In this paper, we focus on knowledge management for critical situation of Fukushima Atomic Ac- cident. For the purpose, we try to clarify the importance of boundary condition management against critical situation. In this accident, the word "Unanticipated" is used so frequently for the excuse. On the other hand, they said that the nuclear reactor is free from danger under the assumed boundary condition.

 Risk management should start from requiring a boundary condition of an "anticipated situation". An evaluation of "anticipated situation" is also required. Many dialogues for Fulushima accident among specialists remain opaque about a boundary condition for "anticipated situation", that are indispensable for inducing a logical result.

 In this paper, we try to clarify the failure of sharing "anticipated situation" among stakeholders. We also focus on establishing a boundary conditions for "anticipated situation" that becomes presupposed grounds for risk estimation. Thus, we introduce the view points of knowledge management and distin- guish two independent issues about Fukushima accident as follows.

 The first one is the evaluation of social risk under the boundary condition about the situation of meltdown. The second one is to estimate "anticipated situation" that causes the situation of meltdown. In addition, we try to clarify the problem of knowledge management where the social risk management under the meltdown is not executed at all. It is also clarified that the second issue is not treated as pure scientific and engineering matter but treated as a matter of cost management. It is a typical organiza- tional failure under the contamination by the viewpoint by cost.

 We investigate these new aspects of the social and organizational crisis management from knowl- edge management point of view.