Posted By Julian Meer (at Work) Craig, Just wondering, what can be so scary about phase seperations at atmospheric pressure? Has someone lost a flammable layer as a result of leaving the valve open and perhaps the regulator is now wanting something concrete instead of relfecting on the actual frequency of this event and what other layers of protection could be brought? I did retro-fit an automated draining mechanism to a knock-out pot of which the microlosses from around the downstream reator PRV's would mount up. The aqueous bottom layer was sent to the effluent, but woe betide the effluent if the top organic layer went in. To limit - not avoid the oragnic layer effluent going in - an actuated valve was fitted with a timer (1 minute burst). Downstream of this block-valve was a ball valve plus another for switching to effluent or other. Operator would drain bottom layer in one minute bursts - an orifice plate kept the turbelence down (but was later successfully moded out) and the pressure in the vessel was maintained by nitrogen supply (with protective devices). As the operator anticpated the phase comming sep comming through the site glass through he would crack back on the ball valve. The advantage of this was if he did walk away only about 2 litres could go the wrong way. This solution did not prove unpopular (not that it could becuase they couldn't over-ride the timer) and didn't take much longer than a full manual op with its various human failings of letting the wrong layer go somewhere its shoudln't. If the valves are hard to operate then do you have large pipe diameters? If so what if a canonical reduction was fitted to enable a smaller valve could be fitted? More time for flow out but might make for a better seperation. I wont mention extension bars becuase the operators have probabley already made them. I am assuming you have an old plant not controlled off any DCS or PLC.