Maximum Flow Rates for PVC Pipes

[mas985]

I have long been skeptical of the flow limits placed upon PVC pipes. Many web sites would post arbitrary limits ranging from 5 ft/sec up to 10 ft/sec indicating that it is not a good idea to exceed these limits but none would explain why or what would happen if you did exceed these limits. I am a big advocate for rules of thumb but I also like to know where they come so I have some trust in them.

There are many often sited reasons for velocity limits in PVC pipe including water abrasion wearing down the inside of the pipe. However, PVC is one of the most abrasion resistant materials and pools generally have small amounts of particulate mater that make their way through the plumbing to the filter. After, the filter there is even less material so return lines should not have any velocity limits under this assumption. Either way it would take an extremely long time to wear down PVC pipe with the average pool water and given that I never heard of PVC plumbing failing because it was worn out, I decided to disregard this reason.

Another reason for the velocity limits is due to stress fractures. Turning the pump on and off over a long period of time will stress the PVC pipe. This appears to be the most likely origin of the velocity limits and being a self-employed engineer with clearly too much time on my hands, I decided to investigate this a bit further. Here is what I found.

Repeated stress cycling of PVC pipe will eventually cause failures according to this paper. The cycles to failure is directly dependent on the average pressure of the pipe and amplitude of surge pressure in the pipe. Surge pressure occurs whenever there is a significant change in the pressure within the pipe over a short period of time (e.g. pump or valve turned on). This creates a pressure wave in the pipe that can be several orders of magnitude greater than the average pressure. Several charts are shown in the Uni-Bell paper which would indicate that most failures occur at very high pressures or large cycle times. So using the formulas in Appendix A, the pressure rating of the pipe can be determined.

The specifications and pressure ratings for schedule 40 PVC piping can also be found here. The pressure ratings shown in the table are for 73 degree water and in the table below shows the correction factors for higher temp water. Again, these pressure limits are designed to lengthen the life of PVC pipe and to be resistant to stress fractures. Also on this site is a reference to the 5 ft/sec velocity limit as to avoid any hydraulic shock. Also on the same page is a method to calculate the surge pressure which is similar to the Uni-Bell paper.

If we assume that the pressure ratings of the pipe are not to be exceeded by a worst case surge pressure, the maximum velocity within the pipe can be determined. This of course assumes an instantaneous change in water flow. It is this transient pressure that I believe has been used to determine limits on velocity. So by cycling a pool pump on and off enough times can create stress fracture within the pipe. If the pool pump was left on all of the time, I do not believe that velocity would not be an issue at all.

Using the pressure limits for schedule 40 piping and assuming worst case water temperature of 105 degrees, the table below shows what the equivalent water velocity and GPM would need to be in order to create a pressure wave from pump cycling that exceeds the pressure rating of the PVC pipe.

The range of water velocities is within what is seen on many web sites which validates the premise that cycle stress is the origin of these limits. However, the limits are based on a very large number of cycles to create a failure. While this method seems sound at first, there are several assumptions that are flawed. First, the required cycles to reach these limits would never within the lifetime of an average swimming pool. Second, the surge pressure requires an instantaneous change in velocity. Starting or stopping a pool pump has a ramp up and down affect on water velocity and is not instantaneous. Also, if surge pressure was as high as the pressure rating for the pipe, it would probably destroy the filter which is typically rated much lower.

After, reviewing the velocity limits, it became clear that many pools exceed these limits including mine. Water velocities in my 2" return plumbing section would definitely exceed the 80 GPM limit by almost 10 GPM. Looking at my filter when turning the pump on reveals that the pressure goes to about 30 PSI before settling to 19 PSI so there is definitely a pressure wave but not a very big one and no where near pressure rating of the pipe.

There are many pools with 2.5" pumps on similar plumbing configurations which would raise the flow to over 100 GPM in 2" pipe so I believe that these limits are exceeded by pools every day without much consequence. Also, for colder water, the limits are much higher so it is even less likely that the limits are exceeded.

It is my belief that the pool industry took recommendations from water distribution system, once again, and applied them incorrectly to swimming pool plumbing systems. So although it may be a good to design pool plumbing such that the velocity limits are not exceed, I believe that too much emphasis is placed on these somewhat arbitrary limits. In reality, a pool's plumbing will never exceed enough cycles and high enough pressures to create pipe fatigue and failure.

I would be happy to hear about any cases where a stress fracture actually occurred but so far, I haven’t heard of any. Most plumbing failures I hear about occur due to poorly welded joints and damage to the pipe from external factors (e.g. shovel).

 

[chem geek]

Very nice analysis.

 

[mas985]

Thanks,

I do my best trying to keep up with you.

 

[JasonLion]

What about energy efficiency? I think dynamic head goes as an exponential function of flow rate. At some point you are going to start getting dramatic reductions in efficiency. Some quick calculations suggest that this will tend to start happening at more then triple the flow rate limits in your table though. Assuming I am doing my math correctly, there is a max practical flow rate to avoid a dramatic loss of energy efficiency, though it is probably much higher than the currently published limits.

