Failure of Shaft

Hi sb,

A picture of what you are trying to describe could be worth literally a thousand words!

After getting a digital image (or images) of this part onto your computer, just click the icon on the editor toolbar to get the picture into your post.

I think you will get more replies.

Mike

As stated above a picture would really help. Possible solution to what I can comprehend may be failure with the set screws. If they were loose upon removal, you may want to double set screw them to keep them from loosening. Then again there may have been a poor machining between the shaft and geari allowing it to wobble and cause stress and fracture.

Please provide more info and pic if possible.

Sounds like a torsional failure to me. How did the crack propagate? Again a picture would really help. Also a description of the machine train, and more information on the type of coupling would help.

It is a gear box between Ball Mill of Pulveriser and Motor. The shaft has broken at the input side (ie on the high speed) coupling end of the gear box.

material is 18CrNiMo6, a standard matl for this application. The interesting part is the zone of failure.

This is one of the two shafts failed at 1000 hrs operation. A third failed only at 200 hrs, but that is a possibly known cause (Multiple propogations, radial from a grrove - notch effect) But this one (rathet two) didn't have notch to begin with. Keyway tip is not a notch. Look at the bottom picture (with broken shaft inside cplg) The keyway on right side (sorry can not put lines or arrows here) at the middle of the keyway.

The coupling is Gear type.

Hi sb

Ouch!

Just a little more info.

Does the coupling allow for shaft mis-alignment?

What is the diameter of the shaft(s)?

How soon after failure were these photos taken?

Hi Friends,

There must be a different dia. steps after coupling step on shaft. There may be chance of sharp edge instead of round edge at steps.

Other reason may be key lengthe might be more and cross the next step of shaft.

Moto

Shaft dia is 140, transmitting approx 2MW power at 1000RPM. It is a gear cplg, dyn balanced to 6.3, slight misalignment is possible. Our erection staff records show the alignment within 0.1mm which is in perfectly acceptable level.

Even in misalignments, the roots are stressed.

Another thing I should mention is the shaft step corner radius is R2 (where it has not broken, total corner radius is visible in first photo all around periphery, So this notch has not failed.

The gap between the keyway end and the collar is 10mm.

(Alignment with in 0.1mm) is this a cold alignment or after heat expansion?

The photographs are a definite help here. In absence of a root groove that would allow stress to concentrate into a shear plane, I think the main cause of the fracture is (a) misalignment and to a lesser extent (b) torsional vibraton from a misalignment, dynamic imbalance, or the sum of harmonics generated by the bearings, the driver, and the load.

Alignment is often neither as easy nor as precise as it appears, and of course I do not know the methods used by your erection crew. Is the outcome of their alignment truely as precise as your standard indicates it must be? Total alignment includes both radial alignment and axial alignment-the radial part is usually quite easy to verify but the latter requires more effort. If one is trying to make good axial alignment in a confined space it can be much tougher.

Me? In absence of part defects I would challenge the values of the "true" alignment. Good luck!

Best Regards,

Ing. Robert Forbus

I agree with you. A and B you listed are likely the total combination. I would go one step more to add that a softer material is needed for the shaft.

sb:

What form of quality assurance, magnaflux etc.; is performed prior to assembly. Is difficult to discern from photo but could there have been a crack before installation.

What is clear representation of terms carburized = case hardening and hardened and tempered?

We purchase this gear box from our vendor.

This is the first time he has supplied this type of gear box, and out of 5 pieces erected 3 have failed (the vendor has supplied total 9 gear boxes, of which 3 have already failed, replaced with gear boxes of the same supplier.

Other gear boxes have not yet clocked the same hours as the first two failed. Third failure is at 200 hrs.

Earlier gearboxes supplied by other vendors are working (we are supplying equipments from 1988 onwards and this is first time in my carrier this shaft has failed. The shafts as old as 20 years are still alive.

The shaft is integral with the pinion (helical 2 stage reducer gear box)

The shaft as far we got information with the manufacturer is rough machined, case caburized in the gear area, quenched and low temperature tempered (to maintain hardness at gear flank to 60HRC), Gear ground, Keywaymilled, Journals ground, Balanced (in the sequence).

