10.04.2015 Views

EQUATIONS OF ELASTIC HYPERSURFACES

EQUATIONS OF ELASTIC HYPERSURFACES

EQUATIONS OF ELASTIC HYPERSURFACES

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

3. CALCULUS <strong>OF</strong> DIFFERENTIAL OPERATORS ON <strong>HYPERSURFACES</strong> 85<br />

In other words, (3.80) is the Euler-Lagrange equation associated with the integral functional<br />

(3.78).<br />

Similarly, minimizers of the energy functional<br />

∫<br />

E [U] :=<br />

S<br />

[‖dU‖ 2 + ‖δU‖ 2 ]<br />

dS, U ∈ Λ l TS , (3.81)<br />

are null-solutions of the Hodge-Laplacian (cf. later (3.124)), while minimizers of the energy<br />

functional<br />

∫<br />

E [U] := ‖∇U‖ 2 dS, U ∈ TS , (3.82)<br />

S<br />

are null-solutions of the Bochner-Laplacian ((cf. later (3.125)).<br />

Our aim is to adopt a similar point of view in the case of anisotropic and isotropic (Lamé)<br />

system of elasticity on S . The departure point is to consider the total free (elastic) energy<br />

∫<br />

E [U] := E(x, D S U(x)) dS, D S U := [ Dj S U k , U ∈ TS (3.83)<br />

]n×n<br />

S<br />

(cf. (3.68)), ignoring at the moment the displacement boundary conditions (Koiter’s model)).<br />

As before, equilibria states correspond to minimizers of the above variational integral (see<br />

[NH1, § 5.2]). First we should identify the correct form of the stored energy density<br />

E(x, D S U(x)). We shall restrict attention to the case of linear elasticity. In this scenario,<br />

E = (S S , Def S ) depends bi-linearly on the stress tensor S S = [ S jk] and the deformation<br />

(strain)<br />

n×n<br />

tensor<br />

Def S = [ D jk] n×n , Djk U := 1 2[<br />

D<br />

S<br />

j U k + D S k U j<br />

]<br />

, j, k = 1, . . . , n (3.84)<br />

which, according to Hooke’s law, satisfy S S = TDef S , for some linear, fourth-order tensor<br />

T. If the medium is also homogeneous (i.e. the density and elastic parameters are positionindependent),<br />

it follows that E depends quadratically on D S U, i.e.<br />

for some linear operator<br />

E(x, D S U(x)) = 〈TD S U(x), D S U(x)〉 (3.85)<br />

T : M n,n (R) −→ M n,n (R), (3.86)<br />

where M n,n (R) stands for the vector space of all n × n matrices with real entries. Hereafter,<br />

we organize M n,n (R) as a real Hilbert space with respect to the inner product<br />

〈A, B〉 := Tr(AB ⊤ ) = ∑ i,j<br />

a ij b ij , ∀ A = [a ij ] i,j<br />

, B = [b ij ] i,j<br />

∈ M n,n (R), (3.87)<br />

where B ⊤ denotes transposed matrix, and Tr is the usual trace operator for square matrices.<br />

A linear operator (3.86 is a tensor of order 4, i.e., T = [ c ijkl<br />

]ijkl , and<br />

[ ]<br />

∑<br />

TA = c ijkl a kl , for A = [a kl ] kl<br />

∈ M n,n (R) . (3.88)<br />

k,l<br />

ij

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!