Page 36 - Root Cause Failure Analysis

Root Cause Failure Analysis Methodology 27

Analyze Sequence of Events
Performing a sequence-of-events analysis and graphically plotting the actions leading
up to and following an event, accident, or failure helps visualize what happened. It is
important to use such a diagram from the start of an investigation. This not only helps
with organizing the information but also in identifying missing or conflicting data,
showing the relationship between events and the incident, and highlighting potential
causes of the incident.

DESIGN REVIEW
It is essential to clearly understand the design parameters and specifications of the
systems associated with an event or equipment failure. Unless the investigator under-
stands precisely what the machine or production system was designed to do and its
inherent limitations, it is impossible to isolate the root cause of a problem or event.
The data obtained from a design review provide a baseline or reference. which is
needed to fully investigate and resolve plant problems.

The objective of the design review is to establish the specific operating characteristics
of the machine or production system involved in the incident. The evaluation should
clearly define the specific function or functions that each machine and system was
designed to perform. In addition, the review should establish the acceptable operating
envelope, or range, that the machine or system can tolerate without a measurable
deviation from design performance.

The logic used for a comprehensive review is similar to that of a failure modes and
effects analysis and a fault-tree analysis in that it is intended to identify the contribut-
ing variables. Unlike these other techniques, which use complex probability tables
and break down each machine to the component level, RCFA takes a more practical
approach. The technique is based on readily available, application-specific data to
determine the variables that may cause or contribute to an incident.

While the level of detail required for a design review varies depending on the type of
event, this step cannot be omitted from any investigation. In some instances, the pro-
cess may be limited to a cursory review of the vendor’s operating and maintenance
(O&M) manual and performance specifications. In others, a full evaluation that
includes all procurement, design, and operations data may be required.

Minimum Design Data
In many cases, the information required can be obtained from four sources: equip-
ment nameplates, procurement specifications, vendor specifications. and the O&M
manuals provided by the vendors.

If the investigator has a reasonable understanding of machine dynamics, a thorough
design review for relatively simple production systems (e.g., pump transfer system)