[Simulink] Discrete equivalent of variable transport delay

Hi,

i wanted to use a 'variable transport delay' in an triggered atomic subsystem but i cannot (because of an invalid sample time), so i tried to make an equivalent discrete thing with a 'variable fractionnal delay' available in the Signal Processing Blockset but I can not. Somedy have an idea please ? I don't know how i could make it.

Thanks a lot ! :)

Answer 1

[Guy Rouleau]

The Delay block should do the job:

http://www.mathworks.com/help/toolbox/simulink/slref/delay.html

You will need to enable the "delay length" option. Inside your triggered subsystem, you will be able to implement a logic to specify the delay length you need.

Answer 2

[Franck]

Hi Guy,

and thanks for your answer !

I apologize but i wasn't precise about the Matlab version i use : i use an old one which is the R2007b. I could use the R2009a (because this version is available in my laboratory) but in both version the "Delay" block you present to me is not present. Moreover, i think this block (when i read the documentation) comportment is like a variable time delay, not a variable transport delay. Am I wrong ?

Thanks again.

[Guy Rouleau]

I agree, the option to specify the delay length through an input port is relatively new. I do not have a good solution for older releases.

The Delay block is discrete. It delays your input by a certain number of samples. It does not do any interpolation like the continuous time delay or transport delay.

Answer 3

[Franck]

I agree, the 'delay' block is discrete and his comportment is similar to the variable time delay when the sample time is very low.

But, sorry to repeat me again, i don't think the delay block can be configure to be similar to the variable transport delay. And that is my problem.

Take a look in the documentation ( http://www.mathworks.fr/help/toolbox/simulink/slref/variabletimedelay.html ). I think the best way would be to understand in the variable transport delay, how the time 'td' is evaluated from 'ti'. If i could understand it, i could reproduce it in discrete.