Skip to main content

differential equations - Solution of nonlinear system with boundary conditions


I'm try solve the following coupled ODEs with boundary conditions:


$I. \ \ \ \ \ \ \ \dfrac{4}{r}[1+A(r)]\left(\dfrac{dH}{dr}\right)^2+\dfrac{dG}{dr}=0$


$II. \ \ \ \ \dfrac{1}{r}\left(\dfrac{dA}{dr}\right)+F(r)+k^2G(r)=0$



$III. \ \ \ \ \left(\dfrac{dG}{dr}\right)^2+4k^2\left(G(r)+\dfrac{F(r)}{k^2}\right)^2\left(\dfrac{dH}{dr}\right)^2-1.6H(r)\left(\dfrac{dH}{dr}\right)^2=0$


$IV. \ \ \ \ \dfrac{d^2F}{dr^2}+\dfrac{1}{r}\dfrac{dF}{dr}+\dfrac{1}{r}\dfrac{dA}{dr}-4k^2F(r)\left(\dfrac{dH}{dr}\right)^2=0$


where $k>0$. The boundary conditions are


$H(0)=1\\A(0)=0\\G(L)=0\\F'(0)=0, \ \ F(L)=0$


with $L$ being the boundary, such that can be adjusted in order to satisfy the boundary conditions above for each value of parameter $k$ choosed. For this physical problem, it is desirable that the ODEs of first order obey other conditions:


$H'(0)=0, \ \ H(L)=H'(L)=0\\A'(0)=0, \ \ A'(L)=0$


Thus, from ODEs, it is possible to conclude


The ODE $I$ guarantees $G'(0)=G'(L)=0$


The ODE $II$ guarantees $F(L)=0$


I tried solve, for example, with k=1 using the shooting method. In this case, it was only possible to execute up to $L = 1.45$ (end=1.45).



k = 1; eps = 10^-10; end = 1.45;
ode1 = 4 (h'[r] ^2) (a[r] + 1)/r == -G'[r];
ode2 = a'[r]/r == -(k^2) (G[r]) - f[r];
ode3 = (G'[r])^2 + 4 (k^2) (h'[r]^2) (G[r] + f[r]/k^2)^2 == (1.6) h[r] (h'[r]^2);
ode4 = f''[r] + f'[r]/r + (a'[r]/r) == 4 (k^2) (f[r]) ((h'[r])^2);
bcs = {h[eps] == (1 - eps), a[eps] == eps, G[end] == eps, f'[eps] == eps, f[end] == eps};

sol = NDSolve[{ode1, ode2, ode3, ode4, bcs}, {a, h, f, G}, {r, eps, end},
Method -> {"Shooting","StartingInitialConditions" -> {f[end] == eps, G[end] == eps}}];


Plot[Evaluate[{h[r], h'[r]} /. sol[[3]]], {r, eps, end},
PlotLegends -> {"H", "H'"}, PlotRange -> All]
Plot[Evaluate[{a[r], a'[r]} /. sol[[3]]], {r, eps, end},
PlotLegends -> {"A", "A'"}, PlotRange -> All]
Plot[Evaluate[{G[r], G'[r]} /. sol[[3]]], {r, eps, end},
PlotLegends -> {"G", "G'"}, PlotRange -> All]
Plot[Evaluate[{f[r], f'[r]} /. sol[[3]]], {r, eps, end},
PlotLegends -> {"F", "F'"}, PlotRange -> All]

enter image description here



However, for values $L>1.45$ appear the error message



"NDSolve::ndsz: At r == 1.294087660532624`, step size is effectively zero; singularity or stiff system suspected.."



But I believe it is possible to find the correct value of $L$ that satisfies all boundary conditions. The behavior of the functions by the plots suggests that this value can be near to 2 or 3. I can not resolve this error message for this problem.


Added_1: Specifically, I am interested in obtaining numerical solutions for the cases of $k=0.1$, $k=1$ and $k=2$. Then, there must be a value $L$ for each $k$ that satisfies all the boundary conditions of the problem.


Any help or altenative solution is welcome.



