Portfolio

Composite design and structural optimisation across two launch vehicle components. Parametric FEM study comparing monolithic, sandwich, and stiffened-monolithic fairing architectures against buckling safety factor, mass, cost, and manufacturability. Payload plate design and FEM validation focused on mass reduction and vibration behaviour under launch loads.

372 parametric FEM simulations across four laminate strategies combining spread-tow woven carbon (XPREG XC130, 88 gsm) and unidirectional carbon (150 gsm). Static load of 0.51 MPa representing a 90 kg rider. Optimal strategy: 6 UD plies, 72 g, 2.7 mm deflection, SF 1.5, best stiffness-to-mass ratio.

Design and FEM simulation of a large CFRP U-shaped component (850 mm depth, 23 kg) for industrial robotics. TFP-optimised multi-layer laminate built as two bonded half-shells to avoid rivets. Adhesive shear limit (11 MPa) verified across all bond lines under 3 load cases. First natural frequency at 19 Hz, nozzle relative displacement <0.33 mm.

Parametric Python-driven FEM workflow coupled to a deep learning algorithm for multi-objective optimisation of composite space structures (fairing, tanks, payload adapters). Automated geometry construction, laminate stacking definition, and batch job submission enabling exploration of complex design spaces.

Zone-partitioned FEM optimisation of a submarine pressure hull across 9 structural zones, each assigned for pressure load, core density class (130/150/200 kg/m³), laminate stacking, and deflection target. 12+ configurations iterated. Mass reduced 14% vs initial target, wet-front zone stiffness increased >60%, tensile SF 2.9 and compressive SF 2.3 achieved.

Structural sizing of a pressurised CFRP shell for a near-space passenger capsule. T800S/3900-2 prepreg UD, laminate sized for 0.1 MPa internal pressure (SF 24.8), out-of-autoclave + AFP manufacturing strategy (void content <2%), phased certification plan including CAI, fatigue, and burst testing.

Composite-native interlocking plate architecture replacing a tubular truss design adapted from metallic construction. All primary adhesive joints designed to work in pure shear. FEM impact study across three severity levels: 500g (elastic survival at 18 m/s), 2000g (delamination risk), and 3500g (sacrificial crumple zone strategy). 15-25% weight saving over reference design.

Development of a novel hourglass-shaped specimen geometry ensuring failure localisation in the gauge section under fibre-direction compression. Experimental characterisation of carbon/PEEK and glass/epoxy laminates by compression and four-point bending, full-field strain measurement by 3D DIC up to fracture, and progressive damage model implementation.

Three-point bending FEM model used to identify axial modulus E11 and failure strain from physical test data on three pultruded GFRP tube geometries. Identified E11: 15 GPa (thin-wall) and 30 GPa (thick-wall). Properties applied to a 350 mm cantilever spacer model under 400 kg sustained load across 150+ parametric simulations. Long-term creep assessed per Eurocomp knock-down factor (k=0.5), leading to a larger-diameter tube recommendation to meet civil safety factor ≥5.

Compact quadrotor with asymmetric sandwich structure (6mm top plate, 3mm bottom plate), vibration-isolated 200x60mm M-LOK mission deck, and integrated prop protection. Validated via 500g impact simulation at 18 m/s, capturing structural strain limits for 2000g and 3500g crash conditions.

Structural optimisation of the central hinge bracket of an autonomous mountain mowing vehicle across three load cases: maximum braking, maximum acceleration, and static 45° slope. Mass reduced 37% (43 kg to 27 kg) in S355 steel, safety factors above 6 against endurance limit. Full ±73° turning angle preserved.

Structural redesign of an injection-moulded ASA guitar case to reduce mass from 9 lbs to 6.7 lbs (26% reduction) while maintaining structural integrity under buckling and material strength requirements. Redesigned internal rib system to improve rigidity and strength while compensating for reduced wall thickness. Design balanced against injection moulding process constraints (draft angles, wall uniformity, tool complexity).

