4 investigations
Nuclear Safety
Four investigations look at different nuclear-safety problems: a control rod moved far beyond its maintenance limit, a relief valve stayed open while the panel suggested otherwise, a reactor fire released iodine, and a lost bomb had to be recovered from deep water.
Control, containment, visibility, and recovery
At Windscale, a late-added chimney filter reduced particulate release but could not prevent iodine from entering the environment. At SL-1, a maintenance procedure assumed a central control rod would move only a few inches; investigators found it had been withdrawn far farther. Three Mile Island added an interface problem: the control room showed that a valve had been commanded closed, not that it was physically closed. Palomares was not a reactor accident at all, but it turned nuclear safety into a recovery problem spread across land and deep water.
The safety problem is different in each case: Windscale is about containment and release pathways; SL-1 about mechanical control, procedure, and reactivity; Three Mile Island about instrumentation and the operator’s picture of the plant; Palomares about accounting for and recovering dispersed nuclear material.
Key cases
The Light Said Closed
The control-room indication showed the command, not the valve’s physical state.
The Rod Was Supposed to Move Four Inches
Procedure, mechanical movement and catastrophic reactivity.
The Filter Was at the Top
The chimney filter trapped much of the particulate while iodine still escaped.
Investigations
They Found the Bomb Twice
After the 1966 Palomares nuclear accident, Alvin found the missing bomb 2,550 feet underwater—then it vanished again during recovery. The Navy had to find…
The Filter Was at the Top
Windscale’s chimney filter was a late addition 125 metres above the reactor. In 1957 it caught most radioactive particulate—but iodine vapour still escaped into…
The Light Said Closed
At Three Mile Island in 1979, a relief valve stayed open while control-room instruments indicated it was closed. Operators did not recognize the coolant…
The Rod Was Supposed to Move Four Inches
At Idaho’s SL-1 reactor in 1961, a maintenance procedure allowed the central control rod to move only four inches. Investigators later found it about…
Limits of the comparison
These four events involve different technologies and regulatory settings; one is a weapons-recovery operation rather than a reactor failure. The dossier does not collapse them into a single causal model. It uses them to compare visibility, containment, control and recovery—four different places where nuclear safety can become fragile.