My boss asked me if theres going to be a vibration problem in a piping configuration.

I have a model and passed static test.

on dynamic/harmonic submenu, I'm lost as to how to proceed. I know the piping is connected to
a 3600 rpm compressor screw compressor. would frequency range be 0 - 60Hz(3600rpm)? what do I look for in the CAESAR printout? Would there be
cyclic stresses on each nodes?

The first pass check I always do in this type of problem is a modal analysis and look at the lowest natural frequency (lnf). There are some norms to give you a feel for risk of vibration in the piping system.

 If nlf less than 1 Hz then risk of large amplitude piping vibration very high

 If nlf less than 5 Hz then risk of large amplitude piping vibration is unlikely unless there is a source of vibration such as rotating equipment.

 If nlf greater than 10 Hz then risk of large amplitude piping vibration is unlikely unless exciting frequency couples with high energy vibration source.

This is a very cost/manhour effective start to any dynamic problem and a good start before jumping into more complex dynamic analysis.

Hope this helps

I've done modal analysis per application guide with eigensolver on and got 5 modes ranging 1 hz to 5 hz. Each mode is different a different direction when you animate it. I guess you look which is the most probable mode for your system and take that as your piping frequency. Next step I took was to simulate machinery vibration using Harmonic analysis, excite the system at 60Hz (3600 rpm) and assume harmonic displacement on compressor discharge connections (since I have no data).

Nowhere in the input where it asked me for the piping frequency i'd like to use but it produced some results and code stresses anyway which is a pass.

Maybe another way to look at this is from the modal frequency results, since max is 5 hz, and machinery runs at 60 hz, we can say there will be no resonance (except when starting the motors).

Jaime;

Using CAESAR, you cannot PREDICT harmonic response, you can only trouble-shoot by simulating a response that you may encounter in the field.

There are several reasons for this:

1) CAESAR does not calculate the EXACT natural frequency, only approximates based on the mass-spring model. We need an infinite number of nodes to generate the exact natural frequencies.

2) You most likely do not know the EXACT forcing function (i.e. what is the magnitude & forcing frequency)

3) Your mathematical computer-approximation will differ from actual installation. For example, exact restraint locations and stiffness will not be the same in reality as in your computer model. Mass may vary as well - insulation and fluid density EXACTLY correct? Is the mass of attached instrumentation or any other attachments considered? Structural steel can contribute mass, where you may model it as a simple restraint (stiffness, no mass).

4) The actual harmonic response curve is VERY narrow (vertical axis = response, horizontal axis = frequency). This means that even very small alterations of calculated natural frequency and/or forcing frequency can give dramatically different predicted system response.

All of this means that the variety and magnitude of differences between your computer model and the real system are far too great to enable prediction with CAESAR (or any other beam-element computer program).

The Harmonic analysis feature in CAESAR is intended as a tool for evaluating field vibrations that already exist. That is, you can fine-tune your model and forcing function such that the results on the computer simulate the vibration problem in the field. Once you simulate the "shape" (displacement solution) you can answer the question: "Is there a problem?". If it is a problem, you can modify the model through changes in stiffness (restraint) or mass, run the analysis with the same excitation, and see what the new system response is, thereby evaluating possible solutions to the problem.
There are times when changing the forcing function can be a solution (eg. slow the compressor), but this is not typically the preferred solution to such problems.

In advance of construction 'Analog studies' can be performed to examine acoustic and other characteristics of a system to help guide design to avoid vibration problems.
There are also basic guidelines that some follow, such as those provided above.

This may be a good time to review relevant pages of sections 4 & 5 of your pipe stress seminar course notes.

Regards,
Jim.

Jaime,
Your modal analysis has only listed the first 5 natural frequencies (eigenvalues). You can change your input to list more by setting the max no. of eigenvalues to be solved (which are in order from the lowest) and/or by setting the max cut off frequency to be solved.
The lowest natural frequency is termed the fundamental frequency and if this is very low as yours is you will get problems when connected to rotating or reciprocating equipment. With refining the restraint system and further analysis and you should aim for a much higher fundamental frequency (at least above 5Hz preferably above 10Hz as my mate Nigel suggested) and for the modal frequencies to be away from your exciting frequency ie. 60Hz.
As Jim has listed above, there are many factors to consider such as restraint stiffnesses and at best your model will be an approximation of this.
I would suggest that you either go on a dynamic analysis course or leave the dynamic analysis to someone who has experience in this field, it cannot be learned by looking at a few notes.

Yes I took the course with Jim Wilcox in May 2001. Problem I have is I'm not a full time pipe stress analyst and this is the first dynamics model analyzed and first analysis for this year.

Jim, I guess we can also model statically just like in a static seismic analysis.

I have worked with many HMI softwares and whenever I get stuck I found discussion forums helpful in finding solutions. Thank you for all your suggestions.