PROJECT_FORMAT.md

# Pipe project and dynamic response · v1.3

A saved project is JSON, not executable code. All stored numbers use SI units.
`schema` is `agentfem.pipe-project/1`; `modelVersion` is `1.3.0`.
Version 1.1.0 and 1.2.0 imports remain supported and receive default excitation settings.
Optional `design.excitation` is `{node:3, axis:2, force:100, frequency:10, zeta:0.02}`.
The harmonic force and probe share a joint and global direction. SI units, peak
amplitude, not RMS. The added damper retains independent position/direction.

```json
{
  "schema": "agentfem.pipe-project/1",
  "modelVersion": "1.3.0",
  "units": "SI",
  "design": {
    "route": [
      {"kind": "straight", "ratio": 5, "roll": 0},
      {"kind": "elbow", "ratio": 2, "roll": 0},
      {"kind": "straight", "ratio": 5, "roll": 0}
    ],
    "diameter": 0.2,
    "wall": 0.06,
    "supports": {"2": {"type": "spring", "k": 100000}},
    "temperature": 100,
    "gravity": true,
    "damper": {"node": 3, "axis": 2, "c": 1000}
  }
}
```

`wall` is t/D; straight `ratio` is L/D, elbow `ratio` is R/D. Elbows also include
two straight tangents, each 2D long. `roll` is degrees about the incoming local
axis. Component count is 1–10. Dimensions are selected from validated catalog
entries; arbitrary dimensions and branching are not supported by this release.

Joint 0 is fixed in all six mechanical DOFs. Support nodes 1…N may be `none`,
`spring` (global Z, k in N/m), `guide` (global Z displacement constrained), or
`fixed` (all six DOFs). Grounded viscous damper axes 0/1/2 mean global X/Y/Z,
with c in N·s/m. It can coexist with a support. A constrained damper direction
has zero relative velocity and therefore no dissipative effect.

Removal policy: deleting segment i removes its downstream joint i+1 and its
attachments. Downstream surviving joints are renumbered. Upstream attachments
remain. A damper attached to the removed joint is disabled (c=0). Undo restores
the entire prior design. Saving stores the design, not cached numerical answers;
opening revalidates and recomputes it. Local drafts stay in browser storage.

## Dynamic experiment

The 12 lowest mass-normalized system modes are retained from the assembled
Craig–Bampton matrices. A single grounded damper gives C=c b bᵀ, where b contains
the selected joint/direction entries of those modes. This full matrix is kept;
cross-modal damping terms are not discarded. There is no background damping.

Initial displacement is the first undamped mode, normalized so its largest
centreline coordinate component is +0.001 m; initial velocity is zero. Motion is
a perturbation about the linear equilibrium. The plotted geometry shows this
perturbation about the undeformed reference to isolate vibration. There is no
thermal geometric stiffness or temperature-dependent material model.

Newmark average acceleration (β=1/4, γ=1/2) integrates for six first-mode periods.
The integration step is at most 1/(80 f_max) and at least 2400 steps per record,
rounded to a multiple of 600 for 601 output samples. Animation playback lasts
8 seconds; chart time and exported time are physical seconds. Forces and force
peaks are sampled at those 601 times. The no-damper reference is analytic cosine.

The independent Python reference uses a matrix exponential of the same modal
equations. This verifies the time solver, not real-plant accuracy. Mode-retention
and energy-balance checks are provided separately. The release does not simulate
hydraulic snubbers, gaps, preload, seismic excitation, pressure, fluid mass,
plasticity, fatigue or piping-code compliance.

## Delivery

The comparison report shows unamplified static results and current support
forces/moments, along with material constants, model version and complete project
settings. Print-to-PDF uses the browser. CSV includes port translations, support
forces/moments and, when a dynamic run exists, its time histories. The reference
and current designs are included in the printable report.

## Power sections

v1.3 additionally accepts (OD,wall thickness) = (0.45,0.06), (0.8,0.03),
(0.8,0.02), (0.45,0.05) metres. Store wall as thickness/OD. Only those exact
pairs are available; ΔT is 0–600 K for these research sections. Generic pipes
retain the 0–200 K range. See POWER_CASES.md for their fixed material constants.