So

we stopped for the last time at this CT sample.

We also want to provide you with a sample that isable to your smoothe improvement.

The Chevron samples come for beta-Неoable-like semaffles,

And the example I showed you is a relatively spruyter tool.

We don't need a chevron cut

we only need a slight swing cut at the tip

because there is not much ductility.

Good, no C-determination and so on, we looked at it last time.

And then we started with the statistical break-mechanics.

Exactly, statistical break-mechanics.

I'll just open everything again, I think that's it.

The last one we still have to drive.

Exactly

statistical break-mechanics means we have real

spruyter materials

that have corresponding spread in the mechanical properties,

especially in the strength,

because we have statistically randomly distributed defects in the material.

That is our basic assumption.

In a real material, they are not statistically randomly distributed.

A good example is when we have powder metallurgy or ceramic powder

and we compact them, and that's a very large green line,

then the compacting is not evenly distributed and also the error frequency is not evenly distributed.

But as a first assumption, let's say it is evenly distributed.

And what we get is that the middle break-tension,

that the break-tension is a distribution,

and that the middle break-tension is smaller the smaller the sample size.

Another reason why technical ceramic parts

structural ceramic parts

have a smaller size

because the defect probabilities are lower.

This also applies to ductile materials, if they are in the standard range.

For steel materials

for example

there is also a diameter dependence on the strength.

Well, that means the source of error is the material itself.

And if we now take a rough look at how the distribution of the strength values looks

then we have a distribution of the failure probability of a sample around a strength.

Around a basic strength.

That means the probability...

That's actually the line.

Sorry, that's in the wrong order.

Good.

So

if we have a volume

and we have a failure probability per volume that has an error value

or has an error distribution

then the probability is that if we have a total volume

that is composed of small cubes, N times this basic volume,