Professional Documents
Culture Documents
14 Quantifying Human Reliability in Risk Assessments PDF
14 Quantifying Human Reliability in Risk Assessments PDF
F
ollowing the Buncefield accident in Associates to develop guidance in this
2005, operators of bulk petroleum area. The aim is to reduce instances of The respective merits of HRA techniques
storage facilities in the UK were poorly conceived or executed analyses. are not addressed in the new guidance,
requested to provide greater assurance The guidance provides an overview of since this information, and more
of their overfill protection systems by important practical considerations, detailed discussion of the concept of
risk assessing them using the layers of worked examples and supporting HRA, are available elsewhere.4,5,6
protection analysis (LOPA) technique. A checklists, to assist with commissioning Attempts to quantify the probability
subsequent review of LOPAs1 indicated and reviewing HRAs.2 of human failures have a long history.
a recurring problem with the use of Early efforts treated people like any
human error probabilities (HEPs) other component in a reliability assess-
HRA techniques
without an adequate consideration of ment (eg what is the probability of an
the conditions that influence these HRA techniques are designed to support operator failing to respond to an
probabilities in the scenario under con- the assessment and minimisation of alarm?). Because these assessments
sideration. It is obvious that the error risks associated with human failures. required probabilities as inputs, there
probability will be affected by a They have both qualitative (eg task was a requirement to develop HEP
number of factors (eg time pressure, analysis, failure identification) and databases. However, very few industries
quality of procedures, equipment quantitative (eg human error estima- were prepared to invest in the effort
design and operating culture) that are tion) components. The guidance focuses required to collect the data to develop
likely to be specific to the situation primarily on quantification, but illus- HEPs, so this led to the widespread use
being evaluated. Using an HEP from a trates the importance of the associated of generic data contained in tools such
database or table without considering qualitative analyses that can have a sig- as THERP (technique for human error
the task context can therefore lead to nificant impact on the numerical results. rate prediction). In fact, the data con-
inaccurate results in applications such Further EI guidance on qualitative tained in THERP and other popular
as quantified risk assessment (QRA), analysis is also available.3 There are a quantification techniques such as
LOPA and safety integrity level (SIL) large number of available HRA tech- HEART (human error assessment and
determination studies. niques that address quantification – one reduction technique) are actually
Human reliability analysis (HRA) tech-
niques are available to support the
development of HEPs and, in some cases,
their integration into QRAs. However,
without a basic understanding of human
factors issues, and the strengths, weak-
nesses and limitations of the
techniques, their use can lead to wildly
pessimistic or optimistic results.
Using funding from its Technical
Partners and other sponsors, the Energy
Institute’s (EI) SILs/LOPAs Working
Group commissioned Human Reliability
Table 1: Generic HRA process Figure 1: Examples of the potential impact of human failures on an event sequence