You may be interested in third photograph also. The shaft unfortunately has a retaining ring groove 4mm wide, between coupling and the Drive End Jrl Brg (ie max load area) It has failed at this groove.

I'm going with Steve S. in this 'a torsional failure'

When ever any failure occurs in Machine or Gear Box the weekest link / component fails.This shows either the output Torque requied is more or the matalurgical properties confirmed to you by supplier is not correct.Go for the further analysis of broken shaft by third party definately it will solve your problem.Also compare the lubrication provision of other G.B supplied to you which havenot failed and output Torgue.

Because we have repetitive failures, I guess a few questions would be: 1- What is the material. An 8620 case hardened may not have required core strength. Look for material upgrade. 2- what stress rising features are on the failing component, prime example: the retaining groove may want a fillet radius at the root. Perhaps an alternate design to eliminate the groove completely. Also, Mag particle inspection may reveal micro cracks in the finished part. Without knowing the specs of this train, these suggestions may or may not address your issue.

Hi all,

First of all: What's your quality requirements imposed to your suppliers?

Obviously, it's a problem of such specific supplier (3 out of 5 is too much high failure rate). Check your requirements on surface crack examination of finished shaft and/or investigate the normal practices of this supplier and compare it with those of else you've had.

According to the last picture it seems a shear fracture but with the ultimate rupture zone quite uncentered. This may be caused by unbalance of torque applied (misalignment problem as everybody have said), but even in this case I would examine the remaining shafts (at least in the keyway zone) with fluorescent magnetic particle method looking for small cracks.

Good luck.

Hi sb,

From what I understand from your description, my experience and knowledge, the failure appears to be due to gear box design problem of your vendor. Some of the troubleshooting steps that I can think of are given below:

1) To start with, if the gearboxes from other vendors working with same parameters such as load and speed are working fine, then it is certainly a problem with your present vendor's gear box. If you replace the gear box on this equipment with that of another of your tested and reliable vendor, and if it works fine, this is confirmed. Of course, you will have to wait for 200 - 1000 hrs of running for this.

2) Finding out why the failure is occurring in the present case will be a little tricky. It should be possible for you to check, record and compare the load on the motor of two similar equipments with different gear boxes (of different vendors) under similar external load conditions. The external load on your present equipment may be more than normal or the frictional loss / FHP of your gear box (with failed shaft) may be higher. This can be due to misalignments, bearing/lubrication problem or inadequate centre to centre distance between matching gears.

3) If the load due to your gear box (in which shaft failure is occurring) is in fact higher than normal, I think the area between the keyway tip and the collar is the weakest since the other portions of the shaft are either inside the coupling or bearing with running or interference fit. If there are reversal of shaft rotation, then the torsional fatigue will aggravate the problem and lead to early failure. There may be torsional impact on the shaft during reversals.

I had a similar shaft failure pattern recently on intermediate pulley shaft of my lathe machine, but shaft diameter was only 1 1/2".

I hope my comments would help you in troubleshooting.

hay guys this sure looks like some drive line failures we had on an auto balance rock crusher that i build. also on some right angle gear boxes we use after about 100,000 bucks worth of replacements out of our pocket we had some of the material analized IT WAS SHITTY MATERIAL would not even pass a sharpie test for 4140---------we started using FLEXOR shaft material have not broken any shafts in over a year

material already tested and is perfect (18CrNiMo6) matching exactly in chamical. Mech is YS=921.9 N/mm2, UTS=1092.0 N/mm2, %E=13.6, RA=51.0, Impact =53 J- 3mm U Notch.

Core hardness 327 BHN

ASTM Grain size 9, with uniform tempered martensite. with a few sulphide inclusions.

These are results from our in house lab.

What still bothers me that the crack has started from an relatively low stress place. In any of the discussed problems the unbalance, alignment etc why the weak area (key root) is untouched by crack? The stress flow lines will always bypass this zone. And this feature has made it interesting.

