The problem I needed to solve

My PhD research investigates whether aperiodic exit guide vanes can mitigate the aerodynamic degradation associated with boundary-layer-ingesting fans. These designs deliberately vary blade geometry around the annulus, but the commercial turbomachinery meshing tools available to the research did not support the required structured, conformal full-annulus grids. I therefore helped develop the missing capability rather than restrict the design study to periodic blades.

What I developed

I wrote most of the Python code for geometry preprocessing, coordinate transformations, passage boundary construction, and surface generation. The passage-centred OH topology enables each blade passage to be constructed individually while adjacent passages share geometric boundaries.

My supervisor developed the OpenFOAM blockMesh dictionary and the grading-related portion of the Python code. We refined the geometric methods through technical discussions and iterative implementation.

Engineering design decisions

Curve robustness. My first cross-passage construction used an ellipse fitted to endpoints and slopes, plus an additional optimized point. The constrained formulation did not reliably yield a unique result. Following my supervisor’s suggestion, I implemented a quadratic Bézier construction that provides direct control of endpoints and tangent directions without that optimization step.

Conformity by construction. Each passage is generated separately, but the topology defines shared geometric boundaries. Corresponding axial cell counts and grading parameters are averaged and applied across passages to obtain compatible node distributions rather than correcting interfaces afterward.

Verification and demonstrated capability

APT-Grid was evaluated using periodic guide vane test cases, with grid-convergence studies and aerodynamic comparisons against Ansys TurboGrid. A full-annulus aperiodic inlet-guide-vane mesh was also generated and used in a blade-resolved CFD calculation.

The manuscript provides detailed convergence and spanwise comparisons. As the manuscript is still being revised, this page does not treat its preliminary numerical differences as final publication figures.

Scope and limitations

The current software assumes compatible meridional blade projections; geometric input requirements and automated curve-processing constraints remain. Its tested capabilities should not be interpreted as universal support for arbitrary blade-row geometries.

Why build a lower-cost model?

Full-annulus blade-resolved turbomachinery simulations are expensive, especially when the inflow and blades vary circumferentially. Body-force models replace explicit blade geometry with volumetric source terms representing flow turning, blockage, profile losses, and shock losses. The aim is to retain the aerodynamic trends needed for design exploration at lower computational cost.

My engineering contributions

My supervisor developed the first version of the OpenFOAM solver; I subsequently cleaned up the implementation and corrected errors. I developed the Python geometry-preprocessing framework, building on an earlier 2D implementation by Amelia George and extending it to 3D. I also developed the Fourier-based interpolation framework, the parametric blockMesh templates, and performed the numerical assessments.

Turning discrete blade geometry into CFD inputs

The preprocessing system extracts blade camber normals, meridional arclength, tangential thickness, and local angular pitch. For aperiodic blade rows, circumferential variations are represented by truncated Fourier series with coefficients estimated by least squares, accommodating irregular circumferential sampling. Coefficients are interpolated across the meridional plane using Delaunay triangulation and barycentric interpolation before reconstruction at CFD cell centres.

Physics and verification

The framework combines established turning-force, compressibility, blockage, profile-loss, and shock-loss formulations. The equations include a spatially varying local angular pitch, supporting geometric variations beyond traditional uniformly pitched blade rows. Assessments cover an idealized zero-thickness guide vane, the ECL5/CATANA transonic fan, and a periodic/aperiodic NAX guide vane.

The ECL5/CATANA assessment captures broad efficiency trends near the design condition but has documented pressure-ratio discrepancies at off-design operating points; the paper examines stator loss modeling as one contributing mechanism. The NAX guide-vane example demonstrates circumferential flow-turning trends and three-dimensional redistribution of mass flow.

Design-study advantage

For the aperiodic NAX IGV, the paper reports a body-force full-annulus grid of approximately 5 million cells versus an estimated 60 million cells for a mesh-independent blade-resolved case. This is a cell-count comparison—not a measured speed-up in runtime.

Limits and next steps

Body-force models deliberately omit blade-scale flow structures. Their accuracy depends on the validity of the underlying formulations, operating regime, and loss models; they complement rather than replace blade-resolved CFD in a verification strategy.

Project context

Hydraulic stab plates house the couplings used to connect flying leads to subsea trees. They require a precise mechanical arrangement for the relevant couplings and mating hardware. Lifting brackets are used to lift and support stab plate assemblies during factory acceptance testing.

My design work

Using established company stab plate templates, I modified designs for individual projects and prepared drawings for manufacture. I also created manufacturing drawings for lifting brackets, documenting geometry, machining features, and assembly requirements. My role involved adapting existing designs to project-specific requirements rather than developing the original stab plate concept.

Testing and equipment verification

I personally performed factory acceptance testing of hydraulic equipment, including operating pressure-testing equipment, pressurizing couplings, monitoring pressure, checking for leakage, and recording results under established company procedures. On other occasions, I supervised the testing activities.

The company's hydraulic coupling FAT procedures cover documentation verification, low-pressure tests, proof-pressure tests, and mated coupling tests, with records of acceptance and test instrumentation. The procedure establishes the testing context; individual acceptance results are not published here.

What this work demonstrates

• Understanding of mechanical and hydraulic interface hardware.
• Project-specific engineering drawing modification and manufacturing documentation.
• Connection between hardware design, fabrication requirements, and equipment verification.
• Hands-on hydraulic pressure testing and supervision in an industrial environment.