“…The analysis development and verification for the required degree of predictability is difficult for a number of reasons: (1) The decompression event is a transient process in which changes in typical phenomena are occurring differently in different parts of the system, thus requiring a "distributed system" as opposed to a "lumped" analysis approach (2) The large size and cost of nuclear power plants prohibit testing of accident response at full scale, thus necessitating incorporating within the analysis the capability of accounting for scaling and of accommodating those f a c t o r s that cannot be scaled…”