Simulating Transient Pressure Pulses in Venturi Model

[brandon.hallinan]

Hi all,

I am setting up a simulation for testing an intake venturi. Of the conditions we are testing for, is that the structure can withstand positive pressure forces, vacuum pressures, and transient pressure pulses.

The first two types of pressure to test for are simple enough to setup, however, I am unsure how to setup the simulation to test for the transient pressure pulses. 

Verbatim from the requirements specifications,

"The Venturi shall be capable of withstanding transient pressure pulses of +44.3 inches water (gauge) to -13.6 inches water (gauge)."

My initial thoughts on this is to set up a load curve where I flip back and forth between the two values, maybe once per second, over a time frame of a minute.

As a note, I am given the maximum mass flow rate through the Venturi and the nominal temperature, but I don't think I need to use those pieces of information for this particular simulation.

Would my methodology for simulating the pressure pulses work, or do I need to look at this another way?

Thanks in advance!

[markdeckerZBQL7]

In a venturi, the pressures will vary across it, with the lowest relative pressures occurring in the throat of the venturi.  That is why you may need to consider the mass flow.  Otherwise, you may just want to assume that the pressures are uniform everywhere.

You are on the right track in using a load curve to vary the pressure over time.  However, the missing component here is the length of the pulse.  The venturi may react differently to a pulse that occurs in 0.01s versus one that spans 1.0s.

[brandon.hallinan]

Thank you for the reply!

Yes, for the other two requirements of the analysis, 

"The entrance and exit sections of the Venturi shall be capable of withstanding an operating depression of 30 inches water (gauge) and a positive pressure of 30 inches water (gauge)."

and

"In all other sections, namely the conical convergence section, the cylindrical throat, and the conical divergent section, the Venturi shall be capable of withstanding a positive pressure of 30 inches water (gauge) and a vacuum pressure of 80 inches water (gauge)."

I set up two case studies; one with the positive pressure applied normal to all internal surfaces of the Venturi, acting outward, and one with the negative pressure, or vacuum, applied normal to all internal surfaces of the Venturi, acting inward toward the center axis.

After thinking more on the transient pressure pulse portion of the analysis, I am thinking that this will actually need to be done in CFD, and that I need more information regarding the frequency of the pulses, as you mentioned in your response.

I imagine that when run in CFD, I will need to incorporate the results back into the mechanical simulation of the model?

[KubliJ]

Hi Brandon,

It is possible to do a transient analysis of a venturi in CFD and use the results in Sim Mech.  However it will be for a single set of results and it would be more or less a snap shot of the pressure applied in sort of a static method in Mech.

If you are needing to consider this as a transient response, especially if pulses are approaching natural frequencies.  Though, thinking about it.  Not sure how important that will be with a venturi versus an orifice plate.

Either way, MES should be able to handle it.  I would recommend that if the model is axisymmetric that you model it as such.  It will save on mesh and analysis time.  Even if it isn't, since this is a new setup (an experimental setup) I would play around with it as an axisymmetric model until you feel comfortable that you have the analysis setup correctly.  It will be quicker and easier to do a mesh convergence on the axisymmetric that you can then apply to a 3D model.

There is linear transient, but I can't recall or find any information on if that supports axisymmetric.

You may still want to model as 3D, see what natural frequency modes there are and their shapes.  Compare that to the frequencies that you are expecting the pressure loads to oscillate at.  Some modes are not axisymmetric and may rule that method out that element type if you are operating near them.

Hope that helps.

James