slow ilapalce calculation for systems with higher polynomial orders
2 views (last 30 days)
Show older comments
Is it possible to calculate the output o(t) of a system h(s) for the input x(s) faster than following codes? Why the following codes run extremely slow? What is wrong?
syms s t
time=linspace(0,4/1e09,401);
h=((5034923382327397*s^3)/627710173538668076383578942320766641610235544446403451289600 + (6766076405664713*s^2)/71362384635297994052914298472474756819137331200 + (210759975314673*s)/64903710731685345356631204115251200 + 376686340223973/281474976710656000000000000)/((5034923382327397*s^3)/12554203470773361527671578846415332832204710888928069025792 + (2284590225627523*s^2)/356811923176489970264571492362373784095686656 + (7512718920571563*s)/649037107316853453566312041152512 + 765850114395651/37778931862957161709568);
x=(100e-03/s)*(1-exp(-s/(2*1e09)))/(1+exp(-s/(2*1e09)));
y=ilaplace(x*h,s,t);
o=subs(y,t,time);
plot(time,o)
2 Comments
Walter Roberson
on 6 Feb 2019
your expression for h is either incorrect or meaningless. Numbers with that many digits will be approximated with floating point which will introduce false behaviour with roots and poles in different places than you expect .
Chances are high that you should be using symbolic constants not floating point .
sym('26271892563873553613')
Accepted Answer
Walter Roberson
on 6 Feb 2019
subs(vpa(y,16),t,time)
No point in calculating with large numbers to indefinite precision only to throw away that precision since plot() only has single precision resolution according to some tests I have done.
0 Comments
More Answers (0)
See Also
Categories
Find more on Calculus in Help Center and File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!