Mechanical Engineering Research

Mechanical Engineering Research

Overview

Use this skill to research thermal-fluid systems with engineering rigor: define the question, collect reliable sources, preserve assumptions and validity ranges, and separate verified evidence from inference.

Workflow Coordination

For full paper, proposal, review-article, thesis chapter, or major manuscript workflows, use an academic-research workflow as the scaffold when one is available, and use this skill as the thermal-fluid/mechanical-engineering judgment layer.

Treat the roles as:

• Academic research workflow: organize the process, checkpoints, outline, drafting sequence, review/revision loop, citation/claim checks, and finalization.
• Mechanical engineering research skill: enforce domain logic, physical reasoning, literature synthesis standards, methodology detail, assumptions, data-analysis rigor, figure discussion, plotting, presentation quality, AI/ML interpretation, reproducible coding, and proposal-specific technical judgment.

When both are available, do not let a generic academic workflow overwrite domain judgment. Apply this skill whenever deciding whether the research question, gap, method, DOE, model assumptions, interpretation, figure narrative, or proposal significance is mechanically and thermally sound.

Research Workflow

1. Clarify the engineering objective.

   • Identify the system, working fluid, operating regime, geometry, boundary conditions, performance metric, and constraints.
   • Ask for missing high-impact values only when they determine the research path; otherwise state assumptions and proceed.

2. Build a source hierarchy.

   • Prefer standards, textbooks/handbooks, peer-reviewed papers, manufacturer datasheets, government or lab reports, and primary patents.
   • Use web search for current papers, standards status, products, prices, regulations, or citations. Cite sources with links whenever browsing is used.
   • Treat blogs, marketing pages, forum posts, and uncited summaries as orientation only unless the user explicitly asks for informal context.

3. Extract engineering substance.

   • Capture equations, correlations, dimensionless groups, material limits, empirical constants, and uncertainty.
   • Record applicability limits: Reynolds/Rayleigh/Nusselt ranges, Prandtl range, Mach/compressibility assumptions, phase-change regime, geometry, roughness, orientation, temperature/pressure range, and fluid property source.
   • Note what was measured, simulated, or assumed.

4. Compare alternatives by mechanism.

   • Explain why each design or model performs differently, not only which is "best."
   • Consider pressure drop, heat-transfer coefficient, pumping power, fouling, manufacturability, instrumentation, maintenance, safety, cost, and scaling behavior.

5. Produce a decision-ready output.

   • Lead with the answer or recommendation.
   • Include assumptions, key evidence, equations/correlations, source quality, uncertainty, and next verification steps.
   • Mark engineering inference explicitly when sources do not directly prove a claim.
   • Use a clear paragraph logic flow: each paragraph starts with a central topic sentence, and each following sentence develops, supports, qualifies, or transitions from that topic.

Output Patterns

For a research brief, use:

• Question: One sentence defining the engineering problem.
• Bottom Line: Concise answer, recommendation, or state of evidence.
• Assumptions: Operating conditions, geometry, fluid, and scope.
• Evidence: Source-backed findings with citations.
• Models/Correlations: Equations, variables, units, and validity limits.
• Tradeoffs: Performance, cost, safety, reliability, manufacturability, and uncertainty.
• Gaps: What remains unverified or standards-dependent.
• Next Steps: Calculations, experiments, simulations, or standards lookups.

For a literature review, read references/literature-review.md. Group sources by mechanism, method, design family, or unresolved question rather than listing papers chronologically.

For federal research proposals, DOE EPSCoR/National Laboratory partnership proposals, full narrative expansions, review-criteria responses, proposal figure planning, preliminary-results integration, or ready-to-submit proposal polishing, read references/proposal-development.md.

For a design comparison, include a compact decision matrix and explain the dominant physics behind each score.

For manuscript-style technical writing, read references/technical-writing-analysis.md before drafting or revising introductions, methods, modeling sections, results/discussion, data analysis, or plot narratives.

For full technical papers, journal manuscripts, or paper-style section drafting, also read references/paper-writing-style.md to match the preferred section logic, abstract style, figure-led results, and conclusion patterns.

For experiments, simulations, or parameter studies, use references/technical-writing-analysis.md to plan a detailed baseline case and hypothesis-driven DOE before proposing broad sweeps or large case matrices.

For research presentations or slide decks, read references/presentation-slides.md before creating slide outlines, slide content, speaker notes, or visual-story plans.

For AI/ML-assisted thermal-fluid research, read references/ai-tools-thermal-fluids.md before recommending computer vision, sequence regression, surrogate modeling, sensor fusion, dimensionality reduction, or data-driven control workflows.

For research code, scripts, notebooks, data pipelines, plotting code, CFD automation, or ML implementation, read references/research-coding.md before writing or reviewing code.

For research tool workflows involving Overleaf, VS Code, GitHub, or git, read references/research-toolchain.md before advising on manuscript collaboration, code/debugging workflow, repository updates, archival releases, branches, commits, tags, or reproducibility hygiene.

For innovation, invention disclosure, provisional patent support, utility patent technical content, or technology commercialization, read references/innovation-commercialization.md before drafting disclosure answers, non-confidential summaries, technical descriptions, figure lists, prior-art comparisons, market/use-case notes, inventor response emails, or commercialization briefs. Support the technical and strategic content, but do not provide legal advice; defer claim scope, filing strategy, assignments, declarations, and prosecution decisions to Technology Ventures and patent counsel.