multi effect evaporator control strategy 1

[CARF]

Dear all,

PS. Brix = sugar concentration

In our tomato paste plant we have a huge triple effect evaporator unit in order to concentrate tomato juice to thick tomato paste (=evaporate the water out of the juice).

The current control works through controlling input Brix changes to a fixed Brix at the outlet by changing the main product flow rate. Altough it works quite well, a disantage is that you change the main flow all the time. Changing the evaporator steam pressure in order to compensate for input Brix variations is to slow for fast fluctuations.

Does anybody have experience with this?
Better control strategies? Feedforward, etc. etc.?
Good evaporator control schematics examles?

Thanks for sharing knowledge!

 

[GBurns]

I have seen control strategies where the steam flow setpoint is set based on the amount of water to evaporate. These evaporators have a varying amount of solids (suspended and dissolved) and the amount of steam is changed based on flow rate and percent solids. This would be a course adjustment. The steam flow is further ratioed to control the final solids to a setpoint. This would be the trim or fine adjustment. These controls will react quickly to flow rate changes and via the trim controls will fine tune the steam flow over the long term. The steam flow is a cascaded loop to the 1st effect pressure controller. Pressure altered to increase/decrease steam flow.

 

[CARF]

Dear GBurns,

I think a have the idea. So you want to do rough adjustment through a sort of feedforward based on the incoming flow and composition of it. This feedforward should adjust the setpoint of the fine-tuning feedbackloop which controls the outgoing flow more precisely.

Did I understand you correctly?

Well thanks, good idea, any comments ?

 

[hacksaw]

Your steam controls sound fine, but your evaportator level controls need to be configured to allow for inlet flow surges instead of setpoint control.

 

[CARF]

I investigated it a bit more and want to add this:

Dynamic feedforward. With this ratio setpoint system, however the product is initially more concentrated when the product in-feed rate is increased. This happens because the ratio system increases steam before the increase in-feed works through the evaporator. A dynamic compensation times the increase in steam rate to nearly coincide with the response to the in-feed increase. With no dynamic compensation, the steam is changed as soon as the feed is changed, and could result in an upset that is actually worse than obtain form an ordinary feedback system.

At this moment I making a dynamic model in Simulink to investigate the proposed control strategy in more detail.

ThX,
MVD

 

[GBurns]

The control strategy I suggested was just the basic strategy used. In some similar strategies I have seen integral only controllers utilized and some DCS vendors permit you to have gain but set them for integral action on setpoint changes. This would prevent the sudden jump in steam flow with the product flow change. Also, you could employ setpoint ramping based on the flow rate and retention time in your evaporators. Much of this depends on what you have for controls (single loop, DCS etc).

At a minimum you would need:
steam flow controller,
steam pressure controller,
outlet consistency/solids controller
flow rate into evaps
consistency/solids into evaps
consistency/solids out of evaps
steam flow
steam pressure

The steam flow setpoint is determined by the calculated amount of water to evaporate from the product based on the inlet flow and consistency/solids. This setpoint can adjusted (fine tuned) by the output of the outlet consistency/solids controller. The output of the steam flow controller controls the setpoint to the steam pressure controller which currently exists.

This is only a suggested control strategy and I personally have never seen it work on tomato juice.