Jjjones98

35 Reputation

4 Badges

2 years, 235 days

MaplePrimes Activity


These are questions asked by Jjjones98

I have written a bit of code which solves a linear system for some quantities which have been Laplace and then Fourier transformed.  

e1 := -2*D*i*k*pi + A*s = 0

e2 := 2*A*i*k*pi + 2*C*i*k*pi + B*s = a

e3 := s*C + 4/5*P*w3*(2*pi*i*k*D - 2*1/3*pi*i*k*D)/w2 = -2*(C + 2*K*(2*pi*i*k*B - 2*1/3*pi*i*k*B))/(w2*K)

e4 := s*D + 2*5/4*P*t4*pi*i*k*C/(t2*K) = -5/(2*P)*D/(t2*K)

sys := {s*C + 4/5*P*w3*(2*pi*i*k*D - 2*1/3*pi*i*k*D)/w2 = -2*(C + 2*K*(2*pi*i*k*B - 2*1/3*pi*i*k*B))/(w2*K), s*D + 2*5/4*P*t4*pi*i*k*C/(t2*K) = -5/(2*P)*D/(t2*K), -2*D*i*k*pi + A*s = 0, 2*A*i*k*pi + 2*C*i*k*pi + B*s = a}

solve(sys, [A, B, C, D])

Linear_System.mw

I get at the end some fractions where everything in the fractions is a constant with some physical meaning except for k which is the only frequency as I am working in one dimension so just need one-dimensional Fourier and Laplace transforms.  s is the corresponding variable from the Laplace transform. 

I was wondering if Maple had some functionality which would enable me to inverse Laplace and then inverse Fourier transform these quantities A, B, C and D from the linear system such that I obtain an algebraic expression at the end and not a numerical result.

I have solved the following linear system for 6 variables on Maple using the following code:

sys := {w = -2*Pi*i*k_2*v + 2*Pi*i*k_2*(4*K^2)/(5*Pi)*u, z = -2*Pi*i*k_1*v + 2*Pi*i*k_1*(4*K^2)/(5*Pi)*u, p*s*x = -2*Pi*i*k_1*u - 4*Pi^2*(k_1^2 + k_2^2)*x + a_1, p*s*y = -2*Pi*i*k_2*u - 4*Pi^2*(k_1^2 + k_2^2)*y + a_2, k_1*z + k_2*w = 0, k_1*x + k_2*y = 0}

solve(sys, [x, y, z, w, v, u])

However, the solution yields z = w = 0, I was wondering if this is correct as I feel that these quantities should not be 0 in the problem which I am studying, is there a way to find a solution which does not involve setting these two to 0?

.

I have a function e^(-\lambda z \sqrt(x^2 + y^2)), is it possible to use Maple to find some sequence of derivatives wrt to x and y which could be applied to this function to get

(z/(1+2*sqrt(x^2 + y^2)*lambda))*e^(-\lambda z \sqrt(x^2 + y^2))

 

 

I am trying to find 6 unknowns A, B, C, L, E, and F and have managed to write down a system of 6 equations involving these unknowns.  However, the equations are long and which I try to put the system together and solve with solve(sys, {A, B, C, E, F, L}) it says `[Length of output exceeds limit of 1000000]`.  Although the equations are long, I feel like something must be going wrong as the output cannot be that long.  The equations are:

C = (((h/(4*Pi*K)*H*e^(-k*h)*I/k - k*h^2/2 + H*A*e^(-k*h)*I/(K*k^2) - B*h - sqrt(Pi/2)*h/(2*Pi)*H*e^(-k*h)*I + sqrt(Pi/2)*K*k*h - sqrt(Pi/2)*H*A*e^(-k*h)*I/k + sqrt(Pi/2)*B*K - sqrt(Pi/2)*K*H*k*h^2*I/2) + sqrt(Pi/2)*H*A*e^(-k*h)*I/k) - sqrt(Pi/2)*H*F*h*K*I - sqrt(Pi/2)*H*A*e^(-k*h)*I/k + sqrt(Pi/2)*K*H*k*h^2/2*I) - sqrt(Pi/2)*H*F*h*K*I,

