# Help with liquid level control system

Thread Starter

#### yunus27

Hi All,

I have been tasked with designing a PI liquid level control system (gravity fed with a valve on the inlet and outlet) and have experienced some problems in regard to the literature I am using, (Process Dynamics and Control, 3rd Edition).

I have been tasked with finding the values of the tuning parameters Kc and ti using the direct synthesis method. When doing so, should I be setting theta (time delay) to 0?

Also, I am confused about the nature of T (tau). I know what Tc, Ti, and Td refer to, but I am so baffled by the T (tau) on the numerator of the below equation as it has no subscript. Also, if Ti = T then why would my teacher ask for Ti to be calculated as well as Kc?

Apologies, this is my first post on a forum. I have spent many days running in circles with this issue. If anyone can help I would be so grateful.

#### ControlsGuy25

Hi All,

I have been tasked with designing a PI liquid level control system (gravity fed with a valve on the inlet and outlet) and have experienced some problems in regard to the literature I am using, (Process Dynamics and Control, 3rd Edition).

I have been tasked with finding the values of the tuning parameters Kc and ti using the direct synthesis method. When doing so, should I be setting theta (time delay) to 0?

Also, I am confused about the nature of T (tau). I know what Tc, Ti, and Td refer to, but I am so baffled by the T (tau) on the numerator of the below equation as it has no subscript. Also, if Ti = T then why would my teacher ask for Ti to be calculated as well as Kc?

View attachment 54

Apologies, this is my first post on a forum. I have spent many days running in circles with this issue. If anyone can help I would be so grateful.
Hi Yunus 27

I woul suggest you to have a look on this tutorial video for PI controller tuning for Tank level control:

Hope this can be an help

James

#### relaxc

If it is not too late.
Your paper give the coefficients of Skogestad ( SIMC), and he does not use exactly the structure of a classic PID.
Theta is the delay and you have choose tauc=theta. and if theta=0 then SIMC does not work well. and finally SIMC work "well" ( relaxc better) when the delay ( dead time ) is close to the time constant of your process. That it why i recommend that you use relaxc ( trial version) https://www.appedge.com/pdffile/relaxc_expertcontrol.pdf if you work with expert control ( unity pro)
relaxc work in all the case and give the best secure response that your process can do.
and if you use relaxc at your school, it will be not possible for your teacher to better.
Ask me if you dont arrive to download relaxc ( inside you will have the full documentation to control all the kinds of process)

Similar threads