Answer



There is a solution satisfying all the boundary conditions for L = 2.665. I'll show you how to build this solution. First, we solve explicitly the first and third equations for the derivatives, we have


ode1 = 4 (h'[r]^2) (a[r] + 1)/r == -G'[r];

ode3 = (G'[r])^2 + 
    4 (k^2) (h'[r]^2) (G[r] + f[r]/k^2)^2 == (8/5) h[r] (h'[r]^2);
s = Solve[{ode1, ode3}, {h'[r], G'[r]}] // FullSimplify

Here we have three pairs of roots:


{{Derivative[1][h][r] -> 0, 
  Derivative[1][G][r] -> 
   0}, {Derivative[1][h][r] -> -((
    I r Sqrt[5 (f[r] + k^2 G[r])^2 - 2 k^2 h[r]])/(
    2 Sqrt[5] Sqrt[k^2 (1 + a[r])^2])), 

  Derivative[1][G][r] -> (5 r (f[r] + k^2 G[r])^2 - 2 k^2 r h[r])/(
   5 k^2 (1 + a[r]))}, {Derivative[1][h][r] -> (
   I r Sqrt[5 (f[r] + k^2 G[r])^2 - 2 k^2 h[r]])/(
   2 Sqrt[5] Sqrt[k^2 (1 + a[r])^2]), 
  Derivative[1][G][r] -> (5 r (f[r] + k^2 G[r])^2 - 2 k^2 r h[r])/(
   5 k^2 (1 + a[r]))}}

The first pair of roots corresponds to the trivial solution $H(r)=1,G(r)=0$. On this branch the second and fourth equations become linear, and their solution can be obtained in an explicit form. We are interested in the second pair of roots, from which we will form a new system of equations:


ode13 = {Derivative[1][G][r] == (
    5 r (f[r] + k^2 G[r])^2 - 2 k^2 r h[r])/(5 k^2 (1 + a[r])), 

   Derivative[1][h][r] == -(( 
     r Sqrt[-5 (f[r] - k^2 G[r])^2 + 2 k^2 h[r]])/(
     2 Sqrt[5] Sqrt[k^2 (1 + a[r])^2]))};
ode24 = {a'[r]/r == -(k^2) (G[r]) - f[r], 
  f''[r] + f'[r]/r + (a'[r]/r) == 4 (k^2) (f[r]) ((h'[r])^2)};

We reduce this system to the dimensionless form by making a substitution r->L*r. Then the parameter $L$ becomes a parameter of the equations, and the system and boundary conditions have the form:


k = 1; eps = 10^-10; L = 1.638;
ode13 = {Derivative[1][G][r] == 
    L^2*(5 r (f[r] + k^2 G[r])^2 - 2 k^2 r h[r])/(5 k^2 (1 + a[r])), 

   Derivative[1][h][r] == -L^2*( 
     r Sqrt[-5 (f[r] + k^2 G[r])^2 + 2 k^2 h[r]])/(
     2 Sqrt[5] Sqrt[k^2 (1 + a[r])^2])};
ode24 = {a'[r]/r == -L^2*((k^2) (G[r]) - f[r]), 
    f''[r] + f'[r]/r + (a'[r]/r) == 4 (k^2) (f[r]) ((h'[r])^2)};
bcs = {h[eps] == 1, a[eps] == 0, G[1] == 0, f'[eps] == 0, f[1] == 0};
sol = NDSolve[{ode13, ode24, bcs}, {a, h, f, G}, {r, eps, 1}, 
   Method -> {"Shooting", 
     "StartingInitialConditions" -> {f[eps] == -0.1, G[eps] == 0.33}}];
{Plot[Evaluate[Re[{h[r], h'[r]}] /. sol], {r, eps, 1}, 

  PlotLegends -> {"H", "H'"}, PlotRange -> All],
 Plot[Evaluate[Re[{a[r], a'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"A", "A'"}, PlotRange -> All],
 Plot[Evaluate[Re[{G[r], G'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"G", "G'"}, PlotRange -> All],
 Plot[Evaluate[Re[{f[r], f'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"F", "F'"}, PlotRange -> All]}

fig1


Another identity transformation of the system allows us to find a solution with L = 2:



k = 1; eps = 10^-10; L = 2;
ode = {G'[r] == (L^2 r (5 (f[r] + k^2 G[r])^2 - 2 k^2 h[r]))/(
    5 k^2 (1 + a[r])), 
   Derivative[1][h][r] == -((
     L^2 r Sqrt[-5 f[r]^2 - 10 k^2 f[r] G[r] - 5 k^4 G[r]^2 + 
       2 k^2 h[r]])/Sqrt[20 k^2 + 40 k^2 a[r] + 20 k^2 a[r]^2]), 
   r*f''[r] + f'[r] - (r*L^2*((k^2) (G[r]) + f[r])) == 
    r*4 (k^2) (f[r]) ((h'[r])^2), 
   a'[r] == -r*L^2*((k^2) (G[r]) + f[r])};
bcs = {h[eps] == 1, a[eps] == 0, G[1] == 0, f'[eps] == 0, f[1] == 0};


sol = NDSolve[{ode, bcs}, {a, h, f, G}, {r, eps, 1}, 
   Method -> {"Shooting", 
     "StartingInitialConditions" -> {f[eps] == -0.1, G[eps] == 0.33}}];


