Research-Stack/3-Mathematical-Models/AMMR/HEAT2D_ADAPTER.md

112 lines
3 KiB
Markdown

# Heat-2D Adapter
This repository is the 2D diffusion / dissipative-residue sandbox for the pressure-cavitation-vibration model catalog.
## Core role
```text
shock / pressure / vibration / combustion / friction event
-> deposited scalar residue
-> 2D diffusion
-> measurable afterimage field
```
In the broader stack:
```text
Burgers handles shock smoothing.
heat-2D handles deposited residue.
```
## Native PDE
The upstream README defines the supported PDE family as:
```math
\partial_t u = \nabla\cdot(\alpha\nabla u)+f
```
where `u(x,y,t)` is the diffusive variable, `alpha(x,y)` is a space-dependent diffusivity, and `f(x,y,t)` is a source term.
For catalog integration, reinterpret these as:
| Field | Catalog interpretation |
|---|---|
| `u` | residual field: heat, deposited shock energy, vibration residue, or pressure-afterimage proxy |
| `alpha` | material / geometry-dependent diffusivity |
| `f` | source term from cavitation, combustion, friction, impact, or acoustic deposit |
| boundary conditions | heat/pressure residue sinks, walls, insulation, or imposed driving surfaces |
## Adapter definition
```math
X_{heat2D}=(u,\alpha,f,\Omega,\partial\Omega,\Delta x,\Delta y,\Delta t,B)
```
```math
\alpha_{heat2D}:X_{heat2D}\rightarrow U_T(x,y,t)
```
The adapter is admissible when:
1. boundary conditions are explicit,
2. source units are documented,
3. geometry masks are kept separate from solver logic,
4. stability requirements are checked for explicit integration,
5. implicit or Crank-Nicolson runs include linear-solver receipt metadata.
## Candidate source projections
### Cavitation collapse deposit
```math
f(x,y,t)=E_{shock}\exp\left(-\frac{(x-x_0)^2+(y-y_0)^2}{2\sigma_s^2}\right)\delta(t-t_0)
```
### Rocket cooling proxy
```math
\partial_t T=\nabla\cdot(\alpha\nabla T)+q''_{combustion}-q''_{coolant}
```
### Passive fluidic residue map
Use `alpha(x,y)` to encode channels, walls, pockets, and damping zones:
```math
\partial_t u=\nabla\cdot(\alpha(x,y)\nabla u)+f
```
### Burgers dissipative afterimage
```math
f_{diss}(x,y,t)=\lambda\nu(\partial_x U)^2
```
## Minimal tests
| Test | Purpose | Expected behavior |
|---|---|---|
| point impulse | pure diffusion sanity check | radial spread, decaying max |
| line source | shock-front residue proxy | front thickens like sqrt(t) |
| geometry mask | passive fluidic damping test | pockets and channels alter residue storage |
| cooling channel | rocket wall proxy | stable gradient under balanced source/sink |
## Failure modes
| Failure | Meaning |
|---|---|
| oscillation in pure explicit heat solve | timestep instability |
| negative values from nonnegative initial/source field | instability or source bug |
| energy growth with no source | boundary/source accounting error |
| claiming CFD from diffusion | category error |
## Scope warning
This repository should not be treated as a full Navier-Stokes, cavitation, rocket-engine, or acoustic solver.
Its correct role is narrower and more useful:
```text
Given a source event, what dissipative residue field remains?
```