Projects / 02 · Undergraduate thesis
Spur Gear Design Optimisation
My thesis took one part of the tilt-rotor mechanism, the spur gear pair, and asked how its tooth geometry changes the stresses it carries. I modelled eight versions of the pair in ANSYS, compared where and how strongly each one was stressed, and carried the selected geometry through to a detailed design.
Original title: Analisis dan Optimalisasi Desain Spur Gear Pada Mekanisme Tilt Rotor eVTOL Dengan Variasi Modul dan Pressure Angle Menggunakan Metode Elemen Hingga
- PERIOD
- Aug 2025 — Jan 2026
- MY ROLE
- Gear analysis and design, Capstone 2
- SOFTWARE
- SolidWorks · ANSYS Static Structural
- SUPERVISOR
- Dr. Eng. Ir. Aditya Rio Prabowo
01
Why the gears
The spur gear pair is what carries the actuator’s torque through to the tilting nacelle. The gears first came up in the Urban Air Mobility project. By Capstone 2 the wider mechanism had been reworked, and my preliminary Capstone 2 calculations pointed to the gear pair as the component with the tightest margin on contact stress. I took it on as the critical part for my thesis.
The question was narrow on purpose. How do module (tooth size) and pressure angle (tooth shape) change the stress at the tooth surface, where the teeth press together, and at the tooth root, where they bend, when everything else is held the same?
02
Eight variants, one controlled comparison
To keep the comparison fair, I fixed the size of the pair. The gear’s pitch diameter stayed at 192 mm, the pinion’s at 96 mm, and the face width at 48 mm, so the centre distance and the 2:1 ratio never changed. Only the teeth changed: a larger module meant fewer, bigger teeth, and each module was modelled at both 14.5° and 20°.
| Module | 2 mm | 3 mm | 4 mm | 6 mm |
|---|---|---|---|---|
| Teeth · gear / pinion | 96 / 48 | 64 / 32 | 48 / 24 | 32 / 16 |
| Pressure angles | 14.5° and 20° for each module | |||
Constant: pitch diameters 192 / 96 mm · face width 48 mm. Source: thesis Tables 33–34.
03
Building the ANSYS model
With the eight geometries defined, I built each gear pair in SolidWorks and analysed it in ANSYS Static Structural. Every model used the same material, supports, load and settings, so I could put the results side by side and see what the tooth geometry alone was changing. Each one went through these six steps.
-
Thesis Fig. 30 · engineering data · select to enlarge01
Material
Linear-elastic steel with E = 200 GPa and ν = 0.29. Through-hardened 400 HB steel is the assumed grade; ANSYS itself only uses the elastic inputs.
-
Thesis Fig. 31 · imported geometry · select to enlarge02
Geometry
I modelled each gear pair in SolidWorks and imported it as Parasolid, to keep the tooth topology and dimensions intact.
-
Thesis Fig. 32 · contact definition · select to enlarge03
Contact
Frictionless contact between the teeth, solved with the Augmented Lagrange formulation. Adjust to Touch closes any small gap so the flanks start in contact.
-
Thesis Fig. 33 · mesh · select to enlarge04
Mesh
Contact sizing refines the elements around the meshing teeth and roots, where the stresses peak, and keeps the rest of the model coarser.
-
Thesis Fig. 34 · supports and moment · select to enlarge05
Supports and load
In the model, the driven gear sits on a fixed support, and the pinion on a frictionless support plus a remote displacement so it can turn. These are simulation constraints, not the mechanism’s actual mounting. A torque of 582.763 N·m, carried over from the preliminary calculations, is applied to the pinion as a common assumed load for all eight.
-
Contact-zone equivalent stresspeak von Mises at the flank contactTooth-root equivalent stresspeak von Mises at the root fillet06
Reading the results
I read equivalent (von Mises) stress in two places on each model. These are FEA stress values, not Hertzian contact pressure or an AGMA tooth rating, and I kept them separate from the hand calculations.
