Hi,

I'm using S7-1200 for steam boiler level control. I'm not able to manage level control with one PID. I tried a lot of P and I tunes, but the best result I got is about a 7% error, which is not acceptable.

I wanted to try out cascade PID. I read about it, but I'm not sure I understand how it works. Another then level meter, I have steam pressure, steam flow, and water flow, so for the second PID, I can use some of these values.

For output values, I can only control the water-fill valve (must be used 3-Step PID, so I suppose it can only be used as second-Slave PID).

As I understood the theory, I can use two PIDs, and the setpoint for the second PID is output from the first PID. But what about the inputs of both PIDs, and the output of the second PID?

Really, any kind of help, comments, and tips will be very helpful.

Thank you in advance!

Hello,

maybe you look here: https://support.industry.siemens.com/forum/de/en/posts/cascade-pid-sample-program-anyone/147380

Or here: https://cache.industry.siemens.com/dl/files/819/103526819/att_993906/v1/103526819_PID_Temp_DOKU_V11_en.pdf

0_Einstein

Steam flow I'm assuming is the same and that no loads are stealing steam from the system (to produce water or via any path that doesn't end up in the boiler).

Steam flow is most of the time the same unless there is no need for greater steam, used for production, but that is not very often.

0_Einstein

The water-fill valve needs to make up for water to steam loss, but really, it just needs to maintain water level within a safe control band - correct?

 Absolutely correct.

0_Einstein

Is steam flow measured in mass flow rate, or something else?

What PID input (feedback) and output are you currently using, and what are you trying to control?  Motor speed or water throttle position or both?

Steam flow and water flow are measured in t/h (tons per hour) and most of the time, the relation between them is water flow = 1.1 x steam flow.

Currently, and only that I can control, is the valve for the water fill, and it is 3-step (via digital outputs). The valve has its potentiometer, but it is partly broken, and it is not showing the correct measurement (position), so my PID is without feedback.

All in all, I think that I don't have enough "components" to make some proper PID, but that is the reality that I can not change, so I must use what I have to make it better as it can be.

If that can be helpful, I have logs in excel where it can be seen how water level changes in relation to other measurements (water flow, steam flow, and steam pressure). Also, the current PID is configured as sluggish - P (about 0.8), I (about 120) and I'm not using the D component, to avoid turbulences.

Again, thank you very much for any help, tip, or advice... If there is a possibility for some kind of cascade PID, I really like to learn how to make it.

SCOTT123

Hello,

maybe you look here: https://support.industry.siemens.com/forum/de/en/posts/cascade-pid-sample-program-anyone/147380

Or here: https://cache.industry.siemens.com/dl/files/819/103526819/att_993906/v1/103526819_PID_Temp_DOKU_V11_en.pdf

 Thank you, I have read this article, but didn't realize how to implement it in my case...

By analyzing all variables that I have available for this process, I actually realized that I can not make cascade PID, as I'm missing the second "target" for control, and also I lack feedback (valve position).

Maybe I can then rephrase the question of this topic - except playing with PI(D) parameters, what can I else try to stabilize the system? The process overall must be slow, but obviously in some cases must react faster, so I don't have any idea how to accomplish that.

To make that PI components are not fixed, and change them depending on, for example, steam flow? So when I detect that flow is increasing, I can increase P and decrease I?

Just giving an example...

Again, thank you in advance!