should be used, but with caution.

Consider what demographic would be living in this apartment building.

If it is near a college/university, then the likelihood of lots of students would drive the flow demand more toward a dormitory, i.e. lots of shower load in the morning, not a lot anytime else. If 70% of the occupants take a shower during an early morning hour period, then you would want pipe sized to handle 35 gpm as the peak.

If the demographic is more mature, then lower flows would be required. Use the fixture unit method and Hunter's curve to arrive at a likely peak flow. Bump it up ~10% to be conservative and design it for less than 8 feet per second.

2" pipe would likely suffice.

 

[davefitz]

Since it is a new building, modern appliances use less water flowrate than the 1950's units seen in "The Honeymooners", which may be the basis of the plumbing code fixture unit flowrates. In particular, the toilets now flush with less than 1 gal/flush, the shower heads are required to have low flow orifices ( but that is the first item the new occupant replaces with a multihead device), point-of-use hot water heaters are now available, some cost conscious folks are now using cold water detergent exclusively, and modern washing machines use less water. So 100 gal/day/person is not a valid number anymore, and is not going to be sustainable in California, Arizona, or Nevada in the near future.

For peak flows, I would assume the toilets and vanities are fed by 3/8" valves and piping, and the showers have the code required orifices installed. The peak flow would actually occur during the last commercial of the superbowl.

"Whom the gods would destroy, they first make mad "

 

[PEDARRIN2]

davefitz,

I would agree with most of what you say, but for the toilets, they do not use less flow rate, they use less flow duration. A tank type toilet may flush 1.28 gallons per flush, but still requires 3-6 gpm per code. A flush valve toilet will require 25 gpm.

Also, with lavatories and showers, we just have the water on longer to compensate for the lower flow. Ask any female how she rinses her long hair out after washing it. She uses the same amount of water, just takes longer to do it.

 

[davefitz]

Pedarrin2,

Maybe so, but if the question is "what is the peak flow", then the answer would be related to the flowrate thru the number of fixtures in service times the number of fixtures simultaneously in service ; if the flow per fixture drops by 50% due to changes in codes and appliance design, then the peak flow also drops by 50%.

If the new building is in a critical water shortage area, then it may be a good idea to consider provisions to later installing water re-use piping .

"Whom the gods would destroy, they first make mad "

 

[PEDARRIN2]

davefitz,

The OP was trying to size his main, which would be directly related to peak flow rate, not flow duration. A building can flow 0.5 gpm for 24 hours flow 720 gallons per day, but the pipe size would only be ~0.75" to handle the flow rate.

He was trying to "reverse engineer" his calculations from peak gpd to peak gpm, which is not the correct way to do it.

I believe Hunter devised his curve in the 1920's - so it is very anachronistic, but since it was a relationship which included flow rate and probability of use - it is not too hard to determine what the actual projected flow rate would be. But it is very difficult to correlate gpm found from the curve and gpd which will relate to total usage. How to do that accurately is still debatable.

Although they may be available, I would not specify a toilet that uses less than 1 gpf. I do not even like the 1.28 gpf. Maybe in an apartment building, with plenty of shower flow (even low flow) to help keep things flowing, but as a general rule, there is a definite lower limit to toilet flow. There are just too many issues with drain line carry and maintenance complaints. You can only put so many clean outs in the piping.

 

[semo]

USDA (as well as Missouri) has an equation that is commonly used for sizing a distribution line. It is 9 * x^0.515 = Q. Based on an x (connections) of 50 you get a peak flow of 68 gpm. This equation takes into account that low numbers of connections use a higher flow per connection; but, also takes into account that larger number of connections won't have them all flowing at the same time.

 

[PEDARRIN2]

semo,

I think there might be a flaw in an equation based on connections.

Say you have a stadium with 50 water closets, which flush 25 gpm (for a short period). 68 gpm would be less than 3 water closets flushing simultaneously - which could happen at breaks.

Say you have a facility with 50 low flow hand washing stations, which use 0.35 gpm. 68 gpm would far exceed the flow with all the fixtures flowing simultaneously, which would only be 17.5 gpm.

If "connections" is defined as the number of dwelling units, it would likely still fall short. If three water closets flush simultaneously, the flow would be 75 gpm. This does not include any shower load or lavatory load or dishwashing/laundry.

 

[Spartan5]

I think this is a great discussion that brings to light some of the problems with mixing prescriptive methods of systems sizing across different types/scales of systems.

Take bimr's 100 gpd/capita for instance at 3.5 people per unit. That's not wrong.

Take Almighty Tom's peaking factor of 4. That's not wrong either.

Put the two together, and you have a pretty conservative method for sizing a sewer system serving a community or region. You look at it though and you see a peak of 1,400 gpd/dwelling and you think, "Gosh! That's a lot of water!!" And it is. And for good reason. Within that number is "routine" infiltration and inflow, and that also associated with a peak storm. It's good to be conservative in sewer system design. When things go wrong, poop winds up where it's not supposed to.

Would I use that to size the water system for a building of 50 units, probably not.

