Beruflich Dokumente
Kultur Dokumente
Aliaksei Maistrou
FEM definition
The finite element method (FEM) (also finite element analysis) is a numerical technique for finding approximate solutions of partial differential equations (PDE) as well as of integral equations. The solution approach is based either on eliminating the differential equation completely (steady state problems), or rendering the PDE into an approximating system of ordinary differential equations, which are then solved using standard techniques such as Euler's method, Runge-Kutta, etc.
where f is given and u is an unknown function of x, and u'' is the second derivative of u with respect to x 1) Convert P1 into its variational equivalent (or weak form). If u solves P1, then for any smooth function v that satisfies the displacement boundary conditions: v = 0 at x = 0 and x = 1, we have
where we have used the assumption that v(0) = v(1) = 0. What we get? What can we do with it? We can choose v as any smooth function, why? Mathematicians say we can! Then we can eliminate infinite dimensional linear problem + minimize dimensionality of finite dimensional space
Here we selected v as a basis with local support The primary advantage of basis with local support is that the inner products
Assumption (that results in approximation): we can represent our functions in the form:
M = ( M ij )
M ij = vi v j
will be zero for almost all j,k. And system matrix becomes sparse
Jakob slides
where f is given and u is an unknown function of x, and u'' is the second derivative of u with respect to x 1) Convert P1 into its variational equivalent (or weak form). If u solves P1, then for any smooth function v that satisfies the displacement boundary conditions: v = 0 at x = 0 and x = 1, we have
Assumption (that results in approximation): we can represent our functions in the form:
2) 3)
Now we have concrete algorithm for a large but finite dimensional linear problem whose solution will approximately solve the original BVP
2D basis functions
1D Acoustical problems
1D Acoustical problems
F0 u z
Problem: find static solution (dynamical also solvable but little more complicated)
u 1 N ( z, t ) = z EA( z )
Solving using FEM
F = ku
u 1 = N ( z) z EA
z1
z2
k2 =
1 1 2 i = k1u1 + k2 (u2 u1 ) 2 2 2
e = F1u1 F2u2
= i + e
EA 2 1 u1 F1 + =0 l / 2 1 1 u2 F2
F1
F2
u 1 u 1 = N ( z) N ( z) = 0 z EA z EA
u z1
n i =1
z2
n i =1
u ( z ) = uii ( z ) F ( z ) =
Wa + Wi = 0
Wa = F1 u1 + F2 u2
z l1
1 l1
EA((1
0
l2
u1 u1 u2 W = [ EA( + ) + F1 ] u1 + l1 l2 l2 u1 u2 +[ EA( + ) + F2 ] u2 = 0 l2 l2
EA 2 1 u1 F1 + =0 l / 2 1 1 u2 F2
Results?
2) 3)
Now we have concrete algorithm for a large but finite dimensional linear problem whose solution will approximately solve the original BVP