thePetroStreet

In ammonia plants that run on natural gas as feed, one of the major control parameter is molar ratio of oxygen to carbon. That ratio is generally known and controlled as steam to dry gas ratio. this ratio is controlled by manually adjusting flows of natural gas and steam. Does any one of you guys comment of controlling such ratio on auto?

Not on Ammonia plant, but at some refineries I have seen such auto controllers. It works just like the auto fuel to air ratio controller in any auxiliary boiler. Let me share the logic, if that you want.

I have developed such logic, with few changes. In general, fuel to air ratio controller consists of a master PID controller with Ratio as a setpoint, set of comparators and slave PIDs for fuel and air flow controllers. I switched the first PID to a calculation block, that calculates the requirement of gas and steam (in my case of Ammonia plant) and set-point is given to respective PIDs of gas and steam flow controllers. What you say about this? How this will impact the system response?

tosif, i couldn't get your point. Can you please share some logic diagrams (CLDs) or flow sheets that may help me and others to understand?

here, i draw it for your reference. Do let me know if you require any further information.

wao thats a complex one. Isn't it some what the same as of auto fuel to air ratio controller in boilers. As frontend consists of calculation blocks, don't you think about this calculations will go into infinite loop. i.e. from comparator to the PID setpoint, then form from PV to comparator and then with a new setpoint. May also result in hunting in any of the flows. Also what about the feedback value of "Ratio" as its seems only uni directional.
Have you experienced such loop in your plant? Why didn't you go with the conventional 'Auto Fuel to air ratio controller' scheme, along more PIDs?

o boy, is it a pen tablet or Ipod like thing? Anyhow I agree with dexter about the feedback of ratio value. Let me share you the air_fuel ratio controller, a typical one.
You can see that this logic is more or less the same as you have devised for your plant. Main set-point to comparator along respective PVs, resulting in set-points to respective PI controllers. Simple, isn't it? Yes there may be a chance of hunting due to some mismatch (resulting in jumping values with-in comparator). That can be resolved either fixing air (or steam in your case) a bit higher in all cases. Or even by introducing dead band in comparator or PIs.

By the way why are you using PID instead of PI controllers. I think PI controller will be best suitable in this scenario.