 

[mas985]

Head is proportional to the square (Head ~ C * GPM^1.852 Hazen-Williams eq.) of the flow rate. So yes head loss goes up with flow rate and will continue to due so as long as there is a pump able to deliver the power. There is no theoretical limit to the increase but you need higher HP pumps (positive head) to get there. There is a limit as to availability of high HP pumps so you probably won't find many over 3 HP. So for a given plumbing system, there is a practical limit of the flow rate but it has to more do with equipment availability and not the physics.

As for efficiency, it is possible to operate at the best efficiency point of a particular pump and at the same time exceed the limits shown in the table. It is a matter of plumbing design and pump choice. In fact, for my pool, I operate very close to the best efficiency point of my pump and get about 88 GPM which is over the limit of the so called maximum velocity. You could probably find a pump and plumbing system combination for each pipe size that would have this same characteristic. This post shows a head curve with different HP pumps and you can see how the best efficiency point head and GPM increases with increasing HP.

The main point here is that many pools operate above the quoted velocity limits without much problem. Some operate near the pumps best efficiency point and some do not.

 

[JasonLion]

Alright, I follow that. I agree, the currently published max flow rate tables have no direct bearing on typical swimming pool applications. However, it would be nice to have some guide to what size pipe should ideally be used in any given situation. With the old, mistaken, understanding of the max GPM table I could say: You are aiming for 70 GPM so (look up in table) you should use 2" pipe. Does your new model invalidate that usage of the table and if so what should we use instead? Obviously no one is going to spec a single 1" pipe with a 3 HP pump for a swiming pool even if the pipe isn't going to crack or burst in that application and I could find a pump that could get 70 GPM through that pipe.

 

[mas985]

I have a few guidelines in the sticky post, but the optimum choice for both pipes and pumps depends on many things. Here is what I outlined as a resonable process (maybe I need to elaborate a bit more) in the sticky.

First choose a pump based upon the maximum flow rates required which would depend on turnover but also spa jets, in floor cleaners, etc.

Once you know the maximum flow you need, you can use the best efficiency point for each HP pump to choose the closest pump and the associated ideal head loss.

Since it is easier to add head loss to plumbing than remove it, you generally want large pipes but you also don't want to spend money you don't have to.

So once you have your ideal head loss, then you can work out what the minimum diameter pipe would be required to get that head loss based upon the head curve of the chosen pump and plumbing curves shown on top. However, this is dependent on distance to the pad and number of fittings in the plumbing.

Each pump manufacture and model has a different head curve so it is difficult to come up with absolute rules of thumb which is why I tried to outline a process to go through.

But in general, here is what I would use as very general guidelines for the new generation of high efficiency pumps.

For pumps less than or equal to a 1.5 HP full rated or 2.0 HP up rated:
One 2.5" suction line or two+ 2.0" suction lines
One 2.0" return line or two+ 1.5" suction lines

For pumps greater than or equal to a 2.0 HP full rated or 2.5 HP up rated:
One 3.0" suction line or two 2.5" suction lines or three+ 2.0" suction lines
One 2.5" return line or two 2.0" suction lines or three+ 1.5" return lines

These plumbing sizes would also work for lower efficiency pumps but be a little oversized. In all cases, I would stay away from single 1.5" lines no matter what since most pumps sold today will over power it and create a lot of head loss.

Also, with very long runs, you may want to go up a size.

Eventually, I may add a few more tables to the sticky to make easier for someone to just look up the best pump size for a given flow rate and recommend plumbing sizes for different run lengths. This just gets into a lot of combinations that need to be covered given the variability in plumbing designs. I may just start with the pumps I have modeled already which are the Northstar, Whisperflo and Intelliflo.

 

[DaveNJ]

mas985,
Would like your comments on the following, pick your brain so to speak.
Thinking of adding solar on roof of garage( in the process of building). I have 2.5" pipe, 2 skimmers, main drain, 4 pool returns, 2 swimout jets, 2 step jets, 6 spa jets. What would be the best configuration for connecting solar i.e between heater, return to spa or pool. Also pipe run to solar 2" or 2.5". Pump pad is 10' from pool and spa, 15' from garage with a rise of 13'. I am thinking of using 6 4x12 panels, thats all I can fit. I don't use a solar cover. Plants, seats and space made me stop using it. Only looking for a 4-5 deg change. We use the pool mainly late in the day.
Thanks
Dave

 

[mas985]

Generally, you want to plumb solar before the valve which selects pool or spa. That way you can heat either. I cut my solar in between the filter and heater but that was mainly because the flow throught the heater must be maintained even when the heat is off. Otherwise, the internal blower kicks in and runs all the time. There are other ways to plumb in solar but the most common is between the filter and heater which also gives you the highest pressure to prime the panels and you could use both the solar and heater if wanted to, assuming the controller allows it.

