AI for Design Quality & DFM
Gear design calculations with AI cover module, ratio, and AGMA and ISO 6336 bending and contact stress, with cited standards so engineers can trust every number.
·
⏱
8 min read
Michelle Ben-David
Michelle Ben-David is a mechanical engineer and Technion graduate. She served in an IDF elite technology and intelligence unit, where she developed multidisciplinary systems integrating mechanics, electronics, and advanced algorithms. Her engineering background spans robotics, medical devices, and automotive systems.
BOTTOM LINE
Gear design calculations reward precision and punish shortcuts. AI will not stamp your drawing or take responsibility for a failed tooth, and it should not. What it can do is give you a faster, better documented first pass: the right method, the correct standard clause, material properties with a source, and a check on the factors that engineers most often miss. Treat AI as a well read assistant that shows its work, keep a responsible engineer in the loop for review, and reuse a proven gear before designing a new one. Handled that way, the math speeds up while the confidence behind each number goes up rather than down.
A gear tooth that looks correct in CAD can still pit, wear, or fracture in service if the bending stress, contact stress, or safety factor was off by a margin no one checked. Gears tend to fail quietly in the math long before they fail loudly on the test rig.
Most gear design still runs on a mix of an old spreadsheet, a printed copy of an AGMA standard, and one senior engineer's memory. That holds up until a cell reference goes stale or the standard on the shelf is two revisions behind the one your customer requires. This guide walks through how AI changes gear design calculations, from sizing the pair with module, ratio, and tooth count, to running AGMA and ISO 6336 bending and contact stress checks, to verifying every number against the actual clause instead of a copied formula.
Start With the Sizing: Module, Ratio, and Tooth Count
Before any stress check, a gear pair is defined by a small set of parameters. Getting these right sets up everything downstream, and getting them wrong quietly poisons every calculation that follows.
Module (m), the ratio of pitch diameter to number of teeth, which sets tooth size in metric designs. Imperial work uses diametral pitch instead, and mixing the two is one of the most common gear mistakes.
Number of teeth on the pinion and gear, which fixes the ratio and drives undercut risk on small pinions.
Gear ratio, the speed and torque relationship between the input and output shafts.
Pressure angle, commonly 20 degrees, which shapes the tooth profile and the direction of the tooth load.
Face width, which spreads the transmitted load across the tooth and strongly affects contact stress.
Center distance, set by the module and tooth counts and usually constrained by the housing.
It is also worth deciding early whether the design is spur or helical, because a helical gear carries an axial thrust load that the shaft and bearings must react, and the helix angle changes the effective geometry used in every strength check. On small pinions, a low tooth count can force a profile shift to avoid undercut, which then feeds back into the module and center distance. None of these choices is isolated, and a change to one ripples through the rest. These parameters are cheap to change on paper and expensive to change in tooling, so the first pass matters. The same discipline applies to the parts that surround the gear, which is why a sound design pairs gear sizing with careful bearing selection and a clear tolerance stack-up analysis from the start.
IN PRACTICE
Leo uses a Large Mechanical Model trained on 1M+ technical sources. It also provides citations, so we don't have to guess whether a material property or tolerance is correct. We see 96% accuracy on technical queries.
Dorian G., AI Engineer
AGMA and ISO 6336: The Standards That Govern Gear Strength
Two standards dominate gear rating. In North American practice, AGMA 2001 covers the strength of spur and helical gears. Internationally, ISO 6336 fills the same role. Both split the problem into two checks that fail for different reasons.
Bending strength at the tooth root, derived from the transmitted load and a geometry factor that accounts for tooth shape, then compared to a material allowable.
Contact strength at the tooth flank, a Hertzian surface stress that drives pitting and often governs on hardened gears.
