Split-Range Valves model with Simulink

J

Thread Starter

Jerome

Hello,

I'm currently studying PID controls and more particulary Split-Range application.

I know the theory of split-range with valves but I haven't been able to create a model with Simulink. So if someone would a simple example of a split range application on simulink, could you please share it?

I just need something very simple, it is jut to understand how it works.

Thanks a lot for your help and support,
best regards,
Jerome
 
If you understand how a split range application works you should not have a problem. This might help.

Split range control is done with two valves installed in parallel in a process line. The one valve that is installed in line with the process line is a small valve and the other one on the bypass line can be a identical small valve or even be twice the size of the small in line valve, depending on the application you want to use it on.
The main purpose of a split range application is to handle big upsets in the process. In some cases these upsets are not much, but bigger than what the small valve can handle by itself and bringing in the second valve is enough to handle the upset but when there are at time a sudden massive increase in demand, a valve twice the size of the small in line valve might be needed.

Another reason why we use a split range application is to do precise control with the small valve alone when there are not much demand. If you only had one big valve installed it will mean that you have to run the valve on one or two percent open all the time to control the normal process and that is obviously not the ideal for precise PID control. Ideally you want any control valve to be open between 40 to 80% during normal operations. In that region you will be able to set your PID settings properly and will be able to do precise control.
For this reason we sometime move over to use split range control since during normal operations we can then do precise and stable control and we can also handle any upsets or increase in demand that might occur in the process from time to time.

The way we do it is to take the 4 to 20mA signal from the PID controller and connect it to both valve's positioners. We then re-calibrate the small valve's positioner to go from fully close to fully open with the first half of the signal and the second valve's positioner to then work with only the second half of the signal.
 
J

Jerome MAZET

Hello,

Thanks a lot for your clear explanations. I was more thinking of the other split range application were the two valves have antagonist behaviours. For example, let's take a tank with an outlet valve and an inlet valve, and I want to control the level.

So, with a split range architecture I can "play" with both valves.
I've tried to do a simulink model of this but I don t see any benefit of this architecture compares to a simple one where I can only play with the inlet valve.

To my mind the split range should be more powerful than the classic one, isn 't it?

Thanks for your help and your time spent to help me :)),

Regards,
Jerome
 
H

Hans H. Eder

One application of split range has been described already. Another widely found one is to steer two valves in different services, like e.g. in the case of batch reactor temperature control: One valve is in the heating medium the other one in the cooling medium.

Split-range looks like an elegant solution at first glance but a common problem is the tuning of the controller since the two valves exhibit different effects (process gains) and also different dynamics w.r.t. the controlled variable. One solution for the above described configuration is to steer e.g. the small valve for “normal” control and to provide a second controller that keeps the opening of the small one within the desired range by steering the large valve. Both configurations are provided in our process control toolkit TOPAS (www.act-control.com/software.html) for learning, tuning and trouble-shooting.

 
I agree, there is no point setting up a level control like that.
What we do have some times is two identical valves working from the same signal but the one is setup as direct and the other as a revers acting. That is pretty much what you, and others - incorrectly, refering to as split range and what you were looking at on your level control application.

In a case like this, or in the case of a true split range application, you cannot realy play with anything since in the true split range application the second valve will only start to react once the first valve is fully open and in the second case the two valve will work in direct opposite of each other based on the input signal.

In both cases you only have one PID controller and one set of PID tuning parameter that you can play with that is why it is always advisable to use identical valves when you do any kind of split range application since that makes the PID tuning easier, but like I said in my previous mail you can also install a bigger valve as your second bypass valve but the control will not be as precise and more unstable as with identical valves due to the fact that you will have to find a happy medium tuning PID tuning set that will not be ideal for either valve.

One of the biggest problems with a split range like the one you are refering to is that the valves tend to fight each other if your tuning set is not set correctly so doing a split range is one thing, doing the final PID tuning is something to consider before you start your modification.

Anyway besides the problems that you could encounter to do the tuning, I for one, still believe the split range to be a extremly powerfull solution to a wide veriaty of control problems and have work on and seen first hand how well it can work.
 
J

Jerome MAZET

Hello All,

Thanks for your comments. I understand now the split range application. I've found this example into a training course and I agree with you all that this not the best way to implement a split range architecture.
So under simulink I've implemented a classic "hot cold" control to simulate a control temperature into a jacketed heater and it works fine.
So once again I would like to thank you all for your time.

Regards,
Jerome
 
B

Bhanwar Singh

Hey Jerome,

I am dealing with a similar kind of problem. I have to control the tank temp and I am using split-range controller through which I can control the coolant flow rate and feed flow rate of the tank. as I am new to this problem, can u Please tell me any simulink simple example for this, that hoe to divide the controller signal to two different values i.e 4-29 mA in to 4-12 and 12 to 20. thnx in advance. plzz reply.

<b>moderator's note:</b> Not sure what word "hoe" is in the phrase "hoe to divide." So I left it.
 
Top