Process models for temperature and pressure

I am planning to simulate a temperature loop and a pressure loop using PIDE functions blocks in ControlLogix. I used a dead time + first order lag to simulate temperature and it works fine. ( is this the right way for temperature loop ? ) Pressure response is faster than temperature response, but not sure how to make a process model. Anybody there to give some inputs ? ( Peter Nachtwey , are you reading this ? )

Yes, I read all the posts with topics like this

Pressure can be very difficult. I can do it but I use a DSP or do it off line using an Excel spread sheet. Pressure is NOT linear. One can not just whip out a difference equation and to simulate pressure accurately. Simulating fluids is a problem because of the square root in the equation that equates a difference in pressure to flow. I must use a system of first order differential equations to do my simulations. I use Excel and each row represents 10 microseconds of time and each column either calculates a change in a pressure or flow or adds the changes and checks limits. This works OK but to do it right one should use Runge Kutta to integrate the differential equations. We are working on the windows version now. It works but the calculation need to be changed over to use the Runge Kutta method if integration and I need to add a few more differential equations to make our hydraulic simulator very accurate.

Temperature is easy. You need to calculate the time constants and a a gain and the difference equation comes easy.

Temperature Simulator and PID using Ron Beaufort's 'Hot Rod' system from last summer.

Here is that hard part. How many time constants are there? How do you measure them? All I need is data from a RSLogix trend or from other source that is a record of the PV and CV as a function of time.
Ron provided the gains and time constants for his system.

If you can wait a bit I have been working on a uLogix simulator for PID control like the one I did for SMC control last summer.

Thanks for your comments Nachtwey.

How many time constants are there? How do you measure them?

Click to expand...

I am not sure about the time constants. All I want to do is to create a PID function in ControlLogix , Give SP,PV and CV tags and simulate the PV using CV. In the actual system PV is read by a prerssure transmitter which is fitted in an air chamber and CV is given to a VFD controlling a blower. Will a second order lag can do this function?

WHY DO I HAVE TO ASK?

iknowsomeplc said:

Will a second order lag can do this function?

Click to expand...

How am I supposed to know? I have know idea what you are trying to do model or control. Chances are NO. For hydraulic cylinders controlled by a servo valve I use a 3rd order system. A hydraulic cylinder can be modeled by a second order system with two imaginary poles and the valve can be modeled by first order system. This works if you just want to play around. A better way is to use a separate PC or DSP board solve a system of first order non-linear differential equations in real time. Best is to have a real system. Nothing beats testing on the real thing.

Even a control logix will have problems controlling pressure of a hydraulic cylinder. A pneumatic system can be controlled because the bulk modulus is lower. ( WHAT IS BULK MODULUS? ).