time constant of typical control valve (urgent help)

HI Guys: I'm lookin for help specially from those who have been in industry. I'm trying to size valves for simulation purpose only,i.e. I'm not going to install the valve, just sizing them to match well with the whole model part of the concerned process. Following the sizing procedure that is available in the literature;I got the following: The flow cofficient of the water control valve is 58. Then if the selected valve is globe valve from fisher design ED of classes 300 with linear characteristics, then the size of the valve will be 2-inch. FoLLOWING THe same procedure for oil and gas phases I got the following sizes: 4-inch for oil and gas. Unfortunately I couldn't find on the fisher cataloque the time constant or response time of these valves. it is very essential in my simulation to include the dynamics of the control valves. So I'm kindly requesting from anyone who has experience with control valves to help me in the following: 1- what is the typical time constant for control valve? 2- The suitable time constants for the control valves with sizes mentioned above? thanx and waiting for help as soon as possible. regards assad.

For a small valve the 'time constant' (it's really more complex than that, and non - linear anyway) may be from a fraction of a second to a second or so. A large valve may go to 10's of seconds. However, it is unlikely to be a factor worth worrying about, since the process lag is likely to be far greater. So the valve 'time constant' will have only little influence on the overal system (Exclude flow control since that is not really the overal system). That's what I reckon, shoot me down if you like as I am probably wrong. Francis
  Quoted Text Here  

Assad,
A first order lag time constant is not really a very helpful approximation of a control valve even if you assume the simplest arrangement of a 4-20 mA signal going to an I/P mounted directly on the valve. It depends a lot on the capacity of the I/P, i.e. how rapidly it can add, or remove, air from the diaphragm. For large valve excursions there will be velocity limiting because of air supply limits. Sticktion and overshoot are not uncommon. Adding a positioner totally changes the dynamics by removing some of the nonlinear affects. Positioners can also destabilize fast loops such as flow.
Walter.
  Quoted Text Here  

The first thing you should do is determine how accurate your valve simulation needs to be. As alluded to by the other posts, if your process is significantly slower than the valve, then you can basically ignore the dynamics, but possibly not the nonlinearity aspects.
A 4 inch valve with a 'typical' pneumatic actuator and air supply I'd expect to look a bit like a time constant of a few seconds, a second or less if there's a booster and the assembly is designed for speed. If your process is of the same order of response time or faster, you'll need to take a closer look. Aspects like velocity limiting (particularly with positioners) can really change the way a fast loop behaves.
One other thing to note: if your process is fast enough for this to matter, then digital sampling times for the control processing elements may also be a substantial aspect to consider.
  Quoted Text Here  

Some information I've seen for small Valtek valves using positioners gives them a second-order response with about 0.8 damping and a period of 2-5 seconds. Velocity limiting is also likely to be a factor.
For a spring-diaphragm actuator, you need to get quite a lot of air into the actuator, and this will dominante the response. The capacity of the I/P to supply air will set the time constant in this case.
Bruce.
Nasser wrote:
  Quoted Text Here