Follow an exact flow. Unfold a toral cover. Let a quantum circuit choose a heat-channel diffusivity. Keep every connection mathematically explicit.
Four finite, reproducible laboratories. The scaling view illustrates exponents; it does not simulate the released singular solution.
The models run locally in this browser.
01
An exact flow on the flat torus.
This Taylor–Green solution satisfies the unforced, incompressible two-dimensional Navier–Stokes equations. Time and viscosity change the amplitude without changing the vortex pattern.
Analytic solution
u = (sin x cos y, −cos x sin y)e−2νt · p = ¼(cos 2x + cos 2y)e−4νt
Teal and ochre encode opposite vorticity signs; arrows give velocity direction. The doughnut is a display of periodic coordinates. The PDE uses the flat metric dx² + dy², not the curved metric induced by this embedding.
Read the exact identity and the geometric boundary
On (ℝ / 2πℤ)², ω = ∂ₓu₂ − ∂ᵧu₁ = 2 sin x sin y e−2νt. The mean energy is E = ½ average |u|² = e−4νt/4, and dE/dt = −νe−4νt. The pressure gradient cancels the nonlinear convective term. The velocity, pressure, divergence, curl and energy are checked independently against symbolic expressions.
The surface display transports coordinate arrows through a torus parameterization. Its visual arrow lengths are not intrinsic speed measurements. This exact two-dimensional flow is not an example of three-dimensional finite-time blowup.
02
A quadratic order becomes a toral cover.
The auxiliary integer matrix J = [[3, 1], [1, 5]] represents 4 + √2 in a faithful realization of ℤ[√2]. Its powers give exact degree counts and transform Fourier characters by their transpose.
Integer structure
All preimages are shown for the selected power. The gold ring marks the selected input; the teal cross marks its image in the same unit square. Opposite sides are identified.
—
Degree det(Jᵏ)
14
Distinct preimages shown
14
Image Jᵏx mod 1
—
Pulled-back frequency
—
Character identity evaluated at the selected point.
See the order, covering degree, and complex-structure obstruction
Set A = J − 4I. Then A² = 2I and M(a,b) = aI + bA = [[a − b, b], [b, a + b]]. Matrix multiplication gives M(a,b)M(c,d) = M(ac + 2bd, ad + bc), while det M(a,b) = a² − 2b². Thus det J = 14, and the toral covering x ↦ Jᵏx has degree 14ᵏ. The complete fibres use integer quotient representatives, rather than randomly sampled points.
J is not complex-linear for any real linear complex structure: no commuting real matrix I satisfies I² = −1. This arithmetic covering is not automatically an endomorphism of a complex elliptic curve. The displayed identities contribute an explicit order-theoretic interpretation of one auxiliary construction; they do not replace the released proof’s analytic estimates.
03
Quantum features, a positive heat channel.
Execute seven gates on two qubits. Three expectation values set a bounded positive diffusivity. A separate, periodic position state then evolves by a completely positive, trace-preserving random-translation channel.
CPU state-vector simulation
The circuit is parameterized and untrained. The position view displays diagonal density matrices on 256 sites. This example establishes no quantum advantage or transfer of the trained ToricGT model to fluids.
q₀ is the most significant basis bit. Four θ parameters each occur once; final rotations 0.4 and −0.3 radians are fixed. Readouts: ⟨Z₀⟩, ⟨Z₁⟩ and ⟨X₀X₁⟩.
Circuit norm²
—
νθ ∈ (0.02, 0.30)
—
r = νθ Δt / Δx²
—
∂νθ / ∂θ₀
—
Three expectation values are computed from all four complex amplitudes.
The 16 × 16 grid represents a diagonal density matrix on a 256-dimensional position space (eight position qubits). It is separate from the two-qubit feature circuit. Darker squares carry greater probability.
Total probability
—
Entropy, bits
—
Entropy increase
—
Quadratic contrast
—
The maximum possible position entropy is 8 bits.
Feature-circuit computational-basis probabilities
Basis
Probability
Amplitude
The curves show this computation. Entropy uses the left panel in bits. Contrast is ½Σ(pᵢ − 1/256)² in the right panel; it is not fluid kinetic energy.
Φr(ρ) = (1 − 4r)ρ + r ∑δ = ±e₁, ±e₂ UδρUδ†, 0 ≤ r ≤ ¼
Inspect the channel, gradient rule, and representation limits
The circuit computes s = 1.4⟨Z₀⟩ − 0.8⟨Z₁⟩ + 0.6⟨X₀X₁⟩ and νθ = 0.02 + 0.28/(1 + e−s). The position grid has Δx = 2π/16 and Δt = 0.1, so r < 0.195 < ¼. Each Uδ is a periodic translation, hence unitary. Kraus operators √(1 − 4r)I and √rUδ give complete positivity and trace preservation. The same mixture is unital. For the displayed diagonal states, von Neumann entropy equals the Shannon entropy computed from the probabilities.
Each parameter-shift feature derivative is [f(θ + πeⱼ/2) − f(θ − πeⱼ/2)]/2. The diffusivity derivative then uses the sigmoid chain rule. The code does not apply parameter shift directly to the nonlinear sigmoid. Independent dense-matrix references check all gates, expectations, norm and derivatives.
The displayed position calculation exactly follows the diagonal sector of this quantum channel; it does not evolve arbitrary coherences. The independent Python checks also exercise a coherent mixed state and Kraus/Choi identities on a separate 3 × 3 position grid. This is a diffusion channel, not a quantum implementation of incompressible Navier–Stokes. Inspect the separately trained quantum ToricGT experiment →
04
Increasing amplitude, decreasing core energy.
A shrinking region can support growing velocity while its integrated energy decreases. Follow the exponents separately to see why a finite-energy condition alone cannot exclude a singular amplitude.
Scaling illustration
These power laws are an auxiliary scaling illustration inspired by the proof’s parameters. They are not an evaluation or simulation of its constructed Navier–Stokes solution.
Every quantity is normalized to one at τ = 1. The plot uses logarithmic coordinates; it does not show resolved velocity samples or an error-controlled approximation to the upstream solution.
Current scaling values and their exponents
Quantity
Power of τ
Current ratio
Limit as τ → 0
U² · volume ∼ τ−1−2h τ3/2−h = τ1/2−3h → 0 for h < 1/6.
Why the scaling calculation is insufficient as a blowup proof
The formal radial and axial scales are τ1/2 and τ1/2−h, velocity is τ−1/2−h, and volume is τ3/2−h. Their energy product has exponent 1/2 − 3h, positive throughout the displayed interval. This accounts for compatibility of growth and finite energy.
A singularity proof must also construct an actual divergence-free velocity and pressure, control the Navier–Stokes residual, and show the forcing has the required smoothness and decay through the singular time. The algebraic exponents do not supply those estimates. The research chapter reviews that construction separately and identifies exactly where the finite complex and order interpretation applies.
COMPUTATION & PROOF
Checks that follow the mathematics.
Independent reference records accompany the executable kernels.
Finite complexes and energy
Periodic grids L = 3,…,8 are checked over the integers for d₁d₀ = 0. Modular rank lower bounds and exact null-vector upper bounds certify Betti numbers (1, 2, 1). Hodge matrices are checked against their sum-of-squares quadratic forms.
Symbolic differentiation checks the manufactured flow. Dense NumPy matrices independently check the browser circuit. Fourier-mode and coherent-state channel tests verify more than a matching picture.