Furnace temperature control


Thread Starter

Blanco, Felix

Hi List,

Could you help me on this please?

We have a temperature control (An outlet Temp PID cascading to a Fuel Gas Flow PID) in the Oil furnace/preheater, most of the time the control work OK but when it rained in the field the temperature of the incoming oil crude would lower significantly. The heater temperature control, would "wind up" on integral and cause the burners to go to max fire, 100% gas flow for extended periods. Maintaining max fire for extended periods would result in High High Bridge Wall Temperatures and a subsequent shutdown of the furnace.

The Process Eng. wants to use a limiter (55%) to limit the output of the gas flow controller to 55% and still maintain reasonable normal operation while avoiding the trip when the field experienced rain.

I don not like this solution because when the weather is sunny we would get a offset problem (it cannot reach it's setpoint).

I am considering to use a low controller output selector {pid.out = LS(master.pid.out; bridge.pid.out)} with a additional temp. controller and transmitter (bridge wall temp.) instead of the limiter. What do you think about this idea?.

Félix Blanco

Amr Elaguizy

Dear Felix,

A similar proplem we encountered in a Paper Mill and one of our electrician built him a little rain wheel from sheet metal, outside a building and whenever it rains the wheel will turn and sends a signal that it is raining outside which
was an input to a PLC and changed the logic accordingly. You might use that simple solution and set your 55% constrain only when it is raining outside.
Hope this might help
Amr Elaguizy

David Bergeron

Sounds good. Basically you want outlet temperature control with Bridge Wall Temperature override. The outlet of the Bridge Wall
Temperature controller and the Outlet Temperature Controller would go through a low select. The output of the low select goes to the setpoint of the fuel flow controller.

Under normal operations, the Outlet temperature controller would control fuel flow. If the Bridge wall temperature becomes higher than the setpoint of the Bridge Wall temperature controller, the Bridge Wall temperature conroller output would override Outlet temperature controller and close off the valve.

Make sure you feedback the output of the low selector to the reset feedback of the two temperature controllers to avoid windup in the
inactive controller.

David Bergeron, PE

Richard Dewees

IMHO these are the steps you should take if one doesn't work go to the next

1. I believe that the answer could be found in the tuning of the Temp loop. Temp is a slow changing process so I bet you could safely
use more derivitive in your tuning and reduce your overshoot.

2. If your PID controller has some form of anti-reset windup employ that.

3. put a temperature sensor on your incoming oil line and set it up in a feed-forward configuration so that the incoming temperature is factored in to the equation and apropriate adjustments are made

Tanweer Ahmed

Hi Blanco,
I had a similar experience at our plant for tin plating. We heat the plated sheet to melt the tin and flow vertically for a shining surface. For feed back of temperature we use a pyrometer. In order to protect the material from over heating resulting from a possible mall function or wrong alignment of pyrometer we added a current limiting algorithm in our program. Whenever the calculated reference current (in your case the fuel) is more than the defined limit (55%) the algorithm simply over writes the limit value to the calculated reference value.

Tanweer Ahmed
[email protected]

Geoff Manchester UK

> Hi Blanco,
This is almost certainly a tuning problem Most out of tune systems are ok most of the time!
The cascade master controller IAT must be longer than the slave IAT. Derivative action should act on the master on process changes only and not on set point changes. If the Integral is saturating and taking the system out of control, it is too fast! More derivative needed!

Paul Salessy COTEMS

Hi Blanco,

Try to tune your cascade loops with Graphidor that takes into account the constraints on the controller's output signal. We may send to you a demo version of the product to try it upon four loops. Our web site address is the following:
For sure Felix your loop is underdesigned versus influencing parameters and their quantification.
If your process is sufficiently constant,try this: no integral, adjust P to a reasonable offset (still close to desired). Then evaluate the corresponding mass flow of crude. Knowing this, crank open the by-pass crude, just near around 25% opening of the control valve. At this point, the upset from variation in temperature of the crude will be reduced. Then you might start quantify the correction. This assumes that the front end system has math capability.
Sounds like your real problem could be to maintain the oil viscosity within set limits.
Large diesel engines burning heavy fuel oil invariably use inline viscometers because of fuel oil quality variation... just one reason why temperature control is unsuitable for engines.
In burners it is more usual to use temperature but increasingly viscosity measurement is being used.
But whichever measurement is used, stable control is essential. Viscosity measurement response is around 2seconds which may be significantly faster than temperature response, particularly if installed in thermowells which means that the problem may get well out off hand before the control circuit can respond. Viscosity control enables very stable heating of the oil. This may give you not only a much better chance to avoid runaway heater control but also yields very siunbstantial savings in large plant operation e.g. power station, petrochemical ... in order to ensure complete combustion using temperature control excess oxygen has often to be increased to 4-4.5% with significant efficiency loss. The option to reduce excess oxygen results in increased soot blowing and increased chance of burning out a tube.
Solartron produce the ViscoMaster which is used in power stations petrochemical plant as well as for large diesel engines.
Visit www.solartron.com or contact via [email protected] (product not yet featured on website) or visit www.dieselpub.com and go to the catalogue section.