ISO 6336-6:2019: Calculating Gear Service Life Under Variable Load Conditions

By Johnny Liu CEO at Dowway Vehicle Published on June 23, 2026 in Shanghai, China

Gears in heavy machinery or vehicle transmissions do not work under constant, uniform torque. Instead, they face sudden torque spikes, start-stop cycles, and unpredictable shifts. If you design a gearbox using only static peak-load calculations, you get one of two things: a system that breaks too soon, or a bulky, heavy component that costs far too much.

To solve this, engineers use ISO 6336-6:2019 (full name: Calculation of load capacity of spur and helical gears — Part 6). This international standard sets up a reliable mathematical framework to calculate the fatigue capacity of gears under varying loads.

At Dowway Vehicle, we build powertrain systems. We use this standard to make sure our gears survive real-world driving conditions without adding unnecessary weight.

The Core Goal and Failure Modes

The main goal of ISO 6336-6:2019 is to find the exact safety margins when load levels and frequencies shift over time.

The standard lets you do two things:

1. Estimate how many hours or cycles a gear will last under a specific load history.
2. Adjust your design to reach a target life with the right safety margins.

This calculation protects gears against the two most common fatigue failures:

• Tooth Flank Contact Fatigue (Pitting): Surface cracks and small craters caused by repetitive pressing stress on the tooth surface.
• Tooth Root Bending Fatigue (Tooth Breakage): Crack formation and eventual snapping at the root of the tooth where tensile stresses peak.

The Core Math: Palmgren-Miner Rule

We calculate cumulative fatigue life using the Palmgren-Miner linear cumulative damage rule.

The concept is straightforward: every single rotation of a gear under load uses up a tiny fraction of its life. This damage accumulates over time. When the total sum of this damage hits 100%, the gear fails.

The Formula

$$U = \sum \frac{n_i}{N_i}$$

• $U$ is the Total Cumulative Damage Index.
• $n_i$ is the actual number of load cycles the gear experiences at a specific stress interval $i$.
• $N_i$ is the theoretical fatigue limit (maximum allowable cycles) at that same stress level $i$, found using the material’s S-N curve (Wöhler curve).

The Boundary

• If $U < 1.0$, the gear should survive its intended design life.
• If $U \ge 1.0$, the gear is highly likely to break or suffer pitting.

The 5-Step Workflow

Calculating variable-load gear life requires a structured, step-by-step process. Here is how we run this calculations.

[Step 1: Load Spectrum] -> [Step 2: Operational Stress] -> [Step 3: S-N Curve Calibration] -> [Step 4: Damage Summation] -> [Step 5: Safety Factor Iteration]

Step 1: Build the Load and Stress Spectrum

First, you must turn continuous, messy real-world torque data into a clean matrix of load levels (bins) and cycle counts ($n_i$).

You can build this spectrum using three main methods:

1. Experimental Measurement: You mount sensors (like strain gauges and torque telemetry tools) directly on physical gearboxes operating in real-world fields. This is the most accurate way.
2. Empirical Analogy: If you do not have a prototype yet, you find data from a similar machine running in similar conditions and scale it to fit your design.
3. Dynamic Simulation: You run mathematical models to calculate internal forces based on external forces. Keep in mind that simulated models can miss transient force peaks. You should always validate these simulation results with physical testing later on.

Step 2: Convert Load to Stress ($ISO\ 6336-2$ & $ISO\ 6336-3$)

Once you have your binned torque spectrum, you must calculate the physical stresses for each individual load bin $i$:

• Contact Stress ($\sigma_{H,i}$): Use ISO 6336-2 to calculate the surface contact stress.
• Bending Stress ($\sigma_{F,i}$): Use ISO 6336-3 to calculate the tensile stress at the tooth root fillet.

Step 3: Calibrate Material Fatigue Strength

Next, look up the raw material limits from base S-N curves. You must adjust these values to match your specific gear setup by applying correction factors. These factors account for:

• Lubricant performance and temperature.
• Surface finish and roughness.
• Size and scale of the gear.
• Operating speeds.
• Reliability targets.

This step gives you the actual, corrected S-N curves for both pitting and bending.

Step 4: Sum the Cumulative Damage

Now, take the actual cycles ($n_i$) from Step 1 and the allowable cycles ($N_i$) from Step 3 to run the summation:$$U_H = \sum \frac{n_i}{N_{H,i}} \quad \text{(For Flank Pitting)}$$$$U_F = \sum \frac{n_i}{N_{F,i}} \quad \text{(For Root Breakage)}$$

Step 5: Run the Safety Factor Iteration Loop

Finally, compare your calculated total damage ($U$) with allowable limits.

If the damage is too high or too low, you adjust parameters like gear size, width, module, or material. You then run the math again. You repeat this loop until you find the exact safety factors ($S_H$ for contact and $S_F$ for bending) that meet your exact life requirements.

Hands-On Takeaways

The accuracy of your final life calculation depends entirely on the quality of your load spectrum from Step 1. No amount of advanced math in the later steps can fix bad load data. If you want reliable gearboxes, spend the time and budget required to get real field measurements or validated dynamic models.

Frequently Asked Questions

Can I calculate variable-load gear life using only ISO 6336-2 and ISO 6336-3?

No, you cannot. ISO 6336-2 and ISO 6336-3 calculate stresses under constant loads only. To evaluate fluctuating, real-world loads over time, you must feed those calculations into the cumulative damage framework of ISO 6336-6:2019.

What should I do if the total damage index (U) is larger than 1.0?

You must redesign the gear to reduce stresses or upgrade the material. When $U \ge 1.0$, the gear is expected to fail. You should return to Step 5 and increase the gear module, widen the face, or use higher-strength case-hardened steel.

How do I build a load spectrum if I do not have a physical prototype?

Combine dynamic simulations with empirical data from older designs. Use multi-body simulation software to model your drive cycles, then adjust the data based on historical logs from similar vehicles. Once you build a physical prototype, run telemetry tests to verify and correct your virtual model.