04
Checking the method first
Before running my own gears, I wanted to know the setup gave sensible numbers. So I applied it to a spur-gear case from a published study, Hamza et al., “Design and Stress Analysis of Spur Gear Using ANSYS and SolidWorks”, and compared my contact stress with the paper’s ANSYS and SolidWorks results and with a Hertz calculation, for three materials.
| Difference from | Paper · ANSYS | Paper · SolidWorks | Hertz |
|---|---|---|---|
| Cast iron | 2.48% | 1.75% | 3.44% |
| Aluminium | 5.57% | 14.25% | 18.13% |
| Copper | 2.04% | 0.20% | 6.97% |
Margin of error in contact stress. Source: thesis Table 36 and Figs. 35–37.
For each material my result came within 10% of at least one reference, which was the threshold I had set for the method, although aluminium sat further from the SolidWorks and Hertz figures. It’s a check against published work rather than a physical test, so I used the eight-variant results mainly to compare the designs with each other.
05
What the eight models showed
In every model the equivalent stress peaked in two places: where the two flanks touch, and in the fillet at the tooth root. The interesting part was how differently those two responded as the geometry changed.
Equivalent (von Mises) stress contours, thesis Figs. 38–45 (Ch. IV). Each shows the whole gear pair; the tooth-contact view opens the enlarged region with the root probe. Every plot has its own legend range, so the colours aren’t comparable between images. Use the table below to compare values.
| Module | Contact zone · MPa | Tooth root · MPa | ||
|---|---|---|---|---|
| 14.5° | 20° | 14.5° | 20° | |
| 2 mm | 529.49 | 572.99 | 265.07 | 196.44 |
| 3 mm | 466.98 | 566.48 | 246.51 | 180.78 |
| 4 mm | 428.83 | 446.94 | 221.64 | 156.12 |
| 6 mm | 504.91 | 556.32 | 181.13 | 149.15 |
Peak equivalent (von Mises) stress from ANSYS, same torque and setup for all eight. Source: thesis Tables 37–38. Lowest contact-zone value highlighted.
Contact-zone stress dipped, then rose
At both angles, contact-zone stress fell from module 2 to module 4 and climbed again at module 6. It was higher at 20° than at 14.5° for every module.
Tooth-root stress fell steadily
Tooth-root stress dropped with every step up in module, and was lower at 20° every time. A larger pressure angle gives a thicker tooth root, but a narrower contact zone.
Peak equivalent (von Mises) values from the same eight ANSYS models. The thesis titles these charts contact and bending stress; they plot the contact-zone and tooth-root equivalent stress, not Hertzian pressure or a separate bending calculation. The original images label only the 14.5° series, so the legends beneath identify both. Select either chart to enlarge.
06
The trade-off at module 4
Module 4 gave the lowest contact-zone stress at both angles, so the real choice came down to pressure angle. Neither option was better on both counts.
In the model, 14.5° lowered the contact-zone stress by about 4% but raised the tooth-root stress by about 42%. I made the contact-zone result the deciding criterion: my preliminary calculations had already shown contact stress at the tooth surface to be the tightest margin, and the tooth-root stresses stayed at levels I judged acceptable in all eight variants. So I selected module 4 at 14.5°, the configuration with the lowest contact-zone stress under the conditions I modelled.
07
From result to gear design
With module 4 at 14.5° selected, I took the geometry back into CAD. Changing the pressure angle reshapes the teeth but keeps the pitch diameters and centre distance, so the revised gears still fit the existing shaft and housing. From the updated models I produced detailed drawings of the gear and pinion, covering their dimensions, tolerances, fits, proposed material and manufacturing notes.
WORKING DRAWINGS
The gear and pinion working drawings (thesis Figs. 54–55) carry confidentiality markings, so they aren’t reproduced here. They document a proposed design; the steel gears weren’t manufactured.
Around the study
While I worked on the gears, the wider Capstone 2 team built a small physical model of the tilt mechanism, with laser-cut acrylic gears, 3D-printed parts, a servo and Arduino control. I helped with its fabrication and electronics.