OctoInformBeta

OpenAI preprint, September 2026 · A variational principle for tangent vector fields

Section 3: divergence-curl energy and weak tangency: Sobolev test spaces determine normal divergence and curl, with a real spectral shift in the vector form

Weak tangency and divergence-curl energy.

Hover over a blue or green term in the text, or a node in the graph, to find it in the other. Select it to read its declaration both in Lean and in words.

The Setting

The variational setting concerns square-integrable vector fields on a planar set Ω⊂ℝ2\Omega\subset\mathbb R^2, with divergence and curl tested against scalar Sobolev functions. The choice of test space specifies the boundary condition in the weak formulation.

Let e1,e2e_1,e_2 be the standard orthonormal basis of the Euclidean plane. All L2L^2 spaces below use Lebesgue measure restricted to Ω\Omega. The space of scalar-vector pairs is L2(Ω)×L2(Ω;ℝ2)L^2(\Omega)\times L^2(\Omega;\mathbb R^2) with the Euclidean product norm in (1): ∥(v,g)∥2=∥v∥L2(Ω)2+∥g∥L2(Ω;ℝ2)2.\|(v,g)\|^2=\|v\|_{L^2(\Omega)}^2 +\|g\|_{L^2(\Omega;\mathbb R^2)}^2.(1)✦

A test function φ∈Cc∞(Ω)\varphi\in C_c^\infty(\Omega) is an infinitely differentiable real function on ℝ2\mathbb R^2 with compact support whose closed support lies in Ω\Omega.✦ For a test function φ∈Cc∞(Ω)\varphi\in C_c^\infty(\Omega) and a direction e∈ℝ2e\in\mathbb R^2, the test derivative of φ\varphi in direction ee is the L2L^2 class of ∂eφ(x)=Dφ(x)e\partial_e\varphi(x)=D\varphi(x)e, and the test vector field associated with φ\varphi and ee is the L2L^2 class of x↦φ(x)ex\mapsto\varphi(x)e.

An H1H^1 weak gradient gg of a real function vv on Ω\Omega is a square-integrable vector field such that vv is square-integrable and the pair satisfies (2): ∫Ωv∂eφdx=−∫Ω(g⋅e)φdx(φ∈Cc∞(Ω),e∈ℝ2).\int_\Omega v\,\partial_e\varphi\,dx =-\int_\Omega (g\cdot e)\varphi\,dx \qquad \bigl(\varphi\in C_c^\infty(\Omega),\ e\in\mathbb R^2\bigr).(2) The Sobolev space H1(Ω)H^1(\Omega) is the weak gradient graph, the real linear subspace of scalar-vector pairs satisfying (2). The Sobolev class ([v],[g])([v],[g]) of a function vv with weak gradient gg is the element of H1(Ω)H^1(\Omega) represented by the pair (v,g)(v,g).✦ The Sobolev value map u↦val⁡uu\mapsto\operatorname{val}u and Sobolev gradient u↦∇uu\mapsto\nabla u on H1(Ω)H^1(\Omega) are the continuous linear projections onto the scalar and vector components, respectively. The Sobolev norm is the restriction of (1) to the graph.

For open Ω\Omega, the Sobolev embedding of test functions sends φ∈Cc∞(Ω)\varphi\in C_c^\infty(\Omega) to its Sobolev class with its classical gradient.✦ For open Ω\Omega, the Dirichlet Sobolev space H01(Ω)H_0^1(\Omega) is the closure in the Sobolev norm of the Sobolev classes of test functions φ∈Cc∞(Ω)\varphi\in C_c^\infty(\Omega) with their classical gradients. The Dirichlet value map and Dirichlet gradient on H01(Ω)H_0^1(\Omega) are the restrictions of the Sobolev value map and Sobolev gradient to H01(Ω)H_0^1(\Omega). Under an integral, a Sobolev function denotes its scalar value.

