set "PID parameters" for power serie of y = 3E-40*x^18.94

[john111smith]

Hi,

How can I set "PID parameters" for the control of the system with the following "power serie" ?

y = 3E-40*x^18.94

X min = 30
X max = 90


PID parameters:
iMax,iMin,
pState,istate,dState,
pGain,iGain,dGain,
MAX,MIN


regards

 

[Skogsgurra]

If I understand you right, you have an extremely non-linear plant there. Do not expect to use one single setting if you are working over more than a few percent of x.

What plant has such an extreme y=f(x)?

Going from x=30 to x=90 means zillions of orders of magnitude change in y. Are you sure you got that relation between x and y?

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[john111smith]

58.9 1.00E-06
68.8 1.00E-05
73.7 8.90E-05
78.6 2.00E-04
83.5 6.40E-04

are some of points.
at main working area X is between 60 and 80

 

[itsmoked]

skogs is correct. You will probably need a a controller that changes the PID constants for every change in set point. An adaptive PID controller.

PH?
This a chemical reaction?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[john111smith]

This is a valve

 

[itsmoked]

Ah! Yep.. Nasty non-linear.

You can often map the valve if things permit then run via a table that takes into account the actual flows that 'will' occur.

Or model the valve and run with the model in the 'plant'.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[john111smith]

thanks
but it will not work with LOOK-UP TABLE
- input is voltage for gas paffing valve control
- turbo vaccum pump run continuously
- output is vaccum

 

[m777182]

Your datapoints seem a part of a titration curve around the inflection poit.Can't you afford a trasformation y--->log and get a (nearly) linear function log=Ax+B?
m777182