thermal-fluid-analysis

Files (2)

• SKILL.md
• openai.yaml · agents/

SKILL.md (first lines)

# Thermal-Fluid Analysis

## Purpose

Develop decision-ready thermal-fluid analysis without separating numerical output from physical meaning. State what is measured, simulated, derived, assumed, or inferred; retain units, sign conventions, property states, system boundaries, and applicability limits.

## Analysis Workflow

1. Define the system: geometry, working fluid, operating regime, boundary conditions, performance metrics, and constraints.
2. Establish the physical model: governing balances, material properties and their evaluation state, dimensionless groups, competing mechanisms, and expected limiting behavior.
3. Select evidence: standards, handbooks, peer-reviewed work, calibrated measurements, validated CFD, or correlations inside their stated range.
4. Test credibility: dimensional consistency, conservation, order of magnitude, limiting cases, uncertainty/sensitivity, and an independent analytical or empirical check where possible.
5. Conclude by mechanism and tradeoff: performance, pressure drop or pumping power, manufacturability, reliability, cost, safety, scaling, and residual risk.

## Experiments

Report the facility and specimen geometry, instrumentation, calibration, sampling, procedure, data reduction, heat-loss treatment, repeatability, uncertainty propagation, and exclusions. Distinguish sensor uncertainty from run-to-run variability. Do not call a preliminary demonstration validation without a stated comparison target.

## CFD And Modeling

List all assumptions before presenting results and justify each one. State the domain, mesh and mesh-independence evidence, models, wall treatment, property models, boundary and initial conditions, solver/convergence criteria, time step where applicable, and validation data. Treat a converged solution as numerical evidence, not experimental proof.

## Results Discussion

For each important figure, move through: what is plotted; what changes; the physical explanation; and comparison with literature, theory, or an independent data source. Do not merely restate a trend. Explain the regime, competing mechanisms, and the conditions under which the conclusion might fail.

## Final Check

Before finalizing, check regime classification, property variation, phase state, coordinate/sign conventions, uncertainty, model validity, and safety or code constraints. Escalate unknown operating conditions that materially change the conclusion.

From hanhuark/mechanical-engineering-research-skill, MIT.

More from hanhuark/mechanical-engineering-research-skill