There is a vibration problem in the piping close to venturi. I think vortex shedding is the vibraton source. Propane gas flows thru venturi. Does anyone know how to calculate the vibration force due to vortex shedding? So I can enter it into caesar2 harmonic analysis.

Using harmonic analysis in CAESAR II to solve vibration problems, your goal is typically to simulate the observed problem so that you can:
1)determine the extent of the problem (eg. stress, forces), and
2)if a problem exists, try to solve the problem on the comuputer and then apply the solution in the field.

Harmonic analysis requires an excitation frequency and a magnitude (in CAESAR II, the magnitude is either force or displacement. In the case of flow-induced vibration, you want to use the Force input, as you have noted).
There are several force/excitation frequency combinations that will generate the same system response, and ideally, you want to get as close to the correct combination as possible (i.e. the correct excitation frequency, and the correct force magnitude).
If you measure the frequency at which the observed vibration occurs, this will correspond to the excitation frequency. This will allow you to apply various force magnitudes at this frequency, scaling up or down until you observe the same system response on the computer as is observed in the field.
If you do not know the excitation frequency, a common calculation predicts a range of 0.2*V/D to 0.3*V/D, (ft/sec * ft, or in/sec & in units). V=fluid velocity, D=venturi dia.
Outside of expensive software simulations, I don't know of a method of calculating the force, but I would suggest that you guess at what seems like a reasonable force magnitude, and apply a range of excitation frequencies (per the above equation), seeking the same system response as in the field.
Perhaps you can restrain the point of application of the load and somehow measure the force applied to the restraint. This may give you a good estimate of the magnitude of the force, but I don't think it is necessary.

Long answer to your simple question; should give you a good start.

Note 1: Vortex shedding is normally considered to be caused by flow around an object (such as wind flowing around a pipe), and the excitation frequency for this is typically estimated by 0.18*V/D to 0.2*V/D.

Note 2: Try to make your model as accurate as possible, stiffness and mass-wise. Extra nodes should be added along straight runs of pipe to help CAESAR II calculate more accurate natural frequencies. This can be done quickly using the Break feature.

Note 3: Non-linear effects (such as opening and closing of a restraint gap during a dynamic event) is not considered by pipe stress programs such as CAESAR II.