`Pratt & Whitney / General Electric jet engine disc HIP vessel failure`

Forensic conclusion. Tom analyzes the 200-ton vessel failure as a three-cause convergence — transient thermal stresses (the operator hadn't waited for thermal equilibration; the design hadn't been modeled in finite element because 1982-era computing couldn't), susceptible microstructure (Japan Steelworks delivered material with toughness below specification, and the un-heated corner regions retained the as-rolled properties), and hydrogen from a corrosion-pit-initiated crack on the water-cooled jacket (the inspection company hadn't been checking the outside often enough; the water treatment company was using out-of-date molybdate chemistry). Critical flaw size: half an inch. Three liable parties: Japan Steelworks, operator, water-treatment / inspection company.

And we knew what the toughness was supposed to be. In fact there were three problems. One is they had residual stresses — actually take that back, the thing had been stress relieved. They had thermal stresses from startup. They had just put a new load in two hours before, and the whole thing was coming up to temperature. I remember coming back after being there the first day and going into Joe Dose's office, and I said, "Joe, beware the thermal stresses." That was my summary, because I had done the calculation. In two hours you would only diffuse heat about halfway in.