0

The PDE in question is, 

sol = DSolve[{epsilon^2*D[u[eta, zeta], eta, eta] + 
      D[u[eta, zeta], zeta, zeta] == -1}, u, {eta, zeta}]

enter image description here

DSolve is unable to find C[1] and C[2] with the following boundary conditions,

bc1 = 2*Kn*epsilon*(D[u[eta, zeta], eta] /. eta -> 1) == -u[eta, zeta];
bc2 = 2*Kn*epsilon*(D[u[eta, zeta], zeta] /. zeta -> 1) == -u[eta, zeta];
bc3 = D[u[eta, zeta], eta] == 0 /. eta -> 0
bc4 = D[u[eta, zeta], zeta] == 0 /. zeta -> 0

Is there an alternate way to find C[1] and C[2] subject to the above conditions?

zhk
  • 11,939
  • 1
  • 22
  • 38
  • …I think the /. eta -> 1 should be at the end of the equation i.e. 2*Kn*epsilon*(D[u[eta, zeta], eta] ) == -u[eta, zeta]/. eta -> 1? – xzczd Aug 17 '17 at 11:44
  • @xzczd Any suggestion on how to get a particular solution for this PDE? – zhk Aug 22 '17 at 14:53
  • Do this system have a finite b.c. at $\infty$ or $-\infty$? If so, the method mentioned here will probably work. – xzczd Aug 23 '17 at 02:40
  • @xzczd The issue is that geometry of the physical model is finite. So we can't have b.c at $\pm\infty$ – zhk Aug 23 '17 at 03:10
  • Then we need 2 more b.c.s, or we can't determine a particular solution. – xzczd Aug 23 '17 at 03:50
  • @xzczd I have included the missing b.cs. This is the same question for which you guyz (user21 and you) helped me with the numerical solution. https://mathematica.stackexchange.com/questions/147089/how-to-input-robin-boundary-conditions-for-nonstandard-laplace-equation – zhk Aug 23 '17 at 04:43

0 Answers0