Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck pendingverified 2026-10-23 (gpt-6.1-sol)
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Bingham yield graph and convex monotonicity

Statement

Let M3(R) be the real 3×3 matrices and I3 the identity matrix, and let S0={D∈M3(R):DT=D,tr⁡D=0}, with Frobenius product D:E=tr⁡(DTE) and norm ∣D∣=(D:D)1/2. Define SO(3):={Q∈M3(R):QTQ=I3,det⁡Q=1}. For D measured in s−1, let μp≥0 be in Pa s and τy≥0 be in Pa; define Φ(D)=μp∣D∣2+τy∣D∣, in Pa s−1, and

∂Φ(D):={τ∈S0:Φ(E)≥Φ(D)+τ:(E−D) for every E∈S0}.

Define the Bingham graph by B(D)={2μpD+τyD/∣D∣} when D≠0, and by B(0)={τ∈S0:∣τ∣≤τy} when D=0; its graph stresses are in Pa. Then Φ is finite, nonnegative, and convex; B(D)=∂Φ(D) for every D; the graph is monotone in the sense that (τ1−τ2):(D1−D2)≥0 whenever τi∈B(Di); every graph value satisfies τ:D≥0; and, for Q∈SO(3), B(QDQT)=QB(D)QT.

Facts & Assumptions

Given: The finite-dimensional real inner-product space S0, the parameters μp,τy≥0, the potential Φ, and the graph B in the Statement.

[A1]

Generalized Newtonian, Bingham and conformation model equations: the Bingham convention is τ=2μpD+τyD/∣D∣ for D≠0, and the zero-rate value is the full trace-free Frobenius ball of radius τy.

Proof

Given: The data in the Statement and Facts & Assumptions.

1.1givenalgebra

The norm is convex by the triangle inequality and positive homogeneity. For 0≤s≤1, ∣sD+(1−s)E∣2≤(s∣D∣+(1−s)∣E∣)2≤s∣D∣2+(1−s)∣E∣2, the second inequality following from s(1−s)(∣D∣−∣E∣)2≥0. Thus Φ is convex; it is finite everywhere and Φ≥0, so the supporting-inequality set defined in the Statement is well-defined.

1.2givenalgebra

Let D≠0. For every E∈S0, expansion of ∣E−D∣2≥0 gives ∣E∣2≥∣D∣2+2D:(E−D), while Cauchy–Schwarz gives ∣E∣≥∣D∣+(D/∣D∣):(E−D). Multiplying by μp and τy and adding shows Φ(E)≥Φ(D)+(2μpD+τyD/∣D∣):(E−D), so the stated nonzero graph value belongs to the set defined in the Statement.

1.3givenalgebra

At D=0, the supporting inequality is μp∣E∣2+τy∣E∣≥ξ:E for every E. If ∣ξ∣≤τy, Cauchy–Schwarz proves it. Conversely, if ∣ξ∣>τy, choose E=tξ/∣ξ∣; the inequality would imply ∣ξ∣≤μpt+τy for every t>0, which fails for sufficiently small t when μp>0, and fails for every t>0 when μp=0. Hence ∂Φ(0) is exactly the stated stress ball, proving B=∂Φ also at zero.

2.1givenstep 1.2algebra

If ξ belongs to the set defined as ∂Φ(D) with D≠0, apply its supporting inequality to E=D+hH. For h>0 division by h gives ξ:H≤(Φ(D+hH)−Φ(D))/h; for h<0 the reversed inequality gives ξ:H≥(Φ(D+hH)−Φ(D))/h. The identity (∣D+hH∣−∣D∣)/h=(2D:H+h∣H∣2)/(∣D+hH∣+∣D∣)→(D:H)/∣D∣ and the quadratic expansion in step 1.2 show that both limits equal (2μpD+τyD/∣D∣):H. Since this holds for every H∈S0, nondegeneracy of the Frobenius product gives ξ=2μpD+τyD/∣D∣.

2.2givenA1step 1.3algebra

If D≠0, then τ:D=2μp∣D∣2+τy∣D∣≥0; if D=0, then τ:D=0 for every point of the stress ball.

3.1givenstep 1.2step 2.1step 1.3

For τi∈∂Φ(Di), the defining inequality gives Φ(D2)≥Φ(D1)+τ1:(D2−D1) and Φ(D1)≥Φ(D2)+τ2:(D1−D2). Adding and cancelling the potential values yields (τ1−τ2):(D1−D2)≥0.

4.1givenA1algebra∎

Conjugation by Q∈SO(3) preserves symmetry, trace, Frobenius norm, and the Frobenius product. Substituting QDQT into each nonzero formula therefore gives QτQT; at D=0, the full ball maps onto itself because ∣QτQT∣=∣τ∣. This proves the stated rotation-equivariance.

Depends on

Dependency tree · 2 levels

Sources