Why centrifugal pumps break shafts and how to prevent it
Though most people who have operated centrifugal pumps are familiar with the problem of broken shafts, few understand the reasons behind why the shafts break. Thus the problem continues...
But it doesn’t have to. Learning why the shafts break can help you prevent the damage from happening.
First, pump shafts break because: they twist, they bend, and they they wear out.
Now consider the possibilities:
- Calculate the diameter of shaft required to carry the maximum horsepower that any particular centrifugal pump could draw. Then add in a generous safety factor for shock loading – which won’t really happen unless you have a jammed pump. You will most likely come up with a shaft diameter that is about one-half of that used in the pump! Most pump shafts are designed on some other basis. You may also deduce that you have never broken a pump shaft by simply twisting it until it breaks off.
- You can bend the shaft until it breaks – though not inside any centrifugal pump we’ve ever seen. There isn’t enough room to bend the shaft that far. But the key here is “inside the pump.” It’s possible to misalign the pump and driver and break the shaft on either side of the outboard bearing. Though this has happened, it’s more likely that the bearings fail before the shaft. So if you have shaft breakage near the bearings, misalignment is probably the culprit.
- Shafts can wear out. This is the cause of the vast majority of centrifugal pump shaft breakage. Conventional wisdom notes that stresses far below the ultimate strength of the shaft – if repeated enough – will cause the shaft to break. But not always. If the stress is low enough, it can be repeated infinitely without causing damage.
For example, you could bend a pump shaft .005” at the center 3,600 times per minute, 24 hours per day without causing it to break. However, if we increase it to .010”, it might break in one week. That’s because you exceeded the shaft’s fatigue limit.
Most centrifugal pump shafts are designed to keep deflections less than the fatigue limit. Design engineers make assumptions to do this – much like how the person who designed your car assumes you wouldn’t drive with a flat tire.
So the question become, which of the design engineers’ assumptions are we violating when the pump shaft breaks?
B1 and B2 are the two bearings supporting the shaft. W1 is the weight of the impeller and shaft and U is the unbalance of the whole assembly. T is a quantity called unbalanced radial thrust. Each one of these forces is acting on a shaft to reverse bend it at every revolution. The design engineer can figure the weights of the shaft and the impeller. Because they don’t change, he can design for these.
Unbalance shouldn’t be an issue because most large impellers are balanced dynamically by manufacturer. But what if a few large air bubbles occasionally pass through the pump? Remember the rest of the impeller is full of liquid. It’s the reverse of a car tire with a misplaced weight on the rim – but there’s just as much unbalance.
Most think that all centrifugal must handle some air. But there’s no way to calculate or estimate how much unbalance a particular amount of air will cause because you don’t know how it will be distributed around the impeller. So if you want a long shaft life, make sure liquid is all your centrifugal pump handles.
The other factor is the unbalanced radial thrust. Take a look at the end view of an impeller and pump volute below. The arrows pointing to the impeller denote the pressure in the volute acting on the impeller.
But the arrows are different sizes. Some will be balanced off by others, but it won’t come even. The result is an unbalanced radial thrust, represented by the big arrow marked “T”. Note that this force doesn’t always act in the same direction and it gets larger when the pump is operated at near shut off and wide open.
This is a curve showing how unbalanced radial thrust is at maximum at shut-off, goes down to a maximum at the pump’s best efficiency point, and then goes up again at wide open.
It makes a big difference. Let’s use a pump in a paper mill as an example. We recently determined that the amount of unbalanced radial thrust on a 24” double suction centrifugal pump – being operated at half of its capacity – totaled more than 1.5 tons.
The lesson here is simple: Don’t run centrifugal pumps at near shut-off or wide open unless you like buying shafts.
And finally, a bit more about shafts. Four factors affect a shaft’s ability to stave off fatigue failure:
- Its size. The distance between the bearings is more important than the shaft’s diameter when it comes to overall stiffness.
- Material. Not much can be gained here. The fatigue limit of common bar stock is the same as the fanciest shafting you can purchase.
- Surface finish and stress concentration points. Shafts should be surface ground, threads undercut and the keyways should have rounded bottoms.
- Wet or dry. Shafts that allow liquid to reach the points of stress concentration fail more quickly than one that remains dry.