Potable Water Reservoir Sizing Criteria 5

I don't know of any rule of thumb, I would think you would have to actually route the recharge/inflow versus the outflow, with diurnal peaks. The volume is the integral of the deficit, then check that it can store without overtopping during lag demand.

Remember: The Chinese ideogram for “crisis” is comprised of the characters for “danger” and “opportunity.”
-Steve

you will also need to consider fire flows and emergency needs and balance this with your proposed pumping / transmission system. Also, consider redundancy so that you can take wells, pumps, tanks or transmission mains out of service for maintenance or repairs and still provide water for your distribution system.

I think I saw once that EPA has done some looking at how reservoir/storage aspects, along with many other things, affect water quality. While I don't know what if anything they say specifically about "sizing", this agency could conceivably also be a resource for your consideration.

Here in Texas, the state mandates a certain minimum storage capacity per connection. Don't have a clue if it's conservative or not.

Pinwards…

You have received some good tips. Let me attempt to tie all this together. Some of what follows you probably already know, but I am including it to be as complete as possible.

The basic storage components in a water tank are: [1] Daily Operational Storage (DOS), [2] Fire Storage (FS), and [3] Emergency Storage (ES). Echoing cvg, depending on your water system, you may or may not need all or part of each component in your tank.

The purpose of DOS (besides running old computers!) is to supply water when demand temporarily exceeds supply due to diurnal fluctuations. Most water systems are designed around the Maximum Day, which is usually defined as the day each year with the highest actual or estimated water usage, though some people use the 95th or 98th percentile day. The 24-hour average flow on a Maximum Day is called--obviously--Maximum Day Demand (MDD). In most municipal systems, demand will exceed the 24-hour average flow during the day and will be less than the average at night. There are exceptions, such as in small systems in arid regions serving large nighttime irrigation demands (e.g. a golf course). You will obviously need to know the diurnal characteristics of your system. As lha said, the Daily Operational Storage can be integrated from the diurnal demand curve.

In one system I worked with years ago, we simplified the diurnal hydrograph--based on lots of data--to a simple rectangular demand curve: Maximum Day Demand-Daytime (MDD-D) was assumed to be 1.5 x MDD for 14 hours and Maximum Day Demand-Nighttime (MDD-N) was assumed to be 0.3 x MDD for 10 hours. Assuming our supply was equal to MDD, the amount of DOS required was then (1.5 - 1.0) x MDD x 14 hrs. This same volume of water would be replaced at night, that is (1.0 - 0.3) x MDD x 10 hrs, so that at the end of a Maximum Day, the DOS was refilled. During the rest of the year, the DOS would be refilled in less than 10 hours. In reality, we sized our supply facilities to be at least MDD plus an amount needed to refill the Fire Storage component in 72 hours. Some of our booster pumping stations were actually sized to take advantage of time-of-use (TOU) electricity rates. Storage tanks supplied with a TOU booster pumping station based on a 10-hour use period required DOS that was 24/10 or 2.4x as large as the above calculation. These calculations are based on a simple system with one storage tank supplied by one booster pumping station or well, with no upper or lower pressure zone to feed.

Many water systems here in Central California are supplied mostly or entirely from wells. The City of Fresno, for instance, serves about 500,000 people with about 230 wells, and there is very little storage. The wells provide such a level of redunancy that storage is not really needed.

The purpose of Fire Storage is to store the volume of water required to meet (usually) the largest fire flow in the system. For instance, a system with a residential zone requiring 1250 gpm for 2 hours and a commercial zone requiring 3000 gpm for 3 hours needs only store 3000 gpm x 3 hrs = 540,000 gallons. You do not need to add an additional 150,000 gallons for the residential fire flow. Fire flow requirements are typically determined by the local fire protection agency. If you have two tanks in a pressure zone, you can usually split the fire flow between the two tanks, rather than allocating the entire required fire flow to each tank separately. However, if your tanks are widely separated in the same zone, it may be wise to size each tank for the maximum fire flow, or at least (using the above example) providing the appropriate FS in each tank based on the nearby fire flow requirement.

The purpose of Emergency Storage is to provide a reserve in the case of supply failure (mechanical, power outage, etc.). This is the storage component where you have the most leeway, and thus it requires the most judgement. If you have a tank in an urban setting supplied by a booster pumping station with an alternative power source (e.g. dual drive, or engine generator), you may only need 2 hours of MDD. If you have a remote tank with a well with only an electric motor to drive it, you may want 5+ hours of MDD. I know of one system where the client wants 12 hours of MDD.

