What do STAMP-based analysts expect from safety investigations?

摘要

STAMP postulates that accidents result from inadequate enforcement of system safety constraints in design, development and operations. It claims that Causal Analysis based on STAMP (CAST) is needed to ensure understanding of the entire accident process and all systemic causal factors involved. STAMP does not specify an accident investigation process. CAST uses information about accidental occurrences in its analyses application, including information, disseminated through investigation reports. When CAST analysts consider an accidental occurrence report as a candidate for analysis, they should be aware of the variations among investigation reports and how they might affect their analysis. They are unlikely to recognize them unless they have conducted and reported actual investigations using more than one method. This paper examines the information CAST requires and its sources, what that various safety investigations now provide, the variances in that information, other sources of input data CAST needs, and describes issues found. Changing investigation reports and methods however is necessary, but insufficient. It requires a non-Tayloristic view on human error and mobilizes other data sources of a systemic nature such as design history, network typologies and business models.