Laplace Transforms in Control Engineering

Hi Peter,

I'm back again. I have been working away at control theory and am happy with my understanding.

I used Matlab to convert my laplace transform to a z transform.

0.00375 X= --------------------- (s+0.15)*(s+0.25)

Which converts it to

0.000438*z + 0.000410 -------------------------- z^2-1.81*z+0.818

then the difference equation is got by dividing the num and dem them by the highest order z

0.000438 + 0.000410z^-1

-----------------------------

1-1.81*z^-1+0.818*z^-2

therefore

x(n)-1.81*x(n-1)+0.818*x(n-2) = 0.000438*u(n)+0.0004104*u(n-1)

therefore

x(n) = 1.81*x(n-1)-0.818*x(n-2) +0.000438*u(n)+0.0004104*u(n-1)

Which is a little different to your result of

x(n) = 1.81*x(n-1)-.819*x(n-2)+.0004387*u(n-1)+.0004104*x(n-2)

Have I got lost somewhere again?

Reply to
Ann
Loading thread data ...

No,I think you are on your way now. However, u(n) cannot affect x(n) as you have indicated in your equation. u(n) can only affect the following positions at x(n-1) and later. Think about it. You have neglected the ZOH ( zero order hold ). If you let us know how you found your answer we can help you with adding the ZOH. However, what you have should be good enough to start testing your fuzzy logic controller. You should add the ZOH to your simulator.

Peter Nachtwey

Reply to
Peter Nachtwey

zzzzzzZZZZZZZZZ

>
Reply to
kieran.jackson

Results of continuous system approximation in discret time domain depends on choosen sample time and method of approximation.

When you take other sampling time you get completely new Z-transform. The same about changing method.

One approach of aprroximation is to get exact time response. ( Same values as time response of continuous system -exact at sampling moments ofcourse). Next approach is to get exact frequency domain response.

You for almost sure are interested in time response maping. We can use step equivalent method. Remember that ideal impulsating device doesn't exist in real life. Read about Zero Order Holder (ZOH, thats what A/D devices do).

Reply to
Mikolaj

Dnia Thu, 11 May 2006 09:19:33 +0200, Mikolaj napisa³:

I meant D/A of course that turns numbers into energy with "sample and hold" method most often.

Reply to
Mikolaj

There is usually a method given in any undergrad control textbook

From an earlier post there are 3 systems using zoh (with various sample times) - which just means that the o/p from your controller is a constant between outputted values - which i assume you want.

the final difference equation should include a z^-q delay in the controller output (as things don't happen instantaneously after the measurement is made) q being the integer number of sample time it takes for the control calculation and includes any delay (buffer code) that is necessary (consider that some control calcs may take longer than others - depending upon the error and the rate of change in error)

Would be surprised if your system is not close to

Transfer function:

1.85e-005 z + 1.826e-005

------------------------ z^2 - 1.96 z + 0.9608

Sampling time: 0.1

Transfer function:

1.873e-007 z + 1.87e-007

------------------------ z^2 - 1.996 z + 0.996

Sampling time: 0.01

Transfer function:

1.875e-009 z + 1.875e-009

------------------------- z^2 - 2 z + 0.9996

Sampling time: 0.001

The above use the simple substitution for 's' and includes a zoh

There are other methods as i've said - an appreciation of these is important.

Kieran

Reply to
Setanta

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.