It has not even moved to the other stress concentrator (the corner radius) and has neatly left it alone.

Hi sb,

Can you draw a simple sketch of the particular shaft / pinion and show the exact positions of the coupling, bearings, gear, etc. which may help every one to understand and analyze the problem better?

Yes, it seems the material is OK from the point of view of complying with chemical and mechanical properties specified.

But think that a crack is a more strong stress concentrator than the corner radius and that if cracks are present, the weak area are there.

Let us know how it goes.

Pl find the drawing of the shaft.

In the drawing you wrote "groove where one shaft failed.."

A groove is also a stress concentrator.

Have you checked if there are some differences in the groove root form between those broken and any of the other supplier?

Have you checked the not used yet shafts for cracks in these zones?

Best regards

I Would have to go back to this question. What has changed since the product worked correctly? That would be the new gearbox. I would think I would be trying to determine alignment and runout of each of the mating parts. Have you changed the coupler that joins the gearbox? Is the coupler concentric? Are either of the shafts that join bent? Do you have access to an infrared camera? The shaft should be operating hotter if it is flexing. Can you overload the equipment in a test to watch for early signs of failure? There are some portable magnaflux systems that should allow you to watch for small cracks while in a test mode. keep the information coming.

Again, this looks like a torsional vibration failure. In that case the crack will initiate in the outer surface at the nodal point in the torsional mode. I expect that there was a stress riser at the crack initiation point, even if it was just a nick or scratch in the surface.

Is there any torsional damping in the system? It doesn't really matter what the torsional natural frequency is, it will be excited by the impulsive energy of the mill.

I would recommend a torsional natural frequency analysis, with attention to the mode shapes, and then a modification that either adds damping to the system, change the shaft strength, or a combination of both.

There are consultants out there who could do this analysis very quickly and for just a few thousand dollars. Sounds like the gear manufacturer needs to contact one...

Any beach lines or inclusion found in the break? You did mention that the material is carburized hardened and tempered. A close up picture of both pieces would help a lot.

The broken piece of shaft is sent to our corporate R&D where they will come out with SEM result. However the problem is they will only give so much. Crack initiation point, propogation etc but we have to guess the root cause. It has happened earlier too. They gave their analysis on propogation etc, we could successfully come out with the cause.

The drive chain in case any body wants to know is

Motor(1000RPM, 3 MW, normal running approx 2MW)

(Grid Coupling-Bibby with metallic spring Grid)

Gear box (1000/121 RPM)

(Gear Coupling)

Pinion / drive shaft + Girth Gear

Horizontal Tube Mill (Running at 19RPM)

Dead Weight of Tube Mill is approx 200 Tons and the Coal is approx 100Tons

In this system, though the shock load exist but is much lesser than the Vertical Milling Systems we manufacture.

The gear boxes internals are all supplier design, and at the moment it will be difficult to open and see the earlier cases.

The shafts have rubbed a bit after failure (due to momentum of the mass of the Mill) before the controller tripped the gearbox on Low Current. You can see the damages taken place on the surfaces of the first gear Box.

In case of the third (200Hrs failure) I was there just after. Infact I visited due to 2nd gear Box failure and the day before I reached, the third has failed.

The observation on failed surface is that the beach lines perse was not visible to naked eye. The propogation(or beach lines) looked to be purely radial. The surface was not polished and was matty and black. The colour has not come out well in the photograph because of the flash.

sb:

The gear boxes internals are all supplier design, and at the moment it will be difficult to open and see the earlier cases.

It is unfortunate because it may be the case itself which is the main factor of failure. In that in resent times gear box cases have been thinned either for weight reduction or competitiveness. The result is often realized as a weakened gear box i.e., the case flexes and the shafts deviate from alignment under load out side of the specifications prior to case thinning. If this is occurring and exceptional amount of force is being placed upon the shaft end at it's weakest point "snap ring groove root".

May it be the shaft failure is and effect rather than a cause.

What do you think...

