I was asked via e-mail what the difference between a PHA and a LOPA is.  I have posted my reply below.

PHA and LOPA are not necessarily related, but some organizations have very standardized methods for implementing both techniques in tandem.  PHA is meant to be a generic term for a study of process hazards.  This terms covers a multitude of methods including HAZOP, What-if, and Checklist.  These studies are essentially “structured brainstorming” techniques where some type of mental prompt, such as a HAZOP guideword or a checklist question are presented to a team for discussion in order to help them to come up with scenarios that could potentially result in a process incident.  Each of these scenarios are then analyzed – usually qualitatively – at a coarse level of analysis to determine if the amount of safeguards that are available are appropriate for the level of risk.
LOPA, on the other hand, is a more detailed level of analysis of a known scenario.  Where PHA is primarily utilized to “identify” hazards that are not currently known, LOPA (and other tools such as fault tree analysis and event tree analysis) are employed to understand a known hazard scenario in a greater level of detail.  LOPA is an “order of magnitude” level of analysis the can semi-quantitatively assess risks and associated safeguards with more rigor than PHA, but not as much as full quantitative risk analysis.  This level of analysis is a good follow up for higher level risk situations, and the establishment of risk performance metrics such as SIL.  LOPA is a generic tool that can be implemented in several ways.  A “HAZOP/LOPA” is a form of LOPA whose goal is to determine if a tolerable level of risk is achieved, considering all of the safeguards that are available.  A “LOPA/SIL” is variation of SIL that is geared specifically to select SIL targets for safety instrumented functions.