{Plot[Evaluate[Re[{h[r], h'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"H", "H'"}, PlotRange -> All],
 Plot[Evaluate[Re[{a[r], a'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"A", "A'"}, PlotRange -> All],

 Plot[Evaluate[Re[{G[r], G'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"G", "G'"}, PlotRange -> All],
 Plot[Evaluate[Re[{f[r], f'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"F", "F'"}, PlotRange -> All]}

fig2


Finally, we point out a solution with h[L] tending to zero. In this case L = 2.665


k = 1; eps = 10^-10; L = 2.665;
ode = {G'[r] == (L^2 r (5 (f[r] + k^2 G[r])^2 - 2 k^2 h[r]))/(
    5 k^2 (1 + a[r])), 

   Derivative[1][h][r] == -((
     L^2 r Sqrt[-5 f[r]^2 - 10 k^2 f[r] G[r] - 5 k^4 G[r]^2 + 
       2 k^2 h[r]])/Sqrt[20 k^2 + 40 k^2 a[r] + 20 k^2 a[r]^2]), 
   r*f''[r] + f'[r] - (r*L^2*((k^2) (G[r]) + f[r])) == 
    r*4 (k^2) (f[r]) ((h'[r])^2), 
   a'[r] == -r*L^2*((k^2) (G[r]) + f[r])};
bcs = {h[eps] == 1, a[eps] == 0, G[1] == 0, f'[eps] == 0, f[1] == 0};

sol = NDSolve[{ode, bcs}, {a, h, f, G}, {r, eps, 1}, 
   Method -> {"Shooting", 

     "StartingInitialConditions" -> {f[eps] == -0.340071, 
       G[eps] == 0.736822}}];


{Plot[Evaluate[Re[{h[r], h'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"H", "H'"}, PlotRange -> All],
 Plot[Evaluate[Re[{a[r], a'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"A", "A'"}, PlotRange -> All],
 Plot[Evaluate[Re[{G[r], G'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"G", "G'"}, PlotRange -> All],

 Plot[Evaluate[Re[{f[r], f'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"F", "F'"}, PlotRange -> All]}

fig3


We indicate the solution algorithm in the case k = 2. We set A = 1, we define two functions


k = 2; eps = 10^-10; L = 1;
ode = {G'[
     r] == (L^2 r (5 (f[r] + k^2 G[r])^2 - 2 k^2 h[r]))/(5 k^2 (1 + 
         a[r])), Derivative[1][h][
     r] == -((L^2 r Sqrt[-5 f[r]^2 - 10 k^2 f[r] G[r] - 

           5 k^4 G[r]^2 + 2 k^2 h[r]])/
       Sqrt[20 k^2 + 40 k^2 a[r] + 20 k^2 a[r]^2]), 
   r*f''[r] + f'[r] - (r*L^2*((k^2) (G[r]) + f[r])) == 
    r*4 (k^2) (f[r]) ((h'[r])^2), 
   a'[r] == -r*L^2*((k^2) (G[r]) + f[r])};
 bcsp = {h[eps] == 1, a[eps] == 0, f'[eps] == 0, f[eps] == p, 
   G[eps] == q};
g = ParametricNDSolveValue[{ode, bcsp}, G, {r, eps, 1}, {p, q}, 
  Method -> {"StiffnessSwitching", "NonstiffTest" -> False}, 
  MaxSteps -> Infinity]


 F = 
 ParametricNDSolveValue[{ode, bcsp}, f, {r, eps, 1}, {p, q}, 
  Method -> {"StiffnessSwitching", "NonstiffTest" -> False}, 
  MaxSteps -> Infinity]

To fulfill the boundary conditions for r = 1, we find the roots of the system of equations


FindRoot[{g[p, q][1] == 0, 
      F[p, q][1] == 0}, {p, -.1}, {q, .18}, Method -> "Secant"]


