Failure Analysis of agitator shaft

Nice example of fatigue. A very similar question was posed recently on here. Does this agitator have a support somewhere in the vessel or does it simply 'hang' from above? The fatigue could either be as a result of misalignment of the bearings, if it is supported, or 'whirling' or imbalance of the agitator paddles. It should have been balanced before installation, so there is the possibility that it has been run at a speed near to one of it's natural frequencies. I'd have thought that a vibration would have been pretty noticeable for it to fail in that sort of timeframe. If it has run 24hrs a day it's done about a million cycles, clearly you were expecting a considerably longer life than that.

Just a thought, clearly there are reasons to use stainless, but it isn't the most fatigue resistant material.

Wrenched,

Great and quick response. The agitator just 'hangs' from above supported by a drive 'spool piece' that is connected to a gear box and through a mechanical seal. Normally 57 RPM does not call for balancing. We certainly do not expect fatigue failure after a few weeks of 24 hour operation. And yes, we have compulsions to use SS 316 L

So would I be right in thinking you have a number of these installed and which have run without a problem over time? It's possible that a flaw in the shaft brought about early failure, but I still wouldn't rule out something in the operating conditions.

A close study is necessary to reveal if the material composition is as specified and if it has been affected by heat (welding?) or if there were any physical features which have compromised it's life like notches, undercuts etc.

PS Just re-read your original and note it mentions welding the shaft to a coupling. I think this needs much more investigation. Any other designs like this? What type of welding is this? Friction?

Yes, there are agitators that run under similar operating conditions. Having said that, this particular agitator had its welds 'come off' from the coupling collar at the top and bottom. The reason was not known. It was assumed that the welding that was done was less than adequate. There was urgency to put the agitator back in operation quickly.The collar was then rewelded to the agitator at the top and bottom with a larger fillet and also holes drilled through the collar (till the shaft) and plug welded as a measure of added protection. So it is possible that the heat during welding could have caused it. However, this is a 100 mm shaft - can you make a shaft this size brittle with 'over' welding?

This sounds plausible. Clearly there's more heat input if the weld fillets are bigger or more in number. The heat affected zone would be deeper. Having said that, whether the load is bending or torsion, the outer surface of the shaft is all important being where the stress is and where the fatigue-critical surface is.

You say it has just been re-welded, how long did it last the first time? When these are manufactured, perhaps normally precautions may have been taken to condition the material appropriately afterwards, like a heat-treatment of some sort. When you welded it, what sort of filler material did you use. The material in and around the weld you made might have been hard and brittle.

Is the assumption that there was no spurious mechanical obstruction in the agitated tank, like a dead cow or a loose aeroplane engine, a correct one to make?

On a tangential path, why use a shaft of solid stainless?? Bit expensive not to mention heavy isnt' it?

First on the tangent.. the agitator is driven by a 25 HP motor.. it stirs up some pretty sticky viscous material.. the agitator has one flight of blades and is 4.16 m long below the in tank coupling. I guess that qualifies for a solid shaft.

and PW Slack... no there was no aircraft engine or even a cow... but a gnome holding the blades from rotating cannot be rules out....!! Seriously, no solids to cause excessive loading..

Welding was SMAW with E316 fillers... standard stuff...no post weld treatment... To me it looks like fatigue.. but ... how and why??

As far as stainless the fracturebility is higher, with very little necking before failure, before one try's to solve for it, it would be good to know the operations.

Is this for a food or pharm. operation. Answer will dictate for appling the fix including sanitary standards, or mat'l type.

Was this a new shaft or was it rebuilt.

Phoenix911,

its for handling synthetic latex.. so not pharma... no need for sanitary finish or designs..

I am not sure if the end was fused or welded.

1. If it was welded, it did not have enough penetration and or the voltage was low, thus, it appears to be porous.

2. The bottom portion of the photograph appears to show oil or heat stain which, in my opinion, it was offset, and the length of it could have produced a pendulum effect and it broke at the weakest point.

In plain english, shaft was not straight.

Thanks,saca03142, I am wondering how the pendulum or wagging effect could have happened... bad fixing of agitator coupling to the drive shaft coupling?

The heat of welding probably distorted the shaft. The "hot side" (generally where the welder is concentrating at the moment) expanded and that would be enough to bend the shaft and create the eccentricity that likely caused the rapid fatigue. You should use a different coupling to avoid welding.

