Villains, victims, and heroes: Accounting for the roles human activity plays in LOPA scenarios

• There are three roles human activity plays in LOPA: initiating cause (“villain”), receptor (“victim”), and IPL (“hero”).
• Human error results from lapses, mistakes, or violations, but LOPA is only suited to random errors—lapses and some mistakes.
• Reducing cause rate requires changing task type, hence probability of error, or reducing frequency of performing the task.
• Reducing occupancy reduces receptor risk, either by reducing number of personnel, or by reducing exposure time.
• While heroics should be discouraged, procedures or planned operator response to emergency conditions may each serve as IPLs.

When a team is analyzing a LOPA scenario, the team needs to consider all three roles played by human interaction in the scenario: that of cause, as a result of human error; that of receptor, both in terms of safety impacts (inside the fence line) and community impacts (outside the fence line); and that of independent layer of protection (IPL), considering both administrative controls and human responses. Frequently, the nature of these three roles are inter-related, and setting guidance that is internally consistent is important to using LOPA to assess risk rather than as a means to game the analyses to simply achieve a wished-for result.A number of criteria have been proposed to quantify human involvement, typically as cause, as receptor, or as IPL. Establishing a framework to look at all three in a unified way is more likely to result in analyses that are consistent from scenario to scenario.This paper describes such a framework and presents it in a way that allows organizations to review their own criteria for quantifying human involvement in LOPA. It also examines some of the published LOPA criteria for human involvement and looks at them in terms of consistency of approach between evaluation of cause, receptor, and IPL. Finally the paper makes suggestions to use in calibrating LOPA methodologies to achieve consistent and believable results in terms of human interaction within and between scenarios that have worked for other organizations.