AM-14-31 Semantic Procedure Analysis

Description:

Generally, information for normal and emergency operating procedures is organized by: (1) unit and (2) purpose. Often there are distinctions for board versus field actions in the procedures. This can lead to situations where an operator with responsibility for multiple units is expected to simultaneously use multiple procedures. In addition, the same actions are often specified in multiple procedures, because one procedure is a subset of another, or the procedures converge on a common set of tasks. Industry is increasing the emphasis on procedures and expanding the depth and breadth of coverage, resulting in a large volume of information that can be unwieldy and hard to maintain. The Center for Operator Performance commissioned Penn State University to examine new ways to improve operating procedures. This research developed an automated Semantic Procedure Analyzer (SPA) to chunk operating procedures into logical “chunks of procedures” (modules). The modules can be combined in any sequence to match user needs for operating procedures, point of action information, or training exercises. Since the modules are used as building blocks, any required updates only need to be made once, and all associated procedures will automatically be compliant. Methods to extract and codify tacit information on procedure use have been used to capture and add tacit information to a module, making it a permanent part of the procedures. Application of SPA to procedures at one plant revealed that only 10% of the content was unique. The SPA modules were loaded into an Access database that is used to generate and maintain procedures. Maintenance of the procedures is only 10% of what it was; and consistency in procedure terminology and depth of content has been improved.