Weibull statistics

Ask ten identical ceramic rods when they snap and you will get ten quite different answers, spread over a wide range. Metals barely scatter, but ceramics scatter enormously, and Weibull statistics is the maths engineers use to describe and design around that scatter. The core idea is the weakest-link picture: a ceramic breaks from its single worst flaw, like a chain that fails at its weakest link, and since no two samples have the same worst flaw, no two break at the same load.

Because failure hunts for the biggest flaw, a bigger component is more likely to contain a dangerous one, so, remarkably, larger ceramic parts are weaker on average. Weibull captures this with a survival probability P = exp( -(sigma/sigma0)^m ), the chance a part survives stress sigma, where sigma0 is a characteristic strength and m is the Weibull modulus. The modulus m is the key number: it measures how tight the scatter is. A high m (say 20 or more, as in a good structural ceramic) means consistent, predictable strength; a low m (5 to 10) means wild scatter and a fat tail of weak parts. To use it you test many specimens, rank their breaking stresses, and fit the line to find m and sigma0.

This is honesty about materials made quantitative. You cannot design a ceramic to its average strength, because half of all parts are weaker than average and would fail. Instead you design to a low survival percentile, for example the stress at which 99.9 percent of parts survive, which for a low-m ceramic can be far below the average. Weibull is why ceramic makers work so hard to shrink and control flaws (fine powders, clean processing, polished surfaces), because raising m matters as much as raising the average strength.