It seems that you may be encountering some form of pressure surges in the main pipework. Approximately 40% of safety relief valves in liquid lines Susceptible to surges exhibited unstable operation. This is a recognised problem which is often solved with a true liquid surge releif valve.
Generally the chatter is caused by a series of steps:
- the PSV set pressure is exceed, but the intial slow movement of the valve results in an over pressure in the main pipework (relief valves are also sized with a minimum of 10% overpressure).
- the losses in the inlet pipework and the valves accumulation also allow the pressure in the main pipework to rise further
- the growing flow capacity as the valves opens, the losses in the inlet pipework and the momentum of the opening valve usually result in a fall in pressure at the relief valve inlet and the valve starts to close
- once the valve has closed the flow towards the valve impinges on the closed valve and the pressure rises rapidly - restarting the cycle.
The main factors which make the chatter worse are;
- the wave travel time form the main pipework to the relief valve. The slower the wave speed generally the longer the chatter cycle. The wave speed is a function of pipewall material, pipe diameter and wall thickness, and also to some extent dependant upon fluid properties
- distance from the the main pipework to the relief. The longer the distance the worse the chatter and overshoot
- the diameter of the feed pipework. High velocities associated with a smaller diameter make the surge pressure and hence the chatter far worse
- the losses (bends, tees etc) in the feed pipework. Increased losses tend to affect the chatter amplitude
- the valve design, large inertia componets like the plug/disc tend to cause overshoot and chatter, the effects of the fluid jet impinging on the plug/disc may also lead to asymmetric operation
- vapour pockets can be formed downstream of the valve as it snaps shuts, which then help acceleration of the fluid once the valve re-opens
A far better solution is to use a purpose designed liquid surge relief valve, like the Mokveld (
pilot operated surge relief valve. It exhibits a faster reponse than common PSV's and modulate stably around the required relief flow. However even this sort of valve could exhibit some instabiity if the distance from the main pipework is inordinately long.
Former Du Pont de Nemours conducted relief valve chatter testing. The study is documented in PVP-Vol 237-2, Seismic engineering - Volume 2 ASME 1992 (attachment 6). The document title 'RELIEF VALVE CHATTER TESTING'.
Hope this helps.
RobV