There is no doubt about the crack being the effect. The snap ring groove is a sure potential cause and we have already asked the supplier to redesign avoiding the snap ring and suggested alternative fastening.

In this our theory was

a) stress concentration due to groove

b) Low corner radius at the grooves. Since a 4mm groove can have only a 0.5 or so corner.

c) As we know that these grooves are made by plunge cutting, there may be chatterring in the groove, which will initiate cracks.

d) The chatterring itself is always associated with rubbing and consequential overheating- Potential defect initiator.

e) MPI can not show anything inside such a thin groove.

But this explains the 200 hrs. As I told for this case we already have hypothesis. But what about other two.

Also I have forgotten to mention, all the bearings (both sides of the shaft) look healthy, and there is no telltale overload indications on the bearings. All the 6 bearings were physicalli inspected. I feel Misalignment/ unbalance must have shown up in brgs.

Photos are for the tuw brgs (first DE- Broken side) 2nd NDE - blind end.

I think I think you are tangling with a cat of many tails and a redesign of gear box including case deflexion testing is a good answer.

If problem were case deflexion I'm not sure any wear factors would be noticeable in the bearings in this case. In other applications I have found unusual wear of gear tooth though when cases have spread under load.

sb,

I think finding out why the shaft failed at the key neck will be difficult than finding why it failed at all. CR4 members have already given many possible reasons.

I am curious to know, why are you so keen to analyze the problem, is it just for academic interest? The designer and manufacturer of the gear box should be able to tell you the reason for the failures. What does he say? He should solve the problem or replace the gear box, if he is interested in getting further orders from you.

I am from Quality Deptt, unless I know what has gone wrong I can not prevent in future. And as in todays condition you know, vendor make a mistake, but the responsibility lies with us in the end as the total package supplier.

Tomorrow another vendor may land up in the same problem. Here it is a massive work- 12 Nos of Gear boxes are affected and that means 1000MW of power is in touch and go.

Engineering and vendors both are still groping in dark after around 10 days of analysis (more than 1 month after first failure)

The matter is critical. Though the Guarantee period is over but we take pride in our equipment giving trouble free performance and here we have with multiple failures. So it can be called academic- to pinpoint the cause and replace all the defective whether failed or not on a war footing. But till we can pinpoint we are not sure. And the corrective action will be taken on all the 12 gear boxes. End of the day it should not happen that we take a wrong action. So I am trying to find from you all if anybody has experienced this problem any where and what are the findings. Our Reaserch team is working and SEM will be taken in a few days.

Sorry if the research looks academic, but to pinpoint the root cause, we have to go deep, and this is a part of brain of the world storming. All your ideas are already being crosschecked with our staff, But the reliability is poor as you know (all the recorded data will be perfect) and broken train is difficult to check for alignment.

However some of the gear boxes still not broken have already been checked and were OK (Also we have replaced the shaft and with new shaft, the alignment was OK, no disturbance was necessary)

Alignment we consider low probability (due to the high frequency of failure) and we are trying to put a gear box from other vendor too. Problem is knowing the problems no customer will allow me to interchange.

"So I am trying to find from you all if anybody has experienced this problem any where and what are the findings. "

Yes I have, and it was a torsional resonance issue.

I have written several times that it appears to be a torsional vibration failure, is there some reason you don't think that is the case? The good bearing condition is consistant with a torsional problem, and as you said it makes mis-alignment an unlikly root cause.

I am interested to know if you have some data or analysis information that has eliminated torsional vibration/resonance as the issue.

Dear Steve

I will look at this problem. Your case is any way true in the 200 hrs shaft. And most likely failure root cause is the groove for the ring and we have already written the vendor to eliminate the groove in all future shafts (the ones that will be used for replacement)

The other shafts I will find out from site. Normally in these motors, we have a motor vibration monitor, and data is retained. I will find from site, if I can get the data. Only it will take a couple of days (today is saturday and I can get by monday only)

Let us see.

Meanwhile we are asking the vendor to carry out flouroscent MPI at the shaft and report the result (I checked the reports submitted and found that LPI is carried out- and that is insufficient for microcracks).

