Air Flow Switch for Glycol Regeneration System

[q3s]

I am engineer responsible for an offshore installation that processes natural gas. We operate a TEG glycol regeneration system with a gas burner to regenerate the glycol. The burner was originally 'normally aspirated'. Some years ago we replaced the burner, and the controls, and this included a new 'forced draft'system (with duty and standby fans). The control system redesign incorporated a pressure switch to detect that the fans were working. But we were generating so little 'head' that the switch would oftem trip and shutdown the system. We changed the switch and put in a dp switch whereby we detected the pressure upstream and downstream of the fans. The results were no better, and the system oftem tripped on start up (it would detect flow sometime, often not though). Then we installed a heat type flow switch (where you have a heated probe and a non haeted probe and the resistance of the heated probe changes as you get air flows across it). The results were no better (still unreliable). The current 'solution' is another dp switch (very low range) measuring dp across the flame arrestor mesh in the burner inlet box (whwre the mesh acts like an orifice plate). It works, but not always.


So I'm looking for a switch, EExd, that will operate (reliably) to detect flow. The duct is 18", the theoritical flow is 2ft/sec. Can anyone suggest a device.

 

[TD2K]

What about using the amps to the fan motor?

 

[q3s]

Thought about either measuring the fan revs or the fan power, but ruled it out for the following reasons. There are dampers on both air inlet ducts (duty and standby fan). These will close when the co2 fire supression system is activated. So under normal circumstances the damper is closed on the standbt fan duct. So we could me measuring fams revs or power and the damper could be closed, thus giving a no-flow scenario.

 

[OilToil]

I ran into almost the same problem, except our issue was that the air flow switch had to clear to allow the unit to start, and the fan just couldn't make enough pressure to trip the DP switch into showing air flow, and the startup sequence would abort. The operators would go out and pull the low pressure side out of the inlet duct and suck on it to clear the trip... you can guess how well that went over! We changed the tubing on the high pressure side (downstream of the blower) so that the high pressure side of the dp switch was tubed so that the inlet to tubing was pointing directly at the fan (like a pitot tube), but that did not help much.

We talked to the heater manufacturer and they said this is a common problem (so why do they design them like that?!) and they suggested replacing the DP switch with a physical flag-type switch, basically a 'wind vane' on a stick. Google "flag switch air flow" to see the kind of things I am talking about. I left that plant before we installed the new switch, and we disabled the DP and just looked at the blower motor status.