Shaft Failure

Aw, that's so sad. Have you replaced it, Mildred?

How long was it in service? Nothing lasts forever.

What is the dark stuff? Corrosion products from service?

If that tip was friction welded to the rest of the shaft, then it was a defective weld; if it is machined, then it lacks a radius that "blends" the two diameters of the shaft.

The twisting of the key way suggests to me that this was not the root cause of the failure but a result of the failure. A lot of force went into bending that shaft before it snapped.

Looks like the shaft was in a bind and twisted causing the shear. Again its apparent in the keyway twist. I am a great believer in shear pins.

If that's what drove it then the rest of the manglatron must be in one hell of a mess!

Ya, I'd bet something is bent. That's probably what's causing the bending fatigue stress.

And what ever event bent the shaft, also caused the keyway initial crack failure. Something stupid happened.

Just guessing, as the OP has not shown the rest of the machine.

bad weld, the keyway shows no unusual stress

If we concentrate only on this detail we can notice:- radius at the grove root is very small almost zero.- fatigue marks are concentric and indicate a start at the corner mentioned aboveI think the sharp corner generated a very high stress due to the torque and on top came a limited bending which brought the stress level over the acceptable limit.It is also possible that the crack did progress even without any bending due to the way the torque is transmitted via the side contact between key and shaft grove wall.The presence of the 2 threaded holes accelerated the crack progress. The marks run almost normal to the hole circular wall.If the part would be cleaned the marks would be easier to analyze.

Is this the driven part of the shaft or the driving part? I ask only to raise the question given the other evidence.

By the pattern it does look as though there has been a bending action causing it to fail It would be advisable to check the alignment before comissoning a new one

Thanks for the comment.

following more details on the issues:

shaft is runing at 30 RPM and we are using vibration analisys to monitor the condition of the unit which didn't reveal any sign or indication of the failure . As discussed with the condition monitoring team it is hard to identify any issue due to the low running speed.

I did PT as weel on the broken shaft and it shows crack propagating from both side of the key way while the shaft is runing only in one direction. I presume it is a sign of vibration/ looseness.

attached here more picture and the coupling drawing with the shaft in.

You don't mention the environment that this is operating in. It could be that the shaft has suffered crevice corrosion, stress corrosion cracking, inter-granular corrosion, Hydrogen embrittlement, or any of a few others that specifically attack Stainless Steel. Stress corrosion cracking and crevice corrosion are major causes of the failure of the Austenitic SSs such as 304 and 316.

For instance, a simple rubber band around a piece of 316 SS shaft in seawater can cut it in half in time. household bleach at normally around 5% NaOCl concentration is particularly aggressive to SS. Add high temperatures to the mix, and stress corrosion cracking is a likely contender.

It is generally recognised that chloride concentrations above 1% will cause corrosion of 316 SS.

SS relies on the formation of a passive layer in order to prevent corrosion. Crevice corrosion takes place predominantly in areas where localised stagnation of the fluid takes place and where the passive layer is prevented from reforming when disturbed, and where the ability to clean the surface of deposits is limited. In your pic. the position of the break is a good indication of lack of circulation and also inhibition of passivation due to close contact with other surfaces and the inability to clean said surfaces.

Whilst 316 SS is generally immune to Hydrogen embrittlement, it can occur with fatal consequences in circumstances such as where improperly controlled cathodic protection systems are employed, in Hydrogen rich environments, or where the solution is low pH. Again, your break has occurred right where this process works best, where the Hydrogen reacts with the Carbon in the steel to form Methane gas which is unable to escape the void and builds up sufficient pressure to form cracking of the metal. If the shaft has ever been repaired using incorrect welding rods, the chance of failure increases by a high magnitude.

Once weakening of the base has occurred, then other stresses can finish the break.

You might consider moving to a duplex type of SS such as 2304 (EN # 1.4362) or 2205 (1.4462) if operating temperatures are above 50°C and/or the solution is brine in nature. These type offer better stress corrosion protection than does 316 (1.44**).

Whilst 316 SS is generally immune to Hydrogen embrittlement.........

I will have to disagree with your statement here, marinel (an ultra high strength copper nickel alloy) is the preferred material used in sub sea and splash zone areas. Iconnel, some super duplex stainless steels and several high strength cupro-nickel alloys can be used, but on fastenings, etc., on oil and gas rigs marinel is becoming the preferred material. 90/10 cupro nickel sheathing has been used to protected the legs on many North Sea oil rigs against hydrogen embrittlement

I fail to see why the use of other alloys in those instances causes you to disagree with my comment. The mere fact that 316 is unsuitable for prolonged immersion in seawater would preclude its use for the purposes that you cite without any regard to Hydrogen embrittlement concerns

To assist you, I have done a search of the web for reliable sites, and draw your attention to a few links that would appear to confirm my comment on 316's general immunity to Hydrogen embrittlement.

this link

This link

This link

This link discusses the effects of Cathodic polarisation on the Hydrogen embrittlement resistivity of 316, which again concurs with what I said in my post.

