The SCARABEE-N in-pile tests were performed between 1983 and 1990. Their main objective was to study the consequences of a hypothetical total instantaneous blockage (TIB) at the entrance of a liquid-metal reactor subassembly at full power. After 14 tests, mainly with fresh fuel, this accident scenario may be fairly well described. It has been shown that no violent, energetic fuel/coolant interactions take place, that almost no fuel is ejected out of the fissile zone, and that the melt penetration into the neighboring subassemblies proceeds rapidly. Progress has been made on understanding the behavior of mixed boiling pools, and hexcan melt-through thresholds can be established.

Codes have been developed (PHYSURA-GRAPPE and SURFASS), and SCARABEE-N tests have been used to help in the evaluation of others (SIMMER II and SABRE). Interesting information for other accident situations has been made available, especially for the transition phase.

Concerning the specific case of the TIB accident, it has been shown that the evolution of the molten zone can be represented by calculations in which energy equations play the essential role. For a Superphénix-type subassembly, PHYSURA-GRAPPE and analytic calculations show that after ~30 s after the beginning of the blockage, there exists a serious risk of propagation beyond a 7-subassembly pool. This can be prevented only by an adequate detection system leading to scram before this time.

The SCARABEE-N program helped to solve the problem for which it was designed; moreover, it yielded scientific expertise for other areas and was a precious training ground for many safety scientists and engineers. It also showed the limits of so-called mechanistic models for very complex problems and the sustained usefulness of simple engineering reasoning.