Conduct an Effective Incident Investigation

Information that can reveal the root cause of an incident resides in many places ― within the plant or process unit, and in control rooms, offices and witnesses' minds. Here's how to find the data and conduct effective witness interviews.

Readers who were listening to the radio or watching television on the morning of February 1, 2003, will remember the loss of the Space Shuttle Columbia over Texas. Within an hour of losing contact with the Columbia, NASA’s Mission Control declared a “contingency” to ensure that all mission data were preserved. All flight controllers had to verify that their logs were up-to-date and institute a hands-off policy with regard to switches, push-buttons, controllers, knobs and the like; all computer data were impounded. This was the start of NASA’s investigation procedure.

Process plants need to develop similar procedures to be carried out following an accident. To successfully determine the root cause of an incident, it is essential to do a throrough job of locating and preserving all the available data and information. It is better to have too much data and too many interviews than to get to the end of the investigation and find that the one key piece of information needed to establish conclusively the incident cause is missing. This is why NASA declared a contingency in the Columbia incident that started with the preservation and collection of all data, not only at Mission Control in Houston, but also at the Kennedy Space Center and shuttle contractor facilities.

This article suggests actions to take following a large incident to preserve data and witness information. These same techniques can be scaled down for smaller events. Smaller incidents will involve smaller investigation teams and smaller areas of impact, but the same steps must be followed to preserve and acquire the data necessary for a thorough investigation.