Other things to keep in mind, if your pressure zone is sometimes required to backfeed a lower zone, you should increase storage. How much depends on operational factors I won't try to address here. If your zone can be backfed from an upper pressure zone, you could theorectically reduce storage, but I don't recommend it. I prefer to assume maximum demand and minimum supply in sizing facilities. A system that is fed by pumping out of a ground level tank working in parallel with direct feed from wells and/or other booster pumping stations can often get by with reduced storage. It all depends on the level of redundancy in your system. I have one system I am working on where the ground level tank will store a fraction of DOS, the full FS, and a 2-hour ES.

Addressing JStephen's comment, I have seen similar requirements. In these cases, I check the storage requirement two ways: their method and my method. I haven't had to deal with this in years, so I can't remember whether it is conservative or not. Regardless, the diurnal hydrograph is best way to determine DOS, the fire flow requirements are the best way to determine FS, and a fortune cookie is the best way to determine ES

I hope by now that I have everyone thoroughly confused. I have sized more than 50 water tanks over the years, ranging from 0.2 MG to 3.0 MG. Most were sized as I explained above. A few required different criteria based on the actual configuration and operation of the system.

Also be aware of water quality issues related to storage tanks. Systems with low diurnal demands and large fire flow requirements may not turn the water over in the tank quickly enough. You may want to add a recirculating pump and chlorinate at the tank. I even saw a tank at Camp Pendleton that has internal walls to force the water through the tank in a zig-zag fashion. If you have an area that is growing, consider phasing storage with two or more tanks.

Fred

Haestad have a book titled Advanced Water Distribution Modelling that is worht investing in. You can then use Epanet (free software) to model the level in your reservoir over a three day or longer period with the demands characterised for diurnal flow patterns. You can then look at the kinematic behaviour of chlorination in pipelines, hold in storage where ageing may be a problem etc.

Your model can be modified for summer/winter conditions where demands change. It maybe you take a reservoir off line to prevent ageing or you have a regime in place that allows the emptying and filling a reservoir. You can also model future conditions with increased surface roughness in pipes, changes in demand for population growth or reduction due to introduction of a recycle water scheme etc etc.

In Pennsylvania, finished water storage is to be sized to meet peak hourly demands with consideration to fire flow demands (the basic sizing computations of which are explained fel3's post. Also, I think there was an article a few months ago in "Opflow" that explained basic sizing computations).


One day's storage is the recommended minimum, which my colleagues and I interpret as ave daily demand.

Ten States call for:

1. Fire flow requirements established by the state insurance services office.

2. Minimum storage capacity of the average daily consumption.

We normally go with 2 days of storage. If you go any longer, you will tend to get water quality deterioration.

AWWA Water Distribution Handbook has an excellent discussion of the various components that go into storage sizing. The AWWA also has several monographs on Fire flow and storage that are useful. These contain references to ISO standards, which are also helpful. One of the most detailed state standards is that of Texas, which addresses various combinations of pump sizes, elevated and pumped storage for different populations.

Several more questions on the same issue :

Pinwards asked for sizing a potable water storage reservoir, In my side of the world this could mean any of the following :
A- Storage reservoir at the source (source being a Treatment plant or well filed etc..)

B- Intermediate Storage reservoir without direct supply to consumers.(for example at intermediate pumping station)

C- Storage reservoir with direct supply to consumers.

D-Storage reservoir at the basement of each building (used mainly for towers).

E-Storage Tank at the roof top of each building.

Fel3 provided a really helpful post. I assumed that he was talking about "C" and the system we are talking about is direct pumping to the reservoir and pressurized gravity network from the tank to the consumers tank on the roof of each building. Now , my questions are ,

1- How do you size "A" and "B" ?
2- We would normally use 2 X MDD(maxdaydemand) for "D", what is the effect on "D" if one single underground reservoir is used for a community instead of a big tank at the basement of each building ? will that convert it to a "B" and fel3 calcs apply ? since this an underground tank then water will be pumped to consumers tank on the roof of each building.
3- same question as in 2, but what if 24 hours pumping is provided and roof tanks are not used ?
4- would you include the losses in the sizing of any tank ?

Best Regards.

Also check this link to the Washington DOH Water System Design Manual and go to Chapter 9.

Thanks mell, the WA DOH is actually the criteria we ended up using to justify the design.

Waseem, our case was C, we pump directly from our water treatment facility up to an elevated storage tank that directly gravity feeds the distribution system to individual homes.

Thanks to everyone else for the great feedback as well.

P

Use at least 30 times the rated capacity (in gpm) of the water supply pumps. This means that you are giving the pump sufficient time to rest before restarting. Frequent start and stopping of pump may damage the motors.