Travis

   
Simulink Problem
travisturkey
Posts: 498
travisturkey Posted Tue 22 Jul, 2008 2:01 PM Quote
Do any of you guys know how to use simulink(matlab)? I have an assignment due for tomorrow,I'm trying to figure it out right now,I'm quite new to it.is there any way where i can simulate,say, 1/a*(e^(-1/a*t))with varying a and t. with one variable it's easy but couldn't find the way for the other. any recommendation will be highly appreciated :)

Mehmet
 
Re: Simulink Problem
michele
Posts: 183
michele Posted Tue 22 Jul, 2008 2:48 PM Quote
well i used simulink just once for simulating an equalizer, but let's try. i suppose that t is the time flowing and a the costant of your system, and maybe you have to simulate your system with varying time and a. so i dunno if this is affordable: you can create many systems how many are the value of a that you treat as a constant in any system (of course every system has a different value of a) so every system will be fed with the same time vector (t) and then you will run the simulation with all the system in parallel. so now it this solution it's feasible it depends on how many values of a you have to consider and how you have to collect the results
 
Re: Simulink Problem
travisturkey
Posts: 498
travisturkey Posted Tue 22 Jul, 2008 2:54 PM Quote
michele wrote:
well i used simulink just once for simulating an equalizer, but let's try. i suppose that t is the time flowing and a the costant of your system, and maybe you have to simulate your system with varying time and a. so i dunno if this is affordable: you can create many systems how many are the value of a that you treat as a constant in any system (of course every system has a different value of a) so every system will be fed with the same time vector (t) and then you will run the simulation with all the system in parallel. so now it this solution it's feasible it depends on how many values of a you have to consider and how you have to collect the results


well I dunno if you have heard anyhing about control/feedback systems but now I have a first order transfer function which is formulated as (1/(a*s+1)),I'm giving in an impulse input (which is the constant command in matlab) and then connect them to the scope to see what happens. the corresponding inverse laplace of 1/a*s+1 is 1/a*(e^-1/a*t) so,first I have to find a way to sketch the inverse laplace of the output(which I couldn't yet) and then simulate it for different a's in one plot...i dunno if i can assign a as,say,a variable varying from 0.1 to 1 with an increment value of 0.1.
 
Re: Simulink Problem
michele
Posts: 183
michele Posted Tue 22 Jul, 2008 3:26 PM Quote
travisturkey wrote:


well I dunno if you have heard anyhing about control/feedback systems but now I have a first order transfer function which is formulated as (1/(a*s+1)),I'm giving in an impulse input (which is the constant command in matlab) and then connect them to the scope to see what happens. the corresponding inverse laplace of 1/a*s+1 is 1/a*(e^-1/a*t) so,first I have to find a way to sketch the inverse laplace of the output(which I couldn't yet) and then simulate it for different a's in one plot...i dunno if i can assign a as,say,a variable varying from 0.1 to 1 with an increment value of 0.1.


i know what are controlfeedback systems but it was long long time ago..., ok, i think that you can treat a as a constant and doing 10 different simulations with different value of a, the result will be a vector of something ist'n it? so once you have the different vectors you can plot them all together this way: plot(x1,t,x2,t, ...)
for having a block that reproduce this:1/a*(e^-1/a*t) sorry but i can't help you, try with the simulink help
 
Re: Simulink Problem
travisturkey
Posts: 498
travisturkey Posted Tue 22 Jul, 2008 3:39 PM Quote
michele wrote:
travisturkey wrote:


well I dunno if you have heard anyhing about control/feedback systems but now I have a first order transfer function which is formulated as (1/(a*s+1)),I'm giving in an impulse input (which is the constant command in matlab) and then connect them to the scope to see what happens. the corresponding inverse laplace of 1/a*s+1 is 1/a*(e^-1/a*t) so,first I have to find a way to sketch the inverse laplace of the output(which I couldn't yet) and then simulate it for different a's in one plot...i dunno if i can assign a as,say,a variable varying from 0.1 to 1 with an increment value of 0.1.


i know what are controlfeedback systems but it was long long time ago..., ok, i think that you can treat a as a constant and doing 10 different simulations with different value of a, the result will be a vector of something ist'n it? so once you have the different vectors you can plot them all together this way: plot(x1,t,x2,t, ...)
for having a block that reproduce this:1/a*(e^-1/a*t) sorry but i can't help you, try with the simulink help


i think I'm pretty much done with simulating different valued a systems simultaneously,I'm now investigating whether i can sort the rest out. Thank you for your help,it worked fine :)
 
Re: Simulink Problem
Moray
Posts: 1918
Moray Posted Tue 22 Jul, 2008 3:47 PM Quote
awww man I huvnae used matlab for years!
 
Re: Simulink Problem
michele
Posts: 183
michele Posted Tue 22 Jul, 2008 3:58 PM Quote
travisturkey wrote:


i think I'm pretty much done with simulating different valued a systems simultaneously,I'm now investigating whether i can sort the rest out. Thank you for your help,it worked fine :)


glad to hear it, your wellcome
 
Pages [1] All Times BST Current Time 9:42 PM
Post Reply