When describing allowable leakage, the standard refers to ".01% of rated valve capacity". Rated valve capacity is defined as "the quantity of test fluid that would pass through the valve at rated travel under the stated pressure conditions as defined by the appropriate equations and manufacturers ratings."

This reads like something the government would come up with. We want to quantify allowable leakage for a metal/metal seated ball valve to meet Class IV. If I assume (yes, I know the repurcussions of that..) that "rated valve capacity" is Cv, then I can have .01% of Cv/min. and meet Class IV.

FCI 70-2 specifies leakage rates on a test stand with when tested according to the FCI procedure (usually 50 psi air at ambient temperature)

The FCI leakage classes are a prety good indication of how carefully the valve parts were fitted together.

How much leakage will you get in service at, say, 450 degrees, steam, at 200 psi?  Anybody's guess.  Aside from the class V procedure which tests for leakage at service pressure, the ANSI classes are not directly related to the expected leakrate in service.  

Some valve types (pilot-balanced globe valves) shut off tighter at elevated differential pressures.  Other valve types (unbalanced globe) require additional actuator force proportional to diff. pressure.  Eccentric rotary valves may be installed with the flow tending to unseat the plug or with the flow tending to add additional seat load. Guess which one shuts off tighter.   And of course the odd bit of weld slag coming down the pipe three minutes after you commission the valve will tend to change everything anyway.

MORE NEWS