Difficult to analyse from a photograph, and I am not a metallurgist, but it looks very much like a fairly low cycle fatigue failure initiating from some sort of surface defect that has propagated across the the heat affected zone (or through the weld) quite quickly and then slightly slower until final fracture. Where the coupling is welded to the shaft would be a point of stress concentration if the coupling was not concentric to the shaft and the shaft was "waggling" around.

I.E. out of balance.

The picture is not showing the coupling side also to give clear picture. Because the shaft is showing welding failure at the end as well as shaft also seems to be broken.

In my experience _ this is caused because of:-

• No/poor weld preparation.
• Improper welding rod - quality and size.
• Improper welding procedure.
• Heat stresses developed during welding process.
  

Also, give the MOC of the coupling to suggest more. The end mark on the photograph also suggest that the shaft was having crack and the centre part broken at the load. What was the idea of welding coupling because normally coupling is not welded to shaft for agitation. Also give idea about the agitator blades and the material inside the reactor because if it is viscouse and blade area is more, it can bend the blades and if blades are made strong, it can break the shaft.

U.M.Patankar

Hey, Uday and others.. see picture of the coupling. MOC of the shaft is 316L.

The coupling is welded on the shaft to so that it couples to the drive shaft. The agitator shaft can then be removed.

The motor is 25 HP and to my mind does not have enough torque to shear off a 'healthy' 100 mm dis SS 316 L shaft even under locked rotor conditions (and even if the mtor protection does not trip) Ofcourse if the shaft has a defect, or been weakened due to weld heat etc, then sure, it could happen.

Await your inputs. Thanks a lot for your interest and suggestions.

I'd agree that it is not thre drive torque which could have done this, and IMHO torsional oscillation is far less likely than bending to be the cause - just looking at the dimensions. I think you said the agitator reaches 4m down into the vessel. That's a big overhang for the bearing and demands some pretty good tolerances such that the thing will run adequately true. What is adequately true is the question, and what would we regard as adequately balanced. It's behaviour will also be governed to a degree by the clearance or play in the bearing(s). We could have a go at working out the critical frequency if we had a dimensioned sketch of the shaft and paddles.

Wrenched, we did the critical frequency calcculation and 57 RPM is well below the first critical frequency. So thats unlikely.

Now on the issue of balancing - the shaft (4.16 m) had a run out of 4 mm at the end of the shaft. Probably thats too much. But any damage should be transmitted through the drive shaft on to the mechanical seal and the seal should have failed. Could this 'imbalance' cause shaft shear?

Consider this.. if the coupling on the agitator shaft and the coupling on the drive shaft are not sitting aligned, could the agitator swing to trace an elliptical pattern at its bottom? This could be the pendulum or swing effect causing low level fatigue failure worsened by heat effect due to excessive welding?

Thanks for coupling photograph. The coupling shows the v groove. Is there any reduction in size? If no then it is clear that welding was not proper. If yes, then you have to strengthen the shaft at the welding joint.

You can go for welding and then putting 20 mm thk sleeve x 250 mm long slide/tight fit on the shaft and weld it to both sides. Use low heat input electrode only. Also if you could make male-female threads in such a way to tighten the joint while in rotation and by making v groove at the joint weld it to give it more strength. It is true that the shaft dia 100 is good enough but the length(4.16 Mtr) is too long. if the reactor is filled fully or maximum and there is only one row of blades at the bottom, there is too much shear load on shaft as the material stirred is viscous.

It is better to repair and run the plant and start making new shaft.

U.M.Patankar

I am at a loss with engineering talk, but being a tech and having the ability to discern flaws, I believe the welding was inadequate and the shaft did not rotate true, having said that my next suggestion would be how to correct it.

1. It can be bored in the center 1.5 -2 inches deep, three of four side holes for bolts to hold it in place and balancing, no more welds.

2. redesigning it this way, you can alwasy replace the end, if needed.

3. Have you also considered using a u joint, all vehicles use it to compensate for uneveness.

I made this drawing hoping that it can be better than what you have.

There doesn't appear to be any beveling of the edges prior to welding, and the welding penetration depth is suspect. Could'a been too low welding heat -- looks like what a "cold" joint.

