With two phase flow, you have a spectrum of flow conditions, based on relative velocities.

1. Mist flow = "high" gas rate, "low" liquid rate
2. Annular flow = "high" gas rate, "high" liquid rate
3. Stratified flow = "low" gas rate, "low" liquid rate, no vertical sections in the line
4. Bubbly flow = "high" liquid rate, "low" gas rate
5. Slug/plug flow = Stratified flow with vertical sections, or at "medium" combinations.

For Mist and Bubbly flow, there are generally no "two phase flow" concerns. They can largely be treated as Gas and Liquid systems accordingly, with the understanding that you can still end up with erosion or cavitation issues. However, I do not classify this as a "two phase flow" problem - just a linespeed problem.

For (steady) stratified flow, there generally should not be any perturbations. This is analogous to the pipes and culverts under the streets carrying rain water.

For annular flow, my gut feel are that vibrations are going to be proportional to the dynamic pressure of the system - i.e. 1/2 density times velocity squared. In a perfectly annular system, you would have a perfectly hollow cylinder of liquid along the walls of the pipe, and a gas cylinder down the annulus of the pipe. The less perfect this flow, the more turbulent waves exist along the radius of the interface. But ultimately, annular flow results in random noise, which I've seen typically dealt with by anchoring the pipe down, everywhere.

I've never seen an empirical method for estimating slug/plug flow size and velocity, though my understanding is that there exist software packages that purport to do so. It would be this impulse load and only this impulse load I would consider designing against, and therefore adding snubbers for.