Jun 25, 2026 · 18 min · 10 segments
***Discover how accurate gear lifetime modeling predicts strength and durability, helping prevent failures while accelerating development and reducing testing time and costs.*** To be successful in…
Which brings us to the topic of today's deep dive.
We're looking at a fascinating white paper from Invalier.
The title is Predicting Fatigue Performance of Fiber-Reinforced Plastic Gears.
And the reason we picked this specific paper is that it tries to do something that historically has been, well, considered almost impossible or at least impossibly expensive to figure out.
Their mission is to stop the guesswork.
They want to move away from the build it, break it, fix it model of engineering and move toward a purely predictive model.
Basically, they claim they can predict the lifespan of a complex plastic gear using just a standard material test and some very clever software.
And we're going to tear apart exactly how they did it.
Because if this works, if this is real, it changes the workflow for basically every mechanical designer out there.
It really, really does.
Because look at the current state of the industry.
If you want to design a plastic gear transmission today, what do you do? You design the part.
You pay for an expensive metal mold, which costs a fortune, takes weeks to machine.
You injection mold the parts, and then you stick them in a test rig and run them for weeks until they snap.
And if they snap too early.
Right.
Or if they strip a tooth.
Then you go back to the drawing board, you tweak the geometry, you modify that expensive mold, and you test again.
It's an iterative loop that just burns cash and slows down your time to market.
I've heard this called dedicated application testing, which sounds very professional, but effectively it's just trial and error with a budget.
Precisely.
And the reason we have to do this trial and error is because we haven't had a reliable way to simulate the failure.
You can run a stress simulation on a computer, sure, but for fiber reinforced plastics, those simulations are often dead wrong.
Okay, let's dig into why they're wrong.
Because this is the core of the problem the paper addresses.
Which brings us to the topic of today's deep dive.
We're looking at a fascinating white paper from Invalier.
The title is Predicting Fatigue Performance of Fiber-Reinforced Plastic Gears.
And the reason we picked this specific paper is that it tries to do something that historically has been, well, considered almost impossible or at least impossibly expensive to figure out.
Their mission is to stop the guesswork.
They want to move away from the build it, break it, fix it model of engineering and move toward a purely predictive model.
Basically, they claim they can predict the lifespan of a complex plastic gear using just a standard material test and some very clever software.
And we're going to tear apart exactly how they did it.
Because if this works, if this is real, it changes the workflow for basically every mechanical designer out there.
It really, really does.
Because look at the current state of the industry.
If you want to design a plastic gear transmission today, what do you do? You design the part.
You pay for an expensive metal mold, which costs a fortune, takes weeks to machine.
You injection mold the parts, and then you stick them in a test rig and run them for weeks until they snap.
And if they snap too early.
Right.
Or if they strip a tooth.
Then you go back to the drawing board, you tweak the geometry, you modify that expensive mold, and you test again.
It's an iterative loop that just burns cash and slows down your time to market.
I've heard this called dedicated application testing, which sounds very professional, but effectively it's just trial and error with a budget.
Precisely.
And the reason we have to do this trial and error is because we haven't had a reliable way to simulate the failure.
You can run a stress simulation on a computer, sure, but for fiber reinforced plastics, those simulations are often dead wrong.
Okay, let's dig into why they're wrong.
Because this is the core of the problem the paper addresses.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.