Skip to content

Thrust Restraint

 I am taking some time out from my campaign for world supremacy to have a rant about something that continues to annoy me. This is the subject of thrust restraint and how a relatively simple subject seems to be confusing to so many, especially structural engineers. I will start by giving an overly condescending explanation of where thrust comes from.

Myth: The force comes from the change in direction as the water goes around the bend. So it is momentum based and the direction of flow is important.

Fact: Almost all the thrust in almost all water pipelines we will be involved in comes from pressure, not momentum. Momentum is only significant if your pressure is very low and the velocity very high. So whether the water is flowing or not is irrelevant in most cases we will encounter. Usually the highest thrust load happens under test pressure when the water is stationary. Go to here and have a play. If you enter in our typical pressures and velocities, then you see that momentum thrust is a very small percentage of pressure thrust.

Myth: the water is acting on the inside of the pipe, so I use the internal diameter in my calculations.

Fact: For RRJ pipes, always use the outside diameter. Sure the water around the bend is acting on the inside, but it is also pushing on the annulus (the area of the ring bounded by the OD and the ID) so the total effect can be calculated using the OD rather than the ID.

Myth: I have to put thrust blocks on a welded PE or steel pipe.

Fact: Any pipeline with restrained joints (welded, flanged, Victaulic, Tyton-lok, etc.) does not need thrust blocks as the thrust is carried in the wall of the pipe. Your hose at home has plenty of bends but it doesn’t need thrust blocks. The only time the hose starts to whip around is when there is an open nozzle on the end, and this is from nozzle thrust.

Myth: I have an unequal tee so I calculate the thrust based on the main diameter, not the branch diameter.

Fact: Always use the branch diameter. It is the hole in the wall of the main line that is resulting in the unbalanced force.

So where does the thrust come from? At a dead end or valve or tee it is easy to envisage the pressure pushing on the disc and easy to work out (T=PA). But in a bend, it is not so easy to see but it is still due to acting on an area. If you look at a bend in plan, the length of the outside curve is more than the length of the inside curve. So the area on one side of the bend is more than on the other. This is where the net thrust comes from on a bend (at least that’s how I envisage it – the formula used to calculate it doesn’t include bend radius so it is based on a single mitre bend as the worst case scenario).

There is one special case with pipe with restrained joints and that is at an unrestrained end. If your PE or steel welded pipe ends in a RRJ at a different pipe material, then this end will need restraint. And the restraint load is not just from pressure thrust. Two additional effects that will be acting are thermal contraction and Poisson’s contraction. Either you need to accommodate this contraction or restrain it. With PE pipe, this contraction can be hundreds of millimetres so the first RRJ will just pop out (it has happened). So you will need an anchor wall on the end of the main. Thermal contraction happens when the operating temperature is less than the installed temperature. Usually this is in the order of 10 to 15 degrees. Poisson’s contraction happens when the pipe is under hoop pressure. The pipe wants to get longer circumferentially so it tries to shorten longitudinally. Fortunately, we have spreadsheets to calculate the loads and design the thrust walls, so email me if you want them. I may even get around to putting them in the water Toolbox.

One last rant is about geotechnical engineers. Any time you ask them for a horizontal bearing pressure to use in your thrust block design, they put so many factors of safety on it you end up designing your block for a soil with the consistency of custard. And it gets even more confusing when they don’t even tell you if it is for working stress design (allowable bearing pressure) or limit state design (ultimate bearing pressure). So treat what they tell you with caution and get a second opinion from someone in the know.


Fail-safe control valves

I have been disappointed that there hasn’t been a groundswell from you peasants, I mean, loyal followers, propelling me to the position of Dictator of PB worldwide. Therefore, I will have to instigate floggings and they will continue until moral improves and I am propelled to glory.

Those of you without Alzheimer’s will recall that last month I wrote about barometric loops and how they can be employed to prevent your line from draining when the pump turns off. I also mentioned another method and that is to put an actuated valve on the end of the line. The way this works is that when the pump wants to stop, a signal is sent to the terminal valve and it shuts, preventing the line from draining. When the pump starts again, the valve is opened and normal service is resumed. The timing of the valve opening and closing and also the time over which the valve closes are critical and a transient pressure model is usually required to ensure the control is correct. If it is not done right, then over-pressure from rapid valve closure or line draining from closing too slowly can occur. Having a VSD on the pump does help considerably as you can slow the flow right down before closing the valve. An emergency pressure relief system just upstream of the valve is also a good idea.

One big issue with a controlled valve is what happens when something goes wrong. If it fails to close, then the line drains and this may not be too big a problem if it is only once in a while. But if it fails to open, the system doesn’t function at all and your sewage pumping station will overflow or your water supply tank will run dry. So you have to make it “fail-safe”. Contrary to popular misconception, this does not mean something that will not fail. In accordance with Murphy’s Law, everything man-made has the potential to fail. What it means is that you accept that it can and will fail, but when it does, it fails in a safe manner. For a terminal valve, this usually means that it should fail open. A common way is to have a pneumatic actuator with air pressure to close and spring return to open. So if the site loses power or air pressure or communications with the pumping station, the valve will automatically open. Hence it has failed into a safe position.

