Waloddi Weibull

1887-1979

The Swedish engineer whose flexible probability distribution describes when things break and wear out, the statistical backbone of reliability testing, failure prediction, and the bathtub curve for electronics.

Ernst Hjalmar Waloddi Weibull was born in 1887 in Vittskövle, Sweden. His was not a conventional academic path. He joined the Swedish Coast Guard in 1905 as a midshipman and rose through the ranks to major, taking courses at the Royal Institute of Technology along the way, graduating in 1924, and earning his doctorate from Uppsala in 1932. He spent much of his career as a consulting engineer to Swedish and German industry, a practical man whose mathematics was always aimed at real failures of real materials.

In 1939 he published his famous paper on what is now called the Weibull distribution, growing out of his statistical theory of the strength of materials. The distribution's power is its flexibility. With just a couple of parameters it can describe many different failure patterns: a shape that captures early infant-mortality failures, one that captures purely random failures, and one that captures wear-out as parts age. The same curve family that fits the fatigue cracking of a steel beam also fits the lifetime of a capacitor or a bearing.

Weibull kept publishing prolifically on strength of materials, fatigue, and rupture, and presented his distribution to American engineers with seven worked case studies in 1951. Much of his later work went into reports for the US Air Force on what is now simply called Weibull analysis. The American Society of Mechanical Engineers gave him their gold medal in 1972, and in 1978 the King of Sweden personally presented him the Great Gold Medal of the Royal Swedish Academy of Engineering Sciences.

In electronics, Weibull's distribution is the mathematics behind reliability engineering. The familiar bathtub curve, high early-life failures, then a long flat random-failure period, then rising wear-out failures, is exactly the kind of behavior a Weibull fit describes by tuning its shape parameter. When you read a component's rated lifetime, design a burn-in screen to weed out infant mortality, or estimate how long a fleet of devices will survive in the field, you are leaning on Weibull statistics. He gave hardware engineers a rigorous way to answer the unavoidable question: when will this break?

Fun facts

Weibull was a career coast guard officer who rose to the rank of major. He took engineering courses and did mathematics on the side, only becoming a full professor and doctorate-holder well into adulthood.
Before his famous distribution, his first paper grew out of ocean expeditions: he developed a technique of setting off explosive charges to determine the type and thickness of seafloor sediments, a method still used in offshore oil exploration today.
The distribution's flexibility is the whole point. By changing a single shape parameter it can describe early failures, constant random failures, or wear-out, which is exactly why it maps so cleanly onto the bathtub curve of electronic component life.