Context
Industry — Honeywell Technology Solutions
Role
Sr. Mechanical Design Engineer
Duration
2019 – 2023
Stack
ANSYS CFX, ICEM, Python, VBA, SAP

Note: all Honeywell work is covered by an IP agreement. This page describes the physics, methodology, and skillset — no proprietary geometry, performance maps, or design margins are shared.

Problem

Honeywell's gas-turbine blades needed continuous aerodynamic refinement under tight mass and tool-readiness constraints. Each geometry variant had to be meshed, simulated, validated against rig data, and cleared for manufacturing — and engineers were spending an outsized share of every cycle on manual data extraction from solver output, slowing iteration velocity.

Approach

  • 3D compressible RANS simulations through rotor / stator passages with SST k-ω and realisable k-ε turbulence closures, chosen based on separation sensitivity and near-wall behaviour.
  • Mesh morphing for parametric geometry sweeps — re-meshing avoided where surface topology was preserved, which cut wall-clock per design point dramatically.
  • Uncertainty quantification across tip-clearance and blade-angle tolerance windows to understand performance scatter before committing to manufacture.
  • CHT-coupled thermal analysis on blade internal cooling passages for the hot-section variants.
  • Validation against rig test data: total pressure ratio, isentropic efficiency, and stage temperature rise.
  • A Python + VBA post-processing layer: CFX solver output → metric extraction → performance-comparison dashboards consumed by the wider design review team.

Tools & setup

  • ANSYS CFX — rotating machinery solver, turbo post-processing, CHT.
  • ANSYS Meshing + ICEM CFD — structured hexahedral grids for periodic passages.
  • Python (NumPy, Pandas, Matplotlib) — automated result parsing, performance dashboards, parametric sensitivity reports.
  • Excel VBA — interactive performance comparison dashboards for engineering review.
  • SAP / ERP — cross-functional coordination with test and manufacturing teams.

Results

26%
Reduction in manual analysis time
$33K
Annual engineering savings
±3%
Stage efficiency vs. rig data
On-time
Mass-optimised geometries, no schedule slip

Tie-in to bachelor's thesis: The turbomachinery work built directly on my VIT undergraduate research — tip-leakage losses in axial compressors and the aerodynamic effect of shrouded blades, squealer tips, winglets, labyrinth seals, casing treatments, and endwall contouring. Understanding the near-tip flow physics from that study shaped how I approached every stage simulation at Honeywell.

Validation

Stage-level metrics were cross-checked against Honeywell rig test campaigns. Total pressure ratio, isentropic efficiency, and stage temperature rise agreed within ±2–3%, which was within the combined measurement and modelling uncertainty window we had quantified up front. Where disagreements appeared, the UQ sweep was revisited rather than tuning the CFD — a practice that kept the models honest.

Lessons learned

  • The biggest lift came from automating the last mile of post-processing, not from a better solver or finer mesh. Engineers ignore results they can't see; the dashboard layer was what moved the needle.
  • Mesh morphing pays off only when surface topology is stable across the sweep. For larger geometry changes, full re-meshing was faster end-to-end.
  • UQ was initially sold internally as "nice to have", but it turned into the most persuasive slide in every design review — quantifying risk made approval decisions cleaner.

Artifacts

  • Technical brief PDFs and automated CFX post-processing workflows (Honeywell proprietary — anonymised summary available on request).