
Dom V. answered 07/07/17
Tutor
5.0
(119)
Cornell Engineering grad specializing in advanced math subjects
This is a homogeneous, linear, 2nd-order equation with constant coefficients. That means we can attempt a trial solution of the form y=eAx, where A is a constant.
Substituting expressions for y, y', and y'' into the differential equation results in the characteristic equation:
A2-2A+(λ+1)=0.
Use the quadratic formula to solve for the possible values of A:
A={2±√[4-4(λ+1)]}/2
A=1±√[-λ]
A=1±i√λ (this is a complex number, which is perfectly okay)
The two possible values of A are complex conjugates:
A1=1+i√λ
A2=1-i√λ
And our general solution is
y(x)=Be(1+i√λ)x + Ce(1-i√λ)x, where B and C are our constants of integration.
Notice that we can simplify each exponential term:
e(1+i√λ)x = ex+ix√λ = (ex)(eix√λ),
Next use Euler's identity [eikx=cos(kx)+i sin(kx) for any constant k]:
e(1+i√λ)x --> ex[cos(√[λ]x)+i sin(√[λ]x)]
e(1-i√λ)x --> ex[cos(√[λ]x)-i sin(√[λ]x)]
Then,
y(x)-->Bex[cos(√[λ]x)+i sin(√[λ]x)] + Cex[cos(√[λ]x)-i sin(√[λ]x)]
Regroup terms:
y(x)=(B+C)excos(√[λ]x) + i(B-C)exsin(√[λ]x)
Rename constant coefficients (notice we can lump the i into the second constant--that is always allowed):
y(x)=a*excos(√[λ]x)+b*exsin(√[λ]x).
All that remains is to apply the two boundary conditions:
- y(0)=0
0=a*e0cos(0)+b*e0sin(0)=a
With a=0 the solution y(x) is reduced to
y(x)=b*exsin(√[λ]x)
- y'(L)=0
0=b*{√[λ]eLcos(√[λ]L)+eLsin(√[λ]L)}
0=beL{√[λ]cos(√[λ]L)+sin(√[λ]L)}
We want to avoid setting b=0 because it reduces our solution to y(x)=0. Notice that eL is just a constant, so we can divide it out of the equation. That means that the trig term inside {...} must equal 0:
√[λ]cos(√[λ]L)+sin(√[λ]L)=0
√[λ]cos(√[λ]L)=-sin(√[λ]L)
√[λ]=-tan(√[λ]L)
At this point we must plot the left side and right side of the above bolded equality as functions of λ and graphically find all intersections (this will depend on L and cannot be solved generally--but you can try plotting for L=1,2,3,etc to observe patterns). Here's an example when L=2: https://www.wolframalpha.com/input/?i=plot%5Bsqrt(x),+-tan(2sqrt(x))%5D
So we can say that the solution to the BVP is y(x)=b exsin(√[λ]x), where b is still arbitrary (and possibly complex) and where λ satisfies the equation √[λ]=-tan(√[λ]L) for a given value of L.