    Out[]= {p -> -0.118505, q -> 0.187442}

Using these roots we find the solution to the original system of equations


bcs = {h[eps] == 1, a[eps] == 0, f'[eps] == 0, 
   f[eps] == -0.1185047596657856`, G[eps] == 0.18744201629754556`};
bcs1 = {G[1] == 0, f[1] == 0};
sol = NDSolve[{ode, bcs}, {a, h, f, G}, {r, eps, 1}, 
   Method -> {"StiffnessSwitching", "NonstiffTest" -> False}];



{Plot[Evaluate[Re[{h[r], h'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"H", "H'"}, PlotRange -> All], 
 Plot[Evaluate[Re[{a[r], a'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"A", "A'"}, PlotRange -> All], 
 Plot[Evaluate[Re[{G[r], G'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"G", "G'"}, PlotRange -> All], 
 Plot[Evaluate[Re[{f[r], f'[r]}] /. sol], {r, eps, 1}, 
  PlotLegends -> {"F", "F'"}, PlotRange -> All]}

fig4 If we want to additionally find L for which h[1]=0, we should move gradually changing the parameters L,p,q.



Comments

Post a Comment

Popular posts from this blog

plotting - Filling between two spheres in SphericalPlot3D

Manipulate[ SphericalPlot3D[{1, 2 - n}, {θ, 0, Pi}, {ϕ, 0, 1.5 Pi}, Mesh -> None, PlotPoints -> 15, PlotRange -> {-2.2, 2.2}], {n, 0, 1}] I cant' seem to be able to make a filling between two spheres. I've already tried the obvious Filling -> {1 -> {2}} but Mathematica doesn't seem to like that option. Is there any easy way around this or ... Answer There is no built-in filling in SphericalPlot3D . One option is to use ParametricPlot3D to draw the surfaces between the two shells: Manipulate[ Show[SphericalPlot3D[{1, 2 - n}, {θ, 0, Pi}, {ϕ, 0, 1.5 Pi}, PlotPoints -> 15, PlotRange -> {-2.2, 2.2}], ParametricPlot3D[{ r {Sin[t] Cos[1.5 Pi], Sin[t] Sin[1.5 Pi], Cos[t]}, r {Sin[t] Cos[0 Pi], Sin[t] Sin[0 Pi], Cos[t]}}, {r, 1, 2 - n}, {t, 0, Pi}, PlotStyle -> Yellow, Mesh -> {2, 15}]], {n, 0, 1}]
Post a Comment
Read more

plotting - Plot 4D data with color as 4th dimension

I have a list of 4D data (x position, y position, amplitude, wavelength). I want to plot x, y, and amplitude on a 3D plot and have the color of the points correspond to the wavelength. I have seen many examples using functions to define color but my wavelength cannot be expressed by an analytic function. Is there a simple way to do this? Answer Here a another possible way to visualize 4D data: data = Flatten[Table[{x, y, x^2 + y^2, Sin[x - y]}, {x, -Pi, Pi,Pi/10}, {y,-Pi,Pi, Pi/10}], 1]; You can use the function Point along with VertexColors . Now the points are places using the first three elements and the color is determined by the fourth. In this case I used Hue, but you can use whatever you prefer. Graphics3D[ Point[data[[All, 1 ;; 3]], VertexColors -> Hue /@ data[[All, 4]]], Axes -> True, BoxRatios -> {1, 1, 1/GoldenRatio}]
Post a Comment
Read more

plotting - Mathematica: 3D plot based on combined 2D graphs

I have several sigmoidal fits to 3 different datasets, with mean fit predictions plus the 95% confidence limits (not symmetrical around the mean) and the actual data. I would now like to show these different 2D plots projected in 3D as in but then using proper perspective. In the link here they give some solutions to combine the plots using isometric perspective, but I would like to use proper 3 point perspective. Any thoughts? Also any way to show the mean points per time point for each series plus or minus the standard error on the mean would be cool too, either using points+vertical bars, or using spheres plus tubes. Below are some test data and the fit function I am using. Note that I am working on a logit(proportion) scale and that the final vertical scale is Log10(percentage). (* some test data *) data = Table[Null, {i, 4}]; data[[1]] = {{1, -5.8}, {2, -5.4}, {3, -0.8}, {4, -0.2}, {5, 4.6}, {1, -6.4}, {2, -5.6}, {3, -0.7}, {4, 0.04}, {5, 1.0}, {1, -6.8}, {2, -4.7}, {3, -1.
Post a Comment
Read more