E = (((h/(4*Pi*K)*J*e^(-k*h)*I/k - k*h^2/2 + J*A*e^(-k*h)*I/(K*k^2) - L*h - sqrt(Pi/2)*h/(2*Pi)*J*e^(-k*h)*I + sqrt(Pi/2)*K*k*h - sqrt(Pi/2)*J*A*e^(-k*h)*I/k + sqrt(Pi/2)*L*K - sqrt(Pi/2)*K*J*k*h^2*I/2) + sqrt(Pi/2)*J*A*e^(-k*h)*I/k) - sqrt(Pi/2)*J*F*h*K*I - sqrt(Pi/2)*J*A*e^(-k*h)*I/k + sqrt(Pi/2)*K*J*k*h^2/2*I) - sqrt(Pi/2)*J*F*h*K*I

0 = -H*(k*z^2/2 - H*A*e^(-k*z)*I/(K*k^2) + B*z + C)*I - J*(k*z^2/2 - J*A*e^(-k*z)*I/(K*k^2) + L*z + E)*I + k*z + A*e^(-k*z)/K + F

0 = ((-H*A*e^(-k*z)*I - H*I*(-2*K*(-H*k*z^2*I/2 - H^2*A*e^(-k*z)/(K*k^2) - H*B*z*I - H*C*I)) - J*I*(-2*K*1/2*(((-J*k*z^2*I/2 - J*H*A*e^(-k*z)/(K*k^2) - J*B*z*I - J*C*I - H*k*z^2*I/2) - J*H*A*e^(-k*z)/(K*k^2)) - H*L*z*I - i*H*E)) - K*k) + H*A*e^(-k*z)*I) + K*H*z*k*I + H*A*e^(-k*z)*I + K*H*F*I

0 = ((-J*A*e^(-k*z)*I - H*(((K*J*k*z^2*I/2 + J*H*A*e^(-k*z)/k^2 + K*J*B*z*I + k*h*K*z^2*I/2) + J*H*A*e^(-k*z)/k^2) + K*H*L*z*I + K*H*E*I)*I - J*I*(-2*K*(-J*k*z^2*I/2 - J^2*A*e^(-k*z)/(K*k^2) - J*L*z*I - J*E*I)) - K*k) + J*A*e^(-k*z)*I) + K*J*z*k*I + J*A*e^(-k*z)*I + K*J*F*I

0 = (-k*A*e^(-k*z) - H*I*(-2*K*1/2*(((k*z + H*A*e^(-k*z)*I/(K*k) + B - H*k*z^2*I/2) + H*A*e^(-k*z)*I/(K*k)) - H*F*z*I - H*G*I)) - J*I*(-2*K*1/2*(((k*z + J*A*e^(-k*z)*I/(K*k) + L - J*k*z^2*I/2) + J*A*e^(-k*z)*I/(K*k)) - J*F*z*I - J*G*I)) - 2*K*k) + 2*k*A*e^(-k*z)

where h takes a constant value, H and J are constants, k is the square root of H^2 + J^2, I is the imaginary unit, and z also takes some value as a parameter.

I am reading a paper which has some useful two-dimensional Fourier transforms in the appendix: for example,

Fourier transform of 1/r = (1/k)*e^(-kz),

where r = sqrt(x^2 + y^2 + z^2) and k =  sqrt(k_1^2 + k_2^2).

My guess is that the author has computed these by taking contour integrals in the upper half-plane and I would like to compute some of these myself but I have many of them to compute and was wondering if it could be done with Maple instead.

For example, could I use Maple to verify that the above 2D Fourier transform is correct and that the inverse 2D Fourier transform takes you back to the original (or almost takes you back).  After that I would then like to feed in the functions which I have to get Fourer and inverse Fourier transforms.

1 2 3 4 5 Page 1 of 5