As an aside, there actually was a Murphy and his law is often misquoted as “If anything can go wrong, it will”. He was a Captain Murphy in the US Air Force. After WW2, they were developing the first ejection seats and it was some poor sod’s job to be strapped into a rocket powered sled, accelerated along a rail track and slammed into a wall. This was to determine what acceleration a human body could withstand (this was in the days before crash test dummies and computer simulations). On the sled were 19 accelerometers to give a reading of the G forces. One day, something went wrong and all the rockets on the sled fired sending the unlucky enlisted man hurtling down the rails and into the wall at full speed. The guy was pulled out of the wreckage bloody, broken and barely alive. But on the upside, they thought that when the get the readings from the accelerometers, at least they will know the absolute upper limit a body can survive

But alas, there were two ways the meters could be mounted and all 19 had been mounted incorrectly and gave no reading. This is when Cpt Murphy coined his law which actually states “When there are two or more ways of doing something and one way will end in disaster, then that is the way it will be done.” So the way to defeat Murphy’s Law is just to make sure there is only one way of doing it. That is, make it fool-proof or fail-safe.

Barometric Loops

There is a new movie out called The Dictator. I haven’t seen it yet but for some reason, I am attracted to it. Puzzling… Anyway, adoring underlings, it is time for your monthly indoctrination. Keep reading; there is no subliminal messaging going on that will turn you into my obedient henchmen. Bwu-ha-ha-ha. (I wonder if there is such a word as henchwomen?)

What is a barometric loop and why would you need one? Basically, it is taking your pipe vertically up out of the ground to the required height and then back down again. At the top, there is vent pipe or some sort of air valve or vacuum breaker. The situation that you would uses one is where you have a pumped main that has a high point somewhere along it that is higher than the outlet point at the end. When the pump turns off, the line drains down from the high point. So you put in the barometric loop to artificially raise the termination level so the line doesn’t drain. The one shown is on two parallel sewage rising mains and is fondly referred to as “Pete Hopman’s Erection”. The fifth leg is a vent pipe.

“I don’t care if my line drains”, you say. In some situations it doesn’t matter if it does drain. But there are a number of disadvantages of turning back on a pump on a partially drained line. This is particularly the case with raw sewage lines. There could be one or more of the following:

  1. The pump see reduced head and goes off the end of its curve with possible cavitation damage or motor overload;
  2. It takes a long time to clear all the air from a line, so stable flow is not established
  3. The air pushed out of the line on restart is malodourous (stinky) if it is a sewage line.
  4. The air pocket is a breeding ground for sulphur reducing bacteria that excrete sulphuric acid that eats away the cement lining of the pipe causing rust and failure (it has happened).

So assuming you do want one, how do you design it? It has been the practice to just make sure the top of the loop is some nominal distance above the high point, usually 2 to 5 metres so the air valve at the high point will hopefully seal. But this simplistic approach can result in a design that does not prevent air pocket forming at the high point. What happens when a pump stops is that flow stops entering the pipe but it continues to leave the pipe until the negative pressure wave reaches the end. If the line is short and made of steel or DI, then this time is short and manageable. If the pipe is long and plastic, then this becomes a problem. If the volume that has left the system is too much, then the loop will not have sufficient water in it to prevent an air pocket at the high point. You may be tempted to make the diameter of the up pipe larger than the pipeline to get more volume in it. This is OK for potable water, but don’t do it with sewage or sediment carrying water as the low velocity will mean all the grit will settle out at the base of the pipe and eventually block the line.

On the down leg you need to aerate it and then possibly put in a de-aeration chamber to remove the entrained air. Another alternative is to put in a vortex drop pipe, but I won’t open that can of worms at this point.

So if you want to put in a barometric loop, please give it some careful thought as it may not solve the problem you were hoping it would.

PS  A barometric loop isn’t the only way of dealing with this problem. You can also use an actuated valve.

Mounting Pumps

After last issue, I got many suggestions, two of which weren’t obscene or impossible or both (I don’t think it is physically possible to do that to myself, Tom). But being the megalomaniacal dictator wannabe that I am, I shall ignore them and do what I want (like most politicians). So here goes.

How many times have you asked yourself “I wonder how thick the footing needs to be for this pump”? Chances are you have never asked yourself that question, but I won’t let that deter me from pontificating on the subject.

Small pumps up to a few kilowatts don’t need anything special – the one I have at home on my rainwater tank just sits happily on the slab without any bolts holding it down. However, big pumps certainly do need consideration. When you have big motors in the hundreds-of-kilowatts range, there is a lot of energy and even small out-of-balance forces can become big vibration problems with a consequent dramatic reduction in pump life between refits. There are a few basic rules of thumb to observe. These are:

  •  The mass of the pump concrete foundation should be five times the mass of the pump, motor and base plate being supported (to provide mass damping)
  •  The foundation should be 75 mm wider than the base plate, all around, up to 375 KW and 150 mm above 375 KW
  • Imaginary lines, extended downward 30 degrees to either side of a vertical through the pump shaft, should pass through the bottom of the foundation and not the sides
  • Ensure there is a good quality flowable grout between the pump frame and the plinth. Epoxy grout is recommended for pumps above 150 kW
  • Pump-motor frames must be levelled using wedges and shims, not levelling nuts. Once the grout is set, the hold-down bolts then need to be tensioned to the foundation (which is not possible if a levelling nut is
    installed – it just tensions the bolt between the two nuts)

Pump slab stiffness – As a bare minimum, the slab should be stiffer and stronger than the pump frame. If it is on a reactive site, as another minimum it should be as stiff as a masonry structure slab under AS2870.

If the pump is up on a structure, then using mass damping is not really possible. In this case, make the support structure very stiff with a natural frequency at least 1.75 times the primary excitation frequency of the pump.