PID simulation of non linear system in Labview

[romiadam]

Hi all,

This is great forum, I always get here help. Hopefully I will get help again...

I am working on a school project, PID controlled heater system. It is not for real aplication, it is only a "jig". Just brief description:

Heater element 24V 40W is in the aluminum heat sink with blades (size 6cm x 6cm x 1cm block plus 5 blades height 3cm). The temperature is controlled by micro-controller (the PID controller is programmed in the micro-controller) by 19.5V PWM (power source used from old laptop).

My plan was that I will measure plant open loop transfer function (input step). Then fill this transfer function in to LabVIEW closed loop simulation with PID controller and setup optimal values for Kp, Ki and Kd. And then using those values for real control system and show that simulated parameters works for the real system (or at least it is close to the tuned real system)

I have got (measured) transfer function of the plant, and when I simulated it in LabVIEW the simulated transfer function was very similar to real measured transfer function. I was surprised.

The problem occurred, when I created closed loop system and added multiplication x5  before plant as Proportional controller and run the simulation. Set point was 80°.  The Steady state value in the simulation was 78°C. In my real control with micro-controller (P control was set to 3 too), the steady state value was only 72°C.

Why there is so big difference? I think 1-2 °C would be acceptable, but 6°C? If in P control I have 6°difference, how much it will be for PI or PID control? I have not tried it yet.

Is the heat sink nonlinear system? That could be the reason?

I am newbie in PID control and Labview simulation.

I know how to simulate linear systems, but never did non linear. Any suggestion?

Sorry for so many questions... 🙂

Thank You for help.

[GerdW]

Hi romiadam,

---

@romiadam wrote:

The problem occurred, when I created closed loop system and added multiplication x5  before plant as Proportional controller and run the simulation. Set point was 80°.  The Steady state value in the simulation was 78°C. In my real control with micro-controller (P control was set to 3 too), the steady state value was only 72°C.

Why there is so big difference? I think 1-2 °C would be acceptable, but 6°C? If in P control I have 6°difference, how much it will be for PI or PID control? I have not tried it yet.

Is the heat sink nonlinear system? That could be the reason?

I am newbie in PID control and Labview simulation.

I know how to simulate linear systems, but never did non linear. Any suggestion?

---

Have you any knowledge on (basic) control theory? Have you heard about properties of PID controls before?

A plain "only P-type" controller will never reach the setpoint! You need to add an I gain to get rid of the error term…

It really makes sense to learn some basics on the topic before trying to solve problems with software!

[romiadam]

Yes, I have the basic/moderate knowledge of control theory (on theory level).

Sorry I was not clear enough with my problem description.

My problem is that Yss (steady state value) of simulated plant in Labview was not close to Yss of the real controlled plant. I am not talking about 0 error. I know that P control always has an error.

With P control I only did first step in to Labview simulation. I wanted to see if Yss of simulated value will be "equal" to  Yss of real valu. I know, it will not be exactly equal, because my measured plant transfer function might not be very accurate and in real control there are external disturbances, but I was expected difference 1-2°C, not 6°C.

[romiadam]

Hi,

I am looking for manual / tutorial / lab / link, where I could  learn how to simulate discrete PID control of heater. I have already measured open loop transfer function of the heater (1-st order transfer function), I recalculated it in to Z-domain. I have transfer function of PID controller in z-domain. I only need some help how to build the block diagram including ZOH. I am not sure where exactly to place it.

I would like to simulate PID control of the heater in digital control, get the correct values of Kp, Ki, Kd, use it for my PID controller compare the output from digitally controlled heater with simulated model in Labview. How close it will be.

I am looking for exactly this:

https://forums.ni.com/t5/Example-Code/Construct-PID-in-Control-Design-and-Simulation-Module/ta-p/353...

but for digital control (transfer functions are in z-domain), not for continuous system

I am searching the internet for few days and still no luck.

Can you please help me?

Thank you.