Difference in simulation results of continuous and discrete transfer function in simulink

Question

Karthik R 2015-7-8

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/229037-difference-in-simulation-results-of-continuous-and-discrete-transfer-function-in-simulink

评论： Reyad El-Khazali 2018-1-16

Hi,

I am trying to compare the continuous (s-domain) and discrete (z-domain) implementation of a system in simulink. My model is simple as shown.

My input signal is a sine of amplitude = '1' and frequency = '2*pi*1e3' rad/sec. The discrete domain TF is obtained by using 'c2d' function of Matlab with a sampling time of '1e-5' seconds. But I am getting entirely different responses to the same. It would be of great help if someone can clarify why this is happening.

   Many thanks in advance
      Karthik R

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Reyad El-Khazali 2018-1-16

Your discrete-time transfer function is wrong. The correct one is: G(z) = [ 0.000986 z^2 + 0.001972 z + 0.000986 ]/[z^2 -1.996z + 0.9999 ] obtained using c2d (N,D, 10^-5, 'Tustin') The simulation of both the continuous and the discrete one were identical.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Azzi Abdelmalek 2015-7-8

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/229037-difference-in-simulation-results-of-continuous-and-discrete-transfer-function-in-simulink#answer_185400

编辑：Azzi Abdelmalek 2015-7-8

Check the poles of your discrete model, they are unstable while your continuous model is stable. That means the discretization was unsuccessful. I think it's due to numeric calculation error. The denominator of your continuous model, maybe is the cause. D=[1 a b] with b very big then 1 and a. Try to make a discretization manually.