High BPF (3x) at Velocity and ENV measurements

[TKL8]

I have a pump with high BPF spike and harmonics symptom and this symptom started to appear since early July. I trended the symptom and requested for a impeller check last month. However, the impeller and wear disc are in good condition but the measurement still indicating the same symptom. In fact, the overall vibration level increased. This time, the 2xBPF spike is the dominant one, standing at 10+ mm/s. The ENV measurements also indicating harmonics of BPF.

1) What does multiple harmonics of 3x in the Enveloping measurement mean? (B2: 3kcpm - 60kcpm and B3: 30kcpm - 600kcpm). Note that the symptom showed up at these two bands.

2) Despite having checked the impeller and wear disc condition (which is in good condition), what other causes that may cause such high BPF and harmonics?
I understand that unequal spacing between the impeller and the wear disc may cause this but I would like to know what is the normal practice to check such clearance?

Your advice is highly appreciated.

 

[vanstoja]

Multiple harmonics of impeller blade passing (IBP) in centrifugal pumps is symptomatic of a distorted fundamental IBP waveform, generally a step or ramp in the sinewave due to an impulse event. The more intensive the impulse the higher is the number of IBP harmonics that may show up in a wideband frequency-amplitude spectrum. Three known sources of IBP peaks are (1) impeller blade trailing edge to diffuser (or volute cutwater tongue) inlet edge interactions across narrow cutwater clearances, (2) impeller blade leading edge interactions with axially streaming inlet vortices coming from either upstream inlet guidevanes (or flow splitters) or upstream piping bend-induced flow distortions,(3) impeller blade trailing edge wake reingestion of meandering vortices produced in vaneless and dump diffusers. The second and third sources are progessively less likely to occur than the first source.
You shold probably examine your wideband spectrum to determine the number of IBP harmonics present. For source #1 we have seen as many as 7 IBP harmonics at 3600 RPM synchronous speed with blade-vane cutwater clearances less than 10% of the impeller outer radius. Increasing cutwater clearance beyond 20% and decreasing pump rotational speed both tend to reduce the number of harmonics present and their signal to noise ratio above the broadband level. If streaming inlet vortices are involved then the problem is upstream of the pump impeller. Impeller wake vortex reingestion can occur with a shrouded vaneless diffuser if if the impeller and diffuser shrouds are misaligned enabling a Coanda wall attachment phenomenon to generate meandering vortices downstream of the impeller.
I am not familiar with the term "friction disk". Do you mean impeller wear rings controlling leakage flow around the impeller?

 

[dgallagher]

You might post this question to Dr. Bob on the SKF Reliability Forums - he's their guru in Enveloping. It's at

http://www.skfcm.com.