flow.FlowStudyResult

flow.FlowStudyResult(
    name,
    kind,
    grid,
    chi,
    velocity,
    density,
    temperature,
    pressure_drop,
    peak_temperature,
    mean_temperature,
    bulk_outlet_temperature,
    thermal_resistance,
    power,
    energy_imbalance,
    reynolds,
    peclet_cell,
    richardson,
)

Everything one conjugate solve produced, fields and scalars alike.

The scalars are JAX arrays rather than floats on purpose: they are what an objective is built from, and converting them to Python floats here would sever the gradient path the whole study exists to provide.

Attributes

Name	Type	Description
name	str	The study’s name.
kind	str	StudyKind.FLOW.
grid	FlowGrid	The lattice the solve ran on.
chi	jax.Array	Solid fraction, (NX, NY, NZ).
velocity	jax.Array	Converged velocity in lattice units, (3, NX, NY, NZ); exactly zero when the inlet speed is zero.
density	jax.Array	Converged density, (NX, NY, NZ); exactly 1 with no flow.
temperature	jax.Array	Temperature rise above the inlet, (NX, NY, NZ).
pressure_drop	jax.Array	Inlet-to-outlet pressure difference, lattice units.
peak_temperature	jax.Array	Hottest cell in the solid.
mean_temperature	jax.Array	chi-weighted mean over the solid.
bulk_outlet_temperature	jax.Array	Mixing-cup temperature of the air leaving.
thermal_resistance	jax.Array	(T_peak - T_inlet) / power.
power	float	Total injected power, the sum over every heat source.
energy_imbalance	jax.Array | None	What the solve failed to conserve, as a fraction of power; None when no power was injected or when a held temperature or an isothermal wall gives heat another way out, in which case the balance is not the study’s to close.
reynolds	float	The Reynolds number solved at.
peclet_cell	float	Largest cell Peclet number in the fluid – how far the energy discretisation is from the diffusive regime.
richardson	float	g beta dT L / U^2, the buoyancy the one-way coupling neglects; see this module’s docstring.

Methods

Name	Description
warnings	Everything about this solve a reader should not have to derive.

warnings

flow.FlowStudyResult.warnings()

Everything about this solve a reader should not have to derive.

Returned rather than logged so a test can assert on them and a caller can decide how loud to be. Empty means every assumption the study makes held on the numbers it actually ran.

Returns

Name	Type	Description
	list[str]	Human-readable strings, one per assumption that did not hold.