The space of square-integrable vector fields is L2(Ω;ℝ2)L^2(\Omega;\mathbb R^2).✦ The coordinate embeddings f↦feif\mapsto f e_i and coordinate projections Z↦Z⋅eiZ\mapsto Z\cdot e_i, for i=1,2i=1,2, are continuous linear maps between the scalar and vector L2L^2 spaces. A Sobolev vector field XX on Ω\Omega is a pair X=(X1,X2)∈H1(Ω)2X=(X_1,X_2)\in H^1(\Omega)^2. For a Sobolev vector field X=(X1,X2)∈H1(Ω)2X=(X_1,X_2)\in H^1(\Omega)^2 and a scalar Sobolev function u∈H1(Ω)u\in H^1(\Omega), the L2L^2 vector value 𝑿\mathbf X, the weak partial derivatives of uu, and the Sobolev divergence and curl of XX are given by (3): 𝑿=(val⁡X1)e1+(val⁡X2)e2,∂iu=(∇u)⋅ei,i=1,2,div⁡X=∂1X1+∂2X2,curl⁡X=∂1X2−∂2X1.\begin{aligned} \mathbf X&=(\operatorname{val}X_1)e_1 +(\operatorname{val}X_2)e_2, &\partial_i u&=(\nabla u)\cdot e_i,\qquad i=1,2,\\ \operatorname{div}X&=\partial_1X_1+\partial_2X_2, &\operatorname{curl}X&=\partial_1X_2-\partial_2X_1. \end{aligned}(3) The quarter-turn rotation JJ in (4) also denotes the continuous linear map on vector L2L^2 fields obtained by applying the rotation pointwise: J(z1,z2)=(−z2,z1).J(z_1,z_2)=(-z_2,z_1).(4)

The Definition

Definition 1. For Z∈L2(Ω;ℝ2)Z\in L^2(\Omega;\mathbb R^2) and f∈L2(Ω)f\in L^2(\Omega), the weak normal divergence condition 𝒟Ω(Z,f)\mathcal D_\Omega(Z,f) is the requirement in (5): 𝒟Ω(Z,f)⇔∫ΩZ⋅∇udx=−∫Ωfudxfor every u∈H1(Ω).\mathcal D_\Omega(Z,f) \quad\Longleftrightarrow\quad \int_\Omega Z\cdot\nabla u\,dx=-\int_\Omega f u\,dx \quad\text{for every }u\in H^1(\Omega).(5)✦ For open Ω\Omega, a scalar field c∈L2(Ω)c\in L^2(\Omega) is a weak curl of ZZ when the condition 𝒞Ω(Z,c)\mathcal C_\Omega(Z,c) in (6) holds: 𝒞Ω(Z,c)⇔∫ΩZ⋅J∇jdx=−∫Ωcjdxfor every j∈H01(Ω).\mathcal C_\Omega(Z,c) \quad\Longleftrightarrow\quad \int_\Omega Z\cdot J\nabla j\,dx=-\int_\Omega c j\,dx \quad\text{for every }j\in H_0^1(\Omega).(6)✦ For a real spectral shift μ\mu, the shifted divergence-curl energy QμQ_\mu on H1(Ω)2H^1(\Omega)^2 is the bilinear form defined for X,Y∈H1(Ω)2X,Y\in H^1(\Omega)^2 by (7): Qμ(X,Y)=∫Ω(div⁡X)(div⁡Y)dx+∫Ω(curl⁡X)(curl⁡Y)dx−μ∫Ω𝑿⋅𝒀dx.Q_\mu(X,Y) =\int_\Omega (\operatorname{div}X)(\operatorname{div}Y)\,dx +\int_\Omega (\operatorname{curl}X)(\operatorname{curl}Y)\,dx -\mu\int_\Omega\mathbf X\cdot\mathbf Y\,dx.(7)✦ The set 𝒯(Ω)\mathcal T(\Omega) of weakly tangent Sobolev vector fields on Ω\Omega consists of the fields X∈H1(Ω)2X\in H^1(\Omega)^2 satisfying the weak normal divergence condition specified in (8): 𝒯(Ω)={X∈H1(Ω)2:𝒟Ω(𝑿,div⁡X)}.\mathcal T(\Omega) =\bigl\{X\in H^1(\Omega)^2: \mathcal D_\Omega(\mathbf X,\operatorname{div}X)\bigr\}.(8)✦

What Is Checked

These definitions are stated over real scalar and planar vector L2L^2 spaces and pairs of H1H^1 functions on Ω\Omega, with μ∈ℝ\mu\in\mathbb R and with Ω\Omega open for weak curl.