Team Charter — Computational Engineering

Published

May 8, 2026

Mission

Predict, before each test, what the rig should do. After each test, compare prediction to measurement and explain the gap. You don’t write the simulation — the teacher does, in Rust — but you own the interpretation and reconciliation of its outputs. This is the analyst role of real engineering practice: the person who runs the model is not the person who runs the test.

Why this team exists

In a real engineering organisation the design analyst, the test engineer, and the manufacturing engineer are three different people who must agree before money is spent on hardware. This team plays the analyst role for the program. Their job is to be sceptical of measured data when it disagrees with prediction, and sceptical of predictions when they disagree with data. That tension is the engineering method.

Roles

Role Responsibility
Lead Schedules team work, attends cross-team standup
Deputy Lead Steps up if Lead absent; owns ICD interface checks
Verification Officer Confirms the prediction sheet matches the test plan before each run
Reader Owns digesting the Rust output; produces 1-page prediction sheets
Recorder Logs all prediction-vs-measurement comparisons
Reconciler Writes the gap analysis: why measured ≠ predicted

What the teacher (Computational Engineering Architect) provides

The teacher runs a Rust simulation — helios_rocket_static_test — that produces per design iteration:

  1. Chamber pressure during the burn (kPa vs t)
  2. Wall temperature at 3 locations (inlet end, throat region, exit cone) for the chosen chamber type (Option H monolithic or Option L laminated copper)
  3. Thrust estimate (N)
  4. Propellant equivalence ratio for the chosen fuel (LPG or ethanol) with optional O₂ enrichment fraction
  5. H₂O₂ / KMnO₄ decomposition rate prediction as a function of [H₂O₂], [KMnO₄], and T (for the school-side rates calibration), plus the Arrhenius extrapolation from 3-6% student data to the 35% teacher demonstration
  6. Stoichiometric check: given the measured rate at 35% and measured ethanol calorimetry, does the chemistry supply enough O₂ for the design ethanol burn at 0.5 g/min?
  7. Safety status — whether the predicted operating point is within the declared school + vendor envelope

These come out as YAML/Markdown prediction sheets and CSV time series. The team’s job is to render them as one-page handouts the rest of the program can read.

Deliverables

ID Deliverable Due Consumer
CE-D1 Prediction sheet for H₂O₂/KMnO₄ rates calibration lab + 35% extrapolation W2L4 (morning of) Rocket Chemistry (compares live), Lab Report 1
CE-D2 Prediction sheet for school LPG atmospheric demo W3L10 Heat Exchanger (sets instrumentation plan), Lab Report 2
CE-D3 Reconciliation: predicted vs measured for school LPG demo W3L12 Lab Report 2
CE-D4 Prediction sheet for vendor burn — both Option H monolithic and Option L laminated thermal profiles, plus chosen propellant (LPG or ethanol) and O₂ enrichment fraction W4L14 Heat Exchanger (reconciliation), Manufacturing (vendor RFQ supports), Lab Report 3
CE-D5 Final reconciliation report: school demo + vendor burn (if data has arrived) + sources of error W4L3 Lab Report 4 + presentation

Syllabus dot points owned

NSW Stage 5 Working Scientifically:

  • WS5.2 Predicting (this team’s whole purpose)
  • WS7.3 Analysing data (residual analysis)
  • WS8 Problem solving (when prediction and measurement disagree)
  • WS9 Communicating (the prediction sheet is a formal handoff document)

Weekly milestones

Week Milestone Evidence
W1 Team understands structure of a prediction sheet Practice sheet on iodine-clock analogue
W2 First two prediction sheets delivered; first reconciliation done CE-D1, CE-D2
W3 Full school-demo prediction + reconciliation cycle CE-D3
W4 Vendor-burn prediction (both H and L thermal profiles); final reconciliation CE-D4, CE-D5

The reconciliation template (this team owns its quality)

Every test produces a row:

Quantity Predicted Measured Residual Explanation
Peak chamber wall T (inlet) 95 °C 88 °C -7 °C Vendor TC mounted 2 mm from gas-side surface — thermal lag
Peak chamber wall T (throat) 510 °C 470 °C -40 °C Copper conductivity higher than literature value used in sim, or hot-zone shorter than modelled
Burn duration 4.5 s 4.3 s -0.2 s Vendor operator closed valve 0.2 s early — log timestamps support
Peak thrust 1.4 N 1.1 N -0.3 N Nozzle expansion not at design Mach number (expansion ratio mismatch with chamber pressure)

A “good” reconciliation explains the residual with a named, plausible mechanism. A “great” reconciliation suggests how the next test could discriminate between two competing explanations.

What the Rust crate now models (v0.2)

Compared to v0.1 (which modelled methane + air + bolted steel laminate):

  • Propellant: Propellant::LPG or Propellant::Ethanol
  • O₂ enrichment fraction: 0.0 (air) to 1.0 (pure O₂)
  • Chamber type: ChamberType::MonolithicCopper (Option H) or ChamberType::LaminatedCopper (Option L)
  • Per-chamber-type thermal properties: copper thermal conductivity (~400 W/m·K) and chamber-type-dependent interfacial resistance for laminated (small but non-zero between layers)
  • Lumped thermal model: 3-zone (inlet end, throat region, exit cone) rather than per-plate
  • Safety envelope: now triggers on vendor limits (vendor’s pressure cert + cell wall T) rather than school limits, since the burn is off-site

The school-side LPG atmospheric demo uses a separate code path that models the brass-tee rig + plastic form study at atmospheric pressure.

Cross-team interfaces

  • From Manufacturing: chamber geometry (Option H or L), TC port locations on TA-04 wall
  • From Materials: copper thermal properties (k, ρ, c_p), laminated vs monolithic interlayer effects
  • From Rocket Chemistry: propellant choice (LPG vs ethanol), enrichment fraction
  • From Heat Exchanger: instrumentation locations and types
  • To all teams: prediction sheets before each test
  • To all teams: reconciliation reports after each test