We had a case similar to your case and we didn't catch any thing regarding the crack but in our case the crack was in the shaft at the key

I believe to catch cracks like that , we cannot rely on ordinary vibration measurement we have to use equipment to measure torsional vibration.


Greetings Mr. Ayman Gamal,

Pleased to note your reply.

Yes your suggestion is Torsional vibration.

Can you share who provides such service of Torsional vibration and how they do and what type of instrumentation they use??

Instrument makers, Training providers in Torsional vibration  please let us know.


Thanks for your feedback.




I have used the previous model of this product for my diagnostic services for torsional, bending and axial shaft vibrations.


Can you show a photo and give more history and details of the impeller failure? Were there any signs of a rub or Cavitation damage? Were any other components failed? Have you considered the possibility that debris entered the intake and caused impact or jamming, but was not present when vibrations were measured? 


You produce a suitable amount of pulses from the rotor by zebra reflective tape, encoder, tacho pulses from gear etc. Enter puls train to a "tacho box" that convert pulses to a signal proportional to to the speed of the pulses a basic tachometer that is I got one from Cop Tech I think it is. Enter that signal to your FFT device and look at the result. "Nothing", it's good, something? Start analyzing and thinking. Other methods exist including a new involving surface speed measurement.... Go Google and and have fun.

If the crack in the photo is the only crack, then it does not look like it could cause unbalance and 1xSS vibrations. I assume you only make survey measurements, so the crack could have occurred anytime.

I doubt that crack could have been caused by torsional vibrations. It could be a stress crack (material flaw) from the casting process, and possibly "helped along" by debris at the inlet.

I have measured vibrations on several ship ballast pumps. I find the picture to be unusual, and I would like to see additional photos. An impact test could identify natural frequencies and if there are any excitation matches with vane pass and harmonics (+/- 1xSS).


Greetings Mr. Walt Strong

Pleased to receive your reply through this forums.

Once you have done some ballast pumps on ships, this is horizontal centrifugal pump driven by motor in Engine room and pump located in pump room of a petro n Chemical tanker.

The Supdt sent pictures after we gave the report of vibration condition and certified in satisfactory condition. They have opened up as this sister vessel pump was opened and found the similar crack and with that doubt they opened and this impeller had similar crack on the impeller tip.

As you know again to get back that ship for next inspection will be after 2 years if that vessel remains with the present ship management.

I will try to procure more photos through the supdt if all go well.

Once again thanks for your expert comments on this subject.







Greeting Sir ,

Premature cracks or primary crack development can not be detected by Vibration analysis until unless it is causing /experiencing flow disturbances at the racked area .

We can detect as a Blade / Vane pass frequencies along with flow induced vibrations .Torsional Vibration might help and it can be detected by various direct contact and non contact methods .Such as strain gauge fixed on shaft , eddy current rings or Laser etc .

I am not sure but Ultrasound noise (on casing)implementation to listen any noise generation (completely guess ) by comparison with normal pumps .

Please refer the below attached paper for your reference .



As you know, vibration measurements cannot catch all faults present. There are many reasons including:

1) Fault not present or fault symptoms not active at time of measurement

2) Fault produces minimal vibrations and possibly hidden by background 

3) Sensor type, sensor location-direction, and sensor mounting

4) Measurement and analysis parameters

5) Data interpretation

Two machine components with same/similar cracked component indicates a class problem worthy of investigation and possibly including:

1) Metallurgy analysis

2) OEM Processes such as casting, heat treating, stress relief machining, and dynamic balancing

3) Natural frequencies and dynamic excitation forces

4) Presence of other defect symptoms such as Cavitation, erosion, corrosion, impact marks, wear ring rubs, and blocked/damaged/missing suction strainer

5) Machine operating procedures that cause excessive pulsations or Cavitation

I have worked on one famous crude oil vessel (Alaska oil spill) that was operating a variable speed ballast pump (turbine drive) at excessive speed with low discharge pressure and high flow. The pump had a lot of Cavitation and was a source of high vibrations and noise throughout the house.



Both papers are interesting, but perhaps not relevant to the small crack in the large impeller shown on OP photo. The test pump had missing pieces of blades, hardly small cracks, so flow was affected and the distortion could be measured with vibrations. It is unlikely that the small crack in the OP photo could cause a flow disturbance and measurable vibration symptoms. 

Mr. Chary did not provide any vibration measurements on the pump.


Add Reply


HACKATHON in association with IMC-2019