Control Engineering Question

Hi, when reading a text book , it got the overall transfer function after using compensator like PI, as G(s) ...xxx/xxx ok ? how then i can use this design to be implemented in the microcontroller ?

An S-plane compensation transfer function implemented in a computer requires that the computer calculate the values of the transfer function many times a second for a typical control loop. This is at least twice the frequency of the highest frequency in the loop, which often means a large number of complex-number calculations per second. If it is too many for standard microcontrollers, than you may need to use a DSP (Digital Signal Processor) controller which is optiminzed for doing such calculations.

Also, if the control loop signals are analog, then they will have to be converted by an ADC to digital for the microcontroller, and then back to analog by a DAC after the calculations. Some DSP chips have the ADC and DAC converters built in for this purpose.

I was told in my controls class that almost no actual control applications use transfer function and instead use state-space models because you can do things like multiple inputs, multiple outputs, non-zero initial conditions, and it's in the time-domain which is more applicable to real problems, so no need for conversions to the frequency domain and back which can be very difficult sometimes. It also does not necessarily require calculations in the complex plane like transfer functions.

(It is possible to change a transfer function into state space form). Maybe it would be more straightforawrd to implement if you did that.

thanks for the reply
but still didnt get it .. what after transforming to state space equations ??
how to impelment it in a microcontroller or processor ??

e.g in digital x(k+1) = G X(K)+ HU(K)
y = C X(K)+ DU(K)

and then ??

Ahmedragia21, do you have MATLAB or SCILAB? Both programs have some build-in functions to help transition from the state space to the transfer function. From there, you can apply your inverse Laplace and build up a discrete time domain controller for your μcontroller.