Composite properties can be fairly accurately measured and tested, just like any other material including plate steel or aluminium.

If a structural failure has occurred, it's related to a design, manufacture or maintenance issue. For example, not enough layers, or a weakness around a joint or fitting.

Metals are uniform and thus predictable. Composites vary cure to cure. We simply don’t have the deep engineering tables that let you know how a material will behave without destructive testing.

I’m by no means a structural engineer, but with bike parts — in steel or aluminum, if you get a tiny 1mm chip in the material, you can more-or-less just treat it as if the steel is 1mm thinner. With engineering tolerances it is irrelevant.

But in a carbon fiber, that 1mm chip could be totally unnoticed for months, while slowly causing an invisible hairline crack, until the entire component suddenly explodes with no warning.

Maybe someone can verify if that is possible at a larger scale, but I feel like it must be to some extent.. especially if there is an invisible manufacturing imperfection due to the higher precision required in carbon fiber

The BOAC Comet failures (cracks propagating from squared-off window corners) is a notorious example, with multiple total hull losses in flight.

I've experienced somewhat less significant failures of bicycle components (though they could easily have proved far worse). It's a good idea to replace components after a wreck or even hard-cornering of a crankset.

https://www.youtube.com/watch?v=-DjnG74DDno

The cracks actually started on the top of the plane, at cutouts for radio antennas. The windows always had rounded corners (although they made them even rounder later, which may have led to the confusion.)

Thanks!

I'm sure the story around composites will improve in due time.

That's not true at all: metals have a grain structure, and how those grains are oriented can change the metal's properties in different directions.

>Composites vary cure to cure.

Metal castings can vary from cast to cast.

These folks, among others, make a lot of them. You are also incorrect about forging being used to introduce uniformity into the grain, it is typically (though not always, as some forging operations may just be for establishing shape) to stretch the grains such that they are oriented along the load path, resulting in higher allowable stresses. Look up highly loaded aircraft parts like landing gear, they are typically forged to establish the grain direction and then machined to final dimension. Nothing stops you from heat treating metal after it has been cast, either. Some heat treatments are also expressly to produce nonuniform conditions in the metal, case hardening for example.

But now that i think about it, I have no clue about performance or even feasibility of testing carbon composites. I'll have to ask around, but I'm pretty sure we have clients that use our hardware for Carbon! :)

But sometimes they are! We have a few scanners specifically for pipelines and even robots for wind blades, yet even then the probe usually uses the same detection technology (ultrasounds, eddy currents, etc). The magic happens on the signal processing side when it comes to the difference between the scanners, and the trend is towards multipurpose all in ones. Our most recent industrial scanner has an FPGA that is explicitly there to allow for extensibility so that it fits most applications. FPGAs have been used quite often in the field but usually merely to implement some specific signal processing, not to be extensible!

I'm rambling here but my point is that some devices can be used for pretty much any application (the Omniscan for example), process specific devices are more for large scale industrial workflows! The probes can be the only thing you need to change to use the same scanner to do stuff as varied as hydrogen poisoning analysis on steel to composite flaw detection on windblades!

Here is an example, again with the Omniscan :

https://www.olympus-ims.com/en/scanners/aerospace-wind-blade...

FPGAs are super powerful for applications like these, I see them used in plenty of sensors where flexibility and being future proof is a must so that you don't need to redesign the entire sensor for every variation on the theme.

Carbon fibre composites are also significantly lighter and more rigid than equivalent steel structures. As with all engineering it is about choosing materials with properties that suit your application.