After seeing and reading all the posts -- I feel :

• You should have nothing more to do with Solid dia 100 SS316L shaft-so expensive!
• Plain Carbon/Constructional Steel Tubular Preller shaft will be your bestchoice
• You can take a standard heavy truck Cardan Shaft from your Auto Dealer,get the matching flange bolt holes machined on the Speed reducer output-and go!
• Latex: Paint Epoxy and never worry any more.
• 25HP,57rpm--the torque is far less than a truck climbing a steep grade.
• CR4 colleagues will give you the torques--compare with Truck part data sheet.
• KISS

You should beef up the diameter at the flange coupling and use a reducing stepped shaft diameters down to your 100mm.

Or fit a lower bearing mounting at the bottom of the agitator tank to stop side deflection forces fatiuging the welded sections

I come a bit late into the discussion but I think it will be interesting.

The first picture was not clear enough but the second gives some hints -at least apparently-:

-the fracture is circular and progresses in radial direction toward the center.

-the second picture shows the radial progress quite well

-when the torque became too important for the remaining section the shaft was destroyed by torsion

- the radial crack and the conical start indicate most probably a "bending load" as crack generator but unbalance gives a radial force which rotates with the shaft that means that a circular crack indicates a bending moment ROTATING with respect to the shaft.

- the question is what could generate in the container a local force at a distance to the centerline so that a bending could load the shaft? It can be imagined that the shaft made an angle and the stuff was not only stirred in radial direction but also "compressed" against the bottom.

- the welding seems to offer a high stress concentration at the root (i estimate according to the second picture about 3..4) so that it is most probably the reason a crack could start. I do not know how big the diameter of the rotor is but the bending moment can be estimated at about 700 Nm.

It seems that the failure is a combination of bad welding and a parasitic effect which was not taken into consideration and thus was not expected.

I would appreciate comments.

Can be due to different reasons:

1.difference in materials composition

2.the nature of the soldering if any

3.inhomogeneity of the assembly accentuated by the forces applied

A solution in terms of investigation would be using an XRAY system prior using

this would show voids, homogeneity of your structure...

Why you are using welded joint for such creatical component u can use some coupling arrangement,eg flange coupling,sleeve coupling,flexible coupling if some miss allignment required.due to improper weld alignment it break from that stress concentration point.And also welded joint strain hard at that jone.Tahnks...

316L is austenitic stainless steel and is a non heatreatable grade, however wrong welding proceduire / filler metal can result in carbide precipitation in the grain boundries acting as stress risers. Shaft is sufficient in size as evidenced by the small rapid overload area(center final failure) as compared to the total shaft area. Misalignment of shaft to coupling welded joint would result in high cyclical stresses.

Thanks for clarifying what 316l is.

Are there 3 or 4 blades on the agitator?

Is the pitch of the blades adjustable, and are the blades equally pitched or aligned?

take it to a welding / metalurgical engineer and evaluate it

I'm a certified welder i do stainless alum and carbon steel been doing it for 15 years .. now just looking at your pictures I say bad weld. Porous and poor penetration. Doesn't look like it was grooved at all to begin with. Take it to a welding inspector or even a qualified welder. If hes good he tell tell you if it was a viable weld. You need someone to put his or her hands on it.

Mabe I'm being thick, but I can't see much evidence of a weld...It would be nice to see the part that came off too.

Go to post #17 by Ash55.

Cheers!

Looks like the join is in a daft place...

I'd have a hole or recess in the big round plate and weld the shaft into that...probably be much cheaper to make too?

But heck what do I know?.... I'm a cat!

Hey Del,

Thats exactly the way it was. There was a hole in the 'plate'. the shaft was welded to the bottom of the plate with a 15 mm weld. There was a recess to the hole on the plate at the top and the shaft was welded on the top as well with a 10 mm weld.

So, there was no 'joint' on the shaft which broke off. The welds came off in the first instance and when it was rewelded with additional welds as plugs on the collar of the plate, the shaft sheared at the top of the weld.

Appreciate all your comments... Thanks a lot guys. Would I be thinking right if I said the summary seems to be that the first weld was not done well (poor fusion etc) and when it was rewelded, the heat affected area caused the shaft to go brittle and shear off with the load?

Dear Mr.Ash55,

You have not mentioned about the fluid/semi-solid which is being agitated by the stirrer and its character of solidification/settling. Once settled or solidified, when you start the motor, the Torque demand will be very high and may exceed the full load torque. More the time lapse, more will be the settling/solidification and hence more the torque demand.

Before starting, you have to instruct the staff to rotate by hand the input shaft of the gear box forward and reverse slowly and make the stirrer to rotate one full round and then start.

The photo of the broken shaft suggests to me - failure due to fatigue.

DHAYANANDHAN.S