Prev: axes tick
Next: problem optimization with fmincon
From: moussams SALEY on 25 Jul 2010 18:56
Hi
I am a new new user of Matlab and have a system of non linear ODE but I am having a problem on the index. I have four variables T1,T2,T3,T4(T is the temperature) depending on time and varying with j. I would like to solve the system using ode45 and need your help on how I have to use the index with a loop to solve it.
The coefficients K1 to K17 are known as well for the initial conditions.
the system is:
dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5
dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11
dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1)
dT4dt(j)= K15*T3(j)+K16*T4(j)+K17
j=1.....N

For N=4, I have used this loop but it does not work:
for j=1:4
if (j==1)
dT1dt(1)= K1*T1(j)+K2*(T1(1)).^4+K3*T2(1)+K4*(T2(1)).^4+K5;
dT2dt(1)= K6*T1(1)+K7*T3(1)+K8*(T1(1)).^4+K9*T2(j)+K10*(T2(1)).^4+K11;
dT3dt(1)= K12*T2(1)+K13*T3(1)+K14*298; %T3(0)=298
dT4dt(1)= K15*T3(1)+K16*T4(1)+K17;
else
dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5;
dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11;
dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1);
dT4dt(j)= K15*T3(j)+K16*T4(j)+K17;
end
end
and i have gotten a message:

??? Attempted to access T1(2); index out of bounds because numel(T1)=1.

Error in ==> HYBRID at 138
dT1dt(j)=K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5;

Thanks in advance for your help
Best regards
From: us on 25 Jul 2010 20:05
"moussams SALEY" <moussa_mounk(a)yahoo.fr> wrote in message <i2ife4$j9i$1(a)fred.mathworks.com>...
> Hi
> I am a new new user of Matlab and have a system of non linear ODE but I am having a problem on the index. I have four variables T1,T2,T3,T4(T is the temperature) depending on time and varying with j. I would like to solve the system using ode45 and need your help on how I have to use the index with a loop to solve it.
> The coefficients K1 to K17 are known as well for the initial conditions.
> the system is:
> dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5
> dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11
> dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1)
> dT4dt(j)= K15*T3(j)+K16*T4(j)+K17
> j=1.....N
>
> For N=4, I have used this loop but it does not work:
> for j=1:4
> if (j==1)
> dT1dt(1)= K1*T1(j)+K2*(T1(1)).^4+K3*T2(1)+K4*(T2(1)).^4+K5;
> dT2dt(1)= K6*T1(1)+K7*T3(1)+K8*(T1(1)).^4+K9*T2(j)+K10*(T2(1)).^4+K11;
> dT3dt(1)= K12*T2(1)+K13*T3(1)+K14*298; %T3(0)=298
> dT4dt(1)= K15*T3(1)+K16*T4(1)+K17;
> else
> dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5;
> dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11;
> dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1);
> dT4dt(j)= K15*T3(j)+K16*T4(j)+K17;
> end
> end
> and i have gotten a message:
>
> ??? Attempted to access T1(2); index out of bounds because numel(T1)=1.
>
> Error in ==> HYBRID at 138
> dT1dt(j)=K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5;
>
> Thanks in advance for your help
> Best regards

well... the error message is rather self-explanatory:
your var T1 is a scalar, ie, one single value - but you try to access values J = 1:4

us
From: Steven_Lord on 25 Jul 2010 22:52


"moussams SALEY" <moussa_mounk(a)yahoo.fr> wrote in message
news:i2ife4$j9i$1(a)fred.mathworks.com...
> Hi
> I am a new new user of Matlab and have a system of non linear ODE but I am
> having a problem on the index. I have four variables T1,T2,T3,T4(T is the
> temperature) depending on time and varying with j. I would like to solve
> the system using ode45 and need your help on how I have to use the index
> with a loop to solve it.
> The coefficients K1 to K17 are known as well for the initial conditions.
> the system is: dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5
> dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11
> dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1)
> dT4dt(j)= K15*T3(j)+K16*T4(j)+K17
> j=1.....N

Why do you have an index at the end of your dT*dt and T* variables? As the
help for the ODE solvers indicate, your ODE function (the function you pass
into the solver as the first input argument) should accept inputs t and y
and return a variable dydt, the values of dy/dt at the time the function
received as input. Therefore, I think all your (j) indices should be
removed.

However, there is one little wrinkle I just noticed. In the expression for
dT3dt, you attempt to use the value of T3 at an earlier time. This means
you do NOT have a system of ordinary differential equations; you have a
system of _delay-differential equations_ instead. In that case, ODE45 is
NOT the right tool for this job; you will need to use DDE23 or DDESD, which
are very similar. Review the help and/or the reference page for those
solvers for more information about how you will need to set up your code.

--
Steve Lord
slord(a)mathworks.com
comp.soft-sys.matlab (CSSM) FAQ: http://matlabwiki.mathworks.com/MATLAB_FAQ
To contact Technical Support use the Contact Us link on
http://www.mathworks.com

From: moussams SALEY on 28 Jul 2010 05:36
"Steven_Lord" <slord(a)mathworks.com> wrote in message <i2ita5$h1e$1(a)fred.mathworks.com>...
>
>
> "moussams SALEY" <moussa_mounk(a)yahoo.fr> wrote in message
> news:i2ife4$j9i$1(a)fred.mathworks.com...
> > Hi
> > I am a new new user of Matlab and have a system of non linear ODE but I am
> > having a problem on the index. I have four variables T1,T2,T3,T4(T is the
> > temperature) depending on time and varying with j. I would like to solve
> > the system using ode45 and need your help on how I have to use the index
> > with a loop to solve it.
> > The coefficients K1 to K17 are known as well for the initial conditions.
> > the system is: dT1dt(j)= K1*T1(j)+K2*(T1(j)).^4+K3*T2(j)+K4*(T2(j)).^4+K5
> > dT2dt(j)= K6*T1(j)+K7*T3(j)+K8*(T1(j)).^4+K9*T2(j)+K10*(T2(j)).^4+K11
> > dT3dt(j)= K12*T2(j)+K13*T3(j)+K14*T3(j-1)
> > dT4dt(j)= K15*T3(j)+K16*T4(j)+K17
> > j=1.....N
>
> Why do you have an index at the end of your dT*dt and T* variables? As the
> help for the ODE solvers indicate, your ODE function (the function you pass
> into the solver as the first input argument) should accept inputs t and y
> and return a variable dydt, the values of dy/dt at the time the function
> received as input. Therefore, I think all your (j) indices should be
> removed.
>
> However, there is one little wrinkle I just noticed. In the expression for
> dT3dt, you attempt to use the value of T3 at an earlier time. This means
> you do NOT have a system of ordinary differential equations; you have a
> system of _delay-differential equations_ instead. In that case, ODE45 is
> NOT the right tool for this job; you will need to use DDE23 or DDESD, which
> are very similar. Review the help and/or the reference page for those
> solvers for more information about how you will need to set up your code.
>
> --
> Steve Lord
> slord(a)mathworks.com
> comp.soft-sys.matlab (CSSM) FAQ: http://matlabwiki.mathworks.com/MATLAB_FAQ
> To contact Technical Support use the Contact Us link on
> http://www.mathworks.com

Dear,
Thank you very much for your suggestions.
Best regards
 | 
Pages: 1
Prev: axes tick
Next: problem optimization with fmincon