Both standards then adjust the nominal stress with load, application, and dynamic factors. The application factor accounts for shock and non uniform operation of the driven machine, and the dynamic factor accounts for load added by mesh error and speed. Each check produces a safety factor, the ratio of the allowable stress to the calculated stress, and the design is only as strong as the lower of the two. Soft or through hardened gears often fail first in bending at the tooth root, while surface hardened gears usually reach their contact limit and pit before they break, so the check that governs shifts with the material and heat treatment. An engineer who reports only one number without saying which failure mode it represents has not finished the calculation. The two standards define these factors differently and use different notation, so a rating computed one way is not directly comparable to the other. This is the same discipline that governs other code driven work, from bolted joint calculations to pressure vessel design, where the edition and clause matter as much as the arithmetic.
Where Gear Calculations Go Wrong
Most gear errors are not exotic. They are small, familiar slips that survive review because a spreadsheet full of numbers looks authoritative even when a single input is wrong.
Confusing module and diametral pitch when moving between metric and imperial data, which scales the whole tooth incorrectly.
Leaving the application factor at one for a shock loaded drive, which understates the real stress.
Ignoring the dynamic factor at high pitch line velocity, where mesh effects add meaningful load.
Reading the geometry factor from the wrong chart, or interpolating it carelessly.
Using material allowables from a stale datasheet or the wrong heat treatment condition.
Checking bending strength but skipping the contact check, even though pitting often governs on hardened gears.
Any one of these can pass a casual review and still put a tooth over its limit in service. The broader problem, and the reason so many teams are rethinking how they run numbers, is covered in this look at AI for engineering calculations.
How AI Handles Gear Design Calculations
This is where an AI assistant built for engineering earns its place. Leo is an AI assistant for mechanical engineers, trained on more than one million pages of standards, textbooks, and technical articles. For gear work, that means it can set up the bending and contact checks in the correct form, cite the AGMA or ISO clause behind each factor, and pull material properties with a source you can click and verify rather than a value you have to trust on faith.
The value driver here is technical accuracy. Leo flags when an application or dynamic factor looks too low for the described duty, and it shows its work so a reviewer can follow the reasoning. Leo also connects to an organization's knowledge base and offers integrations with leading PDM and PLM platforms, including SolidWorks PDM, Autodesk Vault, PTC Windchill, Siemens Teamcenter, and Arena PLM, among others. It sits as an intelligence layer on top of those systems rather than replacing them, so it can surface a gear you already designed and qualified before you draw a new one. That matters for two reasons. First, standards get revised, and a calculation that quietly used an old edition can fail an audit even when the arithmetic is perfect, so a tool that names the edition and clause protects the work. Second, most teams carry more usable prior work than any one engineer can remember, and surfacing a proven gear with its calculation history attached turns tribal knowledge into something the whole team can reuse. The work stays protected too, since Leo is SOC-2 certified, is GDPR compliant, and never trains on customer data.
Building Gear Calculations Into Your Workflow
Speed only helps if the result is trustworthy and repeatable. A workflow that treats AI as a fast, well read assistant, not an unchecked oracle, gives you both.
Start from a reviewed calculation template rather than a blank sheet, so the method is consistent across the team.
Let AI draft the sizing and the two stress checks, then confirm each factor against the cited clause.
Keep the citation trail attached to the calculation, so a reviewer can verify the edition and clause quickly.
Search your PDM or PLM for an existing gear that meets the duty before you release a new part number.
Record the design decision and the governing failure mode, so the next engineer inherits the reasoning and not just the result.
One more habit pays off over time. Note the assumptions that were hardest to pin down, such as the duty of the driven machine or the exact heat treatment, because those are the inputs a reviewer will question and the ones a future revision is most likely to change. Done this way, the calculation gets faster and the paper trail gets stronger at the same time, which is exactly the combination that holds up in a design review.
FAQ
Run Gear Calculations You Can Trust
See how Leo checks gear stress against cited standards.
Leo runs your gear bending and contact checks, cites the AGMA and ISO clauses behind every number, and surfaces gears you already qualified so you reuse before you redesign.
Schedule a Demo →
#1 New AI Software Globally - G2 2026
Enterprise-grade security
Trusted by world-class engineering teams