I like the way fel3 looked at it. You can probably eliminate hose bibs, and maybe even laundry from each apartment depending. But at the end, and it's not "wrong" necessarily, there's the 350 gpd/dwelling hiding in his calculations for his peaking factor. Try to imagine, under any situation, where a single apartment, regardless of how many people are there, could use 350 gallons? Now try to imagine all 50 of them doing it on the same day. How does that happen? Mixing prescriptive methods, that's how.

AlmightyTom looked at the Fixture Unit Method and asked how can a single family house have an instantaneous demand of 20 gpm? I dunno... a hose bib watering the lawn, laundry and dishwasher filling, shower, and a toilet flush (and maybe a factor of safety because, as mentioned, 1 pipe size larger shouldn't break the bank". I could probably pull that off all by myself if I was getting ready to head out of town.

PEDARRIN2 has a good idea to look at the demographic. But bear in mind this building will be there for some time and demographics change.

davefitz has a good point. Fixtures have changed. Prescriptive methods haven't really (with the exception of the Fixture Unit Method). It's not a bad thing though to have pipes too big (to a point). EPAct 1992 changed things pretty drastically in that regard.

davefitz also hit the nail on the head with regards to water-reuse piping and "critical water shortage area". That's going to be a real thing, folks.

Just my thoughts and a hope that people try to take a comprehensive look at this sort of question. I used to design on-site treatment systems for individual houses and small communities and was saddled with 450-600 gpd design "criteria" with little or no basis behind it (aside from "That's how sewer systems are designed", so I've mulled this over a bit.

On a related unrelated note, while PEDARRIN2 is disinclined to specify low flow fixtures, we need to figure out how to make it work. If I had it my way, every new dwelling would have a waterless urinal and/or waste diverting toilet installed somewhere within it. Take it a few crazy steps further and add a tank outside to collect the urine. Nitrogen, phosphorous, and potassium you say? Sounds like fertilizer to me. Pumping water, treating it, pumping it, flushing a toilet, pumping/treating it (somewhat anyway) to deal with the disposal of something with intrinsic value strikes me as shortsighted. I guess as long as we continue to break the nitrogen cycle and make chemical fertilizers to apply to our lawn it's not a big deal

http://news.nationalgeographic.com/...cycling-urine-human-waste-compost-fertilizer/

http://www.scientificamerican.com/article/human-urine-is-an-effective-fertilizer/

I digress.

 

[PEDARRIN2]

The industrialized world worked so hard to get indoor plumbing into the house because to have an outhouse was so primitive, now the current direction is to put the outhouse in the house because to use water is so wrong.

I am not against low flow fixtures, but they have to be balanced with how to get the waste out. The owner that is so happy to get his low flow toilet is not so happy when he has to snake his pipes because there simply is not enough water to get the waste out of the building. Sloping the pipe more and reducing the pipe diameter only goes so far.

I am very much in favor of proper and responsible use of water, but in my opinion, many of the places where there is a critical water shortage is because there are too many people living in an area that does not have the natural resources to support them. Instead of thinking "Should I live there?", too many are thinking, "You have to use less so I can live here."

I digress as well.

 

[semo]

Pedarrin2,

While I agree with your scenario, we are talking apples to oranges. In a stadium like you are discussing, the users are there for a few hours, there is a higher popluation per fixture, and during breaks, there will be times that almost all of the fixture units are flowing.

In a home (apartment complex), this will not be the case. There will be users not home, they will awake and use the shower at different times, eat at different times, etc, and the population per fixture is drastically less. This is why the equation based solutions work for residences.

But, in his original method, it only takes 10% of the units to utilize a fixture and they will be exceeding the design flow.

The safe way is to use the fixture method; but, in some situations (not this instance) oversizing the lines can be just as bad as undersizing. It could lead to stale water, sediment deposits (low velocity), etc which cause concerns.

In this situation, it really doesn't matter or cost much difference to install a 4-inch or a 6-inch line. The main cost is not in the pipe size; but, in the installation.

I also agree that you have to look at the local area when selecting the equation. There are different equations and some of them have different growth variables to use that can correspond with a specific region.

 

[davefitz]

1. exceeding the recommended waste pipe slope will cause the liquids to separate from the solids and the solids will then plug the waste pipe- do not use greater slopes to evacuate wastes.

2. nobody is suggesting we go back to outhouses or defecating in the fields at night. That would likely cause mass migration of all womenfolk to other nations that have plumbing.

3.modern alternative methods to treat wastewater include MBR membrane bioreactors followed by tertiary treatment and re-use of the permeate for fertilizing lawns or for flushing other toilets. Re-used water piping for flushing toilets may be on the way eventually. Grey water waste piping systems can also be used to minimize the size of the MBR.

"Whom the gods would destroy, they first make mad "

 

[bimr]

Getting back to the original problem.

You have to determine what building code the AHJ has adopted. Once you have determined the code requirements, then you can proceed to estimate the water supply requirement.

Attached is an outline of the methods that are used to estimate the water demand.

There is one note in the table that shows each Apartment, multiple family (per resident) is 50 GPD (note this is not peak demand).

USDA also is using a lower water supply value as well. "The value of 100 GPCD shown in Section 6 is a general value and may not be appropriate for many rural systems financed with WWD funds, so in the absence of reliable data, a value of 5000 gallons per EDU per month (approximately 67 GPCD or 167 GPD per EDU) should be used."