Five successive material removal iterations on a steel adapter, each validated by FEM under a 17 kN combined load. Mass reduced 26% (0.422 kg to 0.312 kg) with maximum stress held below 200 MPa throughout. Load method independently validated by virtual node comparison.

Modal and forced-vibration FEM of a 770mm GNSS antenna housing per IEC 60945 sweep requirements (2-100 Hz, 7 m/s² base excitation). Structure redesigned from an initial complex-geometry configuration to a simplified, manufacturability-driven architecture, reducing structural complexity while maintaining dynamic performance. Resonance response improved from a critical peak near the qualification band to a safer margin, with maximum stress and displacement both substantially reduced.

Comparative FEM of four structural variants for a 6.2 m steel chassis under four load cases including eccentric pallet loading and lateral forces. Mass reduced by 44% (1091 kg to 609 kg), maximum stress 134 MPa, safety factor 2.6 on S355 yield strength.

Design and execution of a combined torsion-tension test campaign on full-scale composite helicopter blades. Custom test fixture designed in parametric CAD, full-field strain measurement by 3D DIC (ARAMIS), FEM model construction and test-analysis correlation across all load levels.

Instrumented inflation campaign on large-format CFRP membranes (1500×1500 mm²). Fabrication support, adaptation of pressure test device, full-field strain measurement by 3D DIC, and dedicated data acquisition unit (pressure + gauges, Arduino).

Full nonlinear FEM of a steel beam-to-column connection including 18 M12 Grade 8.8 bolts with 35 kN preload, frictional contact, and geometric nonlinearity. Force-strain correlation against physical test results within <8% deviation. Captured elastic, transition, and yielding phases matching experimental failure load.

Reconstruction of a 10-layer PCB from Gerber files into a CAD model (400 MB STEP), including all copper layers, prepreg, vias, and pins. Homogenised orthotropic material properties derived by thermal simulation. Thermo-mechanical FEM under 2.23 W heat load showed equilibrium rise of 27.4°C and max deformation of 9.2 µm within optical tolerances.

Thermal FEM of an aluminium-substrate PCB assembly for high-power LED lighting. Temperature field prediction under operational heat flux, hotspot identification, and margin analysis against component thermal limits.

Conjugate heat transfer simulation of lithium-ion battery pack with serpentine cold plate cooling under multiple PowerShot discharge cycles. Temperature field extraction, cell-level margin analysis, and Python post-processing pipeline for parametric cold plate design sweeps.

Load-driven laminate design connecting a carbon racing fin to a lead keel bulb via a 6x M10 titanium bolted joint, validated for hard-grounding impact. Staircase ply-drop architecture transitions a 48-ply solid sole into the fin's 23-ply sidewall laminate without a stress-concentrating knife-edge, with each connection element checked against its own governing failure mode.

Single-frustum welded aluminium nose cone and boattail for a demonstrator rocket, designed around manufacturability from the outset: lap-strip plug-weld joints, ring stiffening tuned for compound-curvature forming, and a complete DXF flat-pattern package with full weld callouts ready for direct shop-floor use.

Wall-thickness optimization of an extruded aluminium mounting rail, driven by a von Mises equivalent strain failure criterion validated against physical test data and post-mortem yield-zone correlation. Python-automated geometry iteration swept four cross-section parameters against the validated criterion to hold target load capacity at unchanged linear mass.

2-DOF cardan hinge for an articulated UGV, sized by Goodman fatigue analysis and load-specific bearing selection, built around a spigot-located bolted stack and fully split, dowel-aligned collar and yoke architecture. Every interface is field-serviceable with standard spanners — no press, puller, or heat required at any disassembly step.

Parallel steel-vs-aluminium and welded-plate-vs-bent-sheet trade-off study for an autonomous UGV chassis, with thickness parametric FEM optimisation converging on a 38 kg steel design at SF 1.5. Motor interface, drawer-mounted battery integration, and a removable top plate were built in from the outset for hands-on serviceability.