Also we are to have a face to face meeting with the vendor and our group for sharing our findings. Will share with you what we find on the thread. Wait for a week. Meanwhile if you get any pathbreaking theories let us share them. so that we eliminate the nonpotentials.

The best thing about this type of problems is the sharing of expertise- may be theoritical and also practical- across us and getting new knowledges after more than 25 years of leaving college is something irresistible. This has been far more fruitful than the google search.

May be we should share the solved problems too and create the information base for the Engineers (Knowledge Management Portals)

regards

Hi sb,

I think for a company keen on maintaining it's reputation, best option would be to repalce the gear box with that of a known reliable vendor. From what I understand, whatever the results of your analysis, it should be tried out by trial and error at your customers' end. For this your customer should be willing and cooperative.

I have found two links which may be of some help to you in analyzing the problem.

FAILURE ANALYSIS OF SHEARED SHAFT OF A BRINE RECYCLE PUMP

Hi sb

It sounds like you may have a combination of troubles. It's possible the shafts are becoming flawed at the weak points during the hardening process or during machining. Has the shaft hardness and depth been tested near the points of failure?

I am in agreement with Steve S. as to torsional vibration being the main cause of the failures. This probably is moot, but has anyone checked the internal gears for excessive backlash/mis-alignment or for excessive linear movement of the gear shafts? Any of those conditions could exascerbate the transmission of vibrations to the input shaft and cause it to fail prematurely especially if the shaft is already weak.

Just some thoughts.

Please keep us posted as you discover the true causes.

Jeff

In case you are all still awake-

the problem was the process-

carburised with approx 2-3 mm stock on dia, removed the stock with 0.5mm grinding allowance, quenched and tempered and then the keyway cut, groove machined, journals ground and gear flanks ground.

With so less stock carbon must have migrated to the left out area making it hard on Q&T.

Keyway cut with single pass (milling cutter with exact width) . This must have created abnormally high stress at the keyway tip (where the cutter is upmilling on one side and down milling on other) and compounded with the hard surface is a recipe for microcrack.

If the crack is small enough it will escape the NDT (LPI/ MPI) due to edge effect.

The vendor was suggested to modify the process

- Carburize with > 5mm stock on surface preferably about 10mm.

- Keyway to be cut with approx 5-6mm lesser width endmill.

- The 2-3 mm allowances per surface to be gradually reduced in another 2 cuts - with last cut about 0.5mm- this will takeout the stressed surfaces.

- suggested also to try local stress relief - may be Shotpeening to generated favourable stress.

Now we are keeping the fingers crossed and hope for the best.

Thanks all for the help.

Would it be practical to carbonise and temper after the milling is finished?

distortions and dimension control. carburizing will increase the volume and the amount of increase is almost noncalculable. The Hardening and tempering will distort the dimensions.

The following modifications were applied

1. Material check (micro) for stringent control on the inclusions especially shape. The specifications accordingly modified with controls.

2. The Key way process modified to reduce the residual stress.

3. The stress raising groove eliminated and alternate fixing arrangement done.

No further failure (touch wood ) for last 18 months. Still kept under close watch.

Good. I am happy to hear of your good news.

Hi sb,

I am working for a gear manufacturing company too. Just in case you have found out the reason for such high number of failures, please share them. Very curious about it.

I am of the opinion that Torsional Stress has exceeded beyond its capacity.

The picture posted by you (Refer Post No.4) - damage has not taken place instantly. Few patches are white in color - which shows that was the only portion transmitting power just a second before the failure. The crack/damage has taken place and kept extending gradually and failed.

Now 2 Questions are to be answered.

1. Was the shaft tested UltraSonically during maintanance shut down.?

2. Whether the gear coupling while alligning, the root clearance and backlash was with in limits.?

Your allignment could be with in limits but the root clearance and backlash for the gear coupling may be inadequate.

Also check whether very heavy coupling is provided than required size, and also check for instant kick loads/jerks - which causes large percentage of shaft failure.

DHAYANANDHAN.S