Hello,

I want to hear any advise about workflow for harmonic dynamic analysis of existing piping connect to reciprocating compressor.

There is piping vibration issue at site, a third party had processed vibration velocity measurement versus frequencies base on EI guideline and the lines are falling in concern/problem categories which require further assessment

Below is as far as I know:

#1:
The lines/system have to be modeled by CII with accurate in boundary conditions (support stiffness, clearance, friction, nonlineae support, etc.) then calculate MNF. But there will be error margin -/+20%. Should experiment modal analysis is preffered? How the modal test results will be handle after that?

#2:
Process to measure displacements at some test point locations. But at which frequencies?

#3
Input measured displacements into CII model as Harmonic Displacements. And compare calculated stress with endurance limit of the material.

Looking for comments

Hello,

I would suggest to use harmonic analysis with pulsation forces as the input. Then you can scale up or down those pulsation forces to match the measured vibration at site at a given frequency. Make sure to check if there are stress concentration points such as small bore connections, etc. since these are the ones prone to fatigue failure due to vibration. You can request to check the vibration readings at these points.

But more expert advise is still required from some senior engineers. In our company there is one group specializing in this type of problem. They know what vibration measurement tools to use to have a more reliable results in your stress analysis.

As much as possible it's better to do some modifications in your piping system to have an acceptable vibration level than to prove in your stress analysis that fatigue stress is within the endurance limit since there are many uncertainties in the assumptions that will be made.

Any other opinion is highly appreciated.

Warm Regards,

Hi Borzki,

I agree with you to use pulsation forces (shaking forces) as the input. How to estimate the forces?

I'm thinking to review existing pulsation study as-built document to pick up forces for input.

Hi Van Ha,

First assume say 100N pulsation force and check your displacement output. Compare the displacement output to the measured vibration amplitude at site. Then you can scale up if say the displacement you've got is less than the measured vibration amplitude. Do this exercise until you match the vibration in the field. Also there are three parameters that you can vary to match what is observed at site. The damping ratio, the forcing frequency, & the induced force or displacement. Note that you don't input the measured vibration displacement directly into Caesar II. You can watch the Webinar presented by Mr. Dave regarding "Harmonic Analysis". You will find bountiful information there.

Any other opinion is highly appreciated.

Cheers!!!

Hi Borzki,

Thank you for your advise,

I want to clarify which frequency shall be considered to pick up displacement output to compare with measured data?

For example, the exciting freq is 7.3Hz, since the harmonic analysis will sweep through a range of freq and resonance may not be at 7.3Hz but let's say at 7.5Hz. My understanding is the adjustment is trying to let displacement output at 7.3Hz is same with measured data rather than 7.5Hz. Since there is big different of diplacement output from these two freq (disp at 7.5Hz > disp at 7.3Hz)

Hi Van Ha,

The frequency sweep is to take into account some uncertainties in the Caesar model such as support stiffness, etc. So it's better to use the maximum displacement at resonant mode and check the level of fatigue stress comparing it to the endurance limit of material. If I remember it right, Mr. Dave also explains other means to check the output from Caesar II other than what the program selected. Say if you have failed tee and you want to get the stress result there.

Any other opinion is greatly appreciated.

Cheers!!!