As for pipe size, obviously 2.5" would produce less head loss but I suspect the panels would dominate in head loss anyway. 2.5" has about 1/3 the head loss as 2.0" pipe for a given length and GPM. 120' of 2.0" pipe, I double the pipe length to account for fittings, adds about 5 feet of head @ 100 GPM while 60' of 2.5" add about 2' or 3' difference which isn't much since it is about 1 PSI. The panels are likely to add 3-5 PSI on their own. One thing you could do is use 2.5" for the horizontal and 2.0" up the side of the garage for asthetics.

One more thing, if you run on low speed most of the time, 2.0" is more than enough.

 

[DaveNJ]

Thanks for the info and suggestions. I had read your stickies and links and was concerned about the pressure from my configuration. Again thankyou
Dave

 

[Poolsean]

Mas, great info.
I was thinking about this and came to an asumption as to where the max velocity came from. In the good old days, pools were plumbed with copper pipes and in some case galvanized pipes. Copper, I hope you will agree, definately does have a max velocity at which the copper will strip away with excessive velocity.
Again, this is an asumption, but it seems to make sense. And we, the pool industry, seem to use what's been generally accepted for years.
I do agree that PVC is not influenced by excessive flow, other than pressure build ups (increase head pressure) but there's little likelihood of damage to the pipe itself from excessive flow.

 

[mas985]

Sean,

I just realized that one of the references that I had in my original post requires a password to get to so I have included it here:

This is from the Harvel site which indicates the origin of the velocity limits for pipe.

HYDRAULIC SHOCK


Hydraulic shock is the term used to describe the momentary pressure rise in a piping system which results when the liquid is started or stopped quickly. This pressure rise is caused by the momentum of the fluid; therefore, the pressure rise increases with the velocity of the liquid, the length of the system from the fluid source, or with an increase in the speed with which it is started or stopped. Examples of situations where hydraulic shock can occur are valves, which are opened or closed quickly, or pumps, which start with an empty discharge line. Hydraulic shock can even occur if a high speed wall of liquid (as from a starting pump) hits a sudden change of direction in the piping, such as an elbow.The pressure rise created by the hydraulic shock effect is added to whatever fluid pressure exists in the piping system and, although only momentary, this shock load can be enough to burst pipe and break fittings or valves.

And they go on to explain how to "avoid" it.

Proper design when laying out a piping system will eliminate the possibility of hydraulic shock damage.

The following suggestions will help in avoiding problems:

1. In a plastic piping system, a fluid velocity not exceeding 5 ft./sec. will minimize hydraulic shock effects, even with quickly closing valves, such as solenoid valves.

2. Using actuated valves which have a specific closing time will eliminate the possibility of someone inadvertently slamming a valve open or closed too quickly. With pneumatic and air- spring actuators, it may be necessary to place a valve in the air line to slow down the valve operation cycle.

3. If possible, when starting a pump, partially close the valve in the discharge line to minimize the volume of liquid, which is rapidly accelerating through the system. Once the pump is up to speed and the line completely full, the valve may be opened.

4. A check valve installed near a pump in the discharge line will keep the line full and help prevent excessive hydraulic shock during pump start- up.

Copper is much stronger for hydraulic shock so you can actually have higher velocities than PVC. It is a soft material so it may be more prone to abrasions than PVC but I have not seen any data that would indicate so and I would really be surprised if any metal pipe could be worn out from inside out due just to water velocity. Corrosion is another issue.

Here is another paper which covers hydraulic shock.

 

[chatcher]

From a practical standpoint, the effects of "undersized" piping on the suction side are considerably different from those on the discharge side. On the discharge side the problem of "too much pump" for a given size pipe is self-correcting. Smaller, longer pipe, with more fittings and directional changes (more friction) increases dynamic head, reducing flow rate, which actually increases the rpm of the pump and reduces the electrical current flowing in the motor (thereby using less electricity and costing less money to run). You can turn a 2HP pump into a pretty good 1HP pump by just throttling down on the discharge side (though it won't be quite as efficient as a properly designed 1HP pump). The operation of centrifugal pumps is counter-intuitive to many people, who assume that TRYING to pump water against a restriction is harder work than ACTUALLY pumping water through an open pipe. The opposite is true, since it is less work for the pump to spin the same water around and around in a circle than it is to accelerate new water to the same speed.

On the suction side, too much head loss will result in cavitation which can actually destroy a pump. Those bubbles created when a pump cavitates are water vapor (steam) caused by too little pressure (too much of a vacuum) at the pump inlet, and as they move from the low pressure side of the pump where they formed to the high pressure side they violently implode as they condense back into liquid water, and hammer away at the impeller, eroding it over time. That rattling sound you hear when your pump cavitates is the sound of the pump self-destructing. (Note there is a big difference between cavitation and sucking air in through leaky pipes or pump covers.)

All this is why any valve you use to reduce the flow through your pump must be in the discharge line, and your suction piping should have as little restriction as possible.