My apologies spades, I an afraid I jumped in, boots and all, due referencing all things to seawater. I guess I am not thinking very logically at this point in time............my wife of fifty years passed away on Thursday and I was "trolling" around to try and keep my mind occupied...................My apologies once again

My sincere condolences for your loss, I can't begin to imagine the pain that you are feeling.

With regard to the topic:- I have no problem with being questioned regarding my post. If I couldn't justify what I had said, then it is quite possible that I was wrong.

My condolence for your loss. My or my wife will be facing that situation the future.

How long is the agitator shaft? Is there an outboard bearing that supports the shaft?

Brand new material will not be straight, so before installing a replacement, check it for runout. A good shaft shop should be able to heat straighten it to less then .001" (.025mm) over it's length.

I see it's a tapered shaft end, and the coupling washer pulls the taper tight.

Is there any creeping of metal under the bolt heads, or the washer? Your indication of the crack location is where the 2 coupling bolts end. This would make sense as this would be where the max stretching and hence stress would be.

A shaft with a small amount of runout, driven this way, with a support bearing that restricts the shaft from running it's natural eccentric would certainly induce cyclical fatigue stress failure.

How many hours of run time did it take for this to fail? Just curious.

1- I still think it would be good to totally clean the surfaces and show the resulting pictures.

2- You last explanation brings some more light on what could have happened. But first it should be of great importance to know how was the contact between the 2 conical surfaces. Where was it ? toward the smaller or toward the bigger diameter ? I bet it was more on the smaller side and the shaft was like a cantilever beam free in the coupling for part of it. The shaft broke where the diameter was small enough to lead to the fatigue limit. Most probably the angular tolerance was not matched to the deformation of the 2 parts and the pulling force of the 2 bolts. Such conical assemblies are very tricky and require a lot of understanding on how they work and how they fail.

3- Bending could come from different sources:

- the agitator can generate different axial forces at different radii which can generate a variable bending moment. It would be good to have an info about the agitator geometry and about the fluid/paste in the reactor.

- It is possible that the agitator inertia and the shaft compliance build an oscillating system and superpose this to the one direction movement. this could be the reason the 2 cracks have different sizes since the constant torque will increase the crack developing speed in a preferential direction. This should be checked computing the shaft compliance and the agitator inertia.

- if the medium is a paste it can build on the agitator an eccentric mass which leads to a centrifugal (sorry centripetal! oh make your choice any way which ever the force name is it pulls in radial direction on the shaft direction from center to periphery) force and this generates an important bending moment. If the shaft would have been in the right position this moment would have loaded the region before the key slot and the key would have had to transmit ONLY torque what it is thought for.

May be if we get more informations we can come to the right explanation till then all are suppositions.

It would appear that the break occurred at the bottom of the tapped holes. If the bolts were over-torqued and created a stress point there, combined with higher torque being required by the mixing blades/paddles and/or if the 316 shaft properties were not up to snuff or the stress analysis done at that point didn't properly account for the material removal by the 3 holes, a combination of some of these would likely lead to failure at the root of the bolt holes since that is the first location to receive the applied torque from the mixing load.

I would check the chemistry of the shaft and ascertain whether someone had re-installed the shaft/coupling connection and whether the recommended torque was applied or if someone just snugged it up using "a helper".

If the load has a high viscosity range, was there a temperature fluctuation just preceding the break that could have drastically increased the applied torque?

Dear Mr.augustin,

Spot 1 has developed breakage, as first failure, and oil dirt entered have become utter black, thus most weekest Spot, which has developed first failure zone, may be due to a minute crack.

Spot 4 has followed to crack further and disintegrated. Spot 2 and Spot 3 were holding intact and transmitted power and then failed. Among these two spots, No. 2 has failed and then No.3

An ultrasonic testing could have helped to identify and replace in time.

DHAYANANDHAN.S

it's done its job on its useful life and fatigue took it. If the shaft key & slut would have been cylindrical in shape, it's life would have been extended for a little while.

The product being agitated may have contributed to this. Lumps on one paddle, additives put in as a bolus ( particularly if added at the periphery ), variable consistency, etc.