Environmental Engineering Reference
In-Depth Information
edge. Confidence assessment could be used to
assign a condition of 3 with a confidence of
2to
reflect that the PF's condition could be anywhere
within the 1-5 conditionvalue range. The problem
with assigning such a lowmeasure of confidence is
that the condition index produced using this
method would have a very large uncertainty asso-
ciated with it, and the 3 assigned to condition
would skew the Failure mode likelihood Indices
(FIs) and Confidence Indices (CIs). It was deemed a
better option to grade non-inspectable PFs with a
zero for their condition representing that no con-
dition was assigned for those features. The calcu-
lation of FI andCI would exclude the zero-rated PF.
sets for the development of the measured step
forwards proposed by this project.
Confidence
There is widespread use of confidence/uncertainty
assessment in inspection methodologies, as it pro-
vides feedback about the precision of the inspec-
tion. Confidence, in terms of the method for visual
inspection proposed, refers to the inspector's as-
sessment of the accuracy of the condition value
given to a performance feature. The confidencewill
be affected by:
.
the experience of the inspector on the specific
type of asset;
.
the access that the inspector has to the asset;
.
the prevailing local weather conditions at the
time of inspection;
.
tidal conditions at the time of inspection (where
appropriate).
In the majority of instances the likely uncer-
tainty range should be low and be spread above or
below the assigned condition score. There may be
instances where the uncertainty is high, for ex-
ample when the performance feature is difficult to
inspect due to poor visibility.
A number of methods that could be applied to
assign confidence values to the condition score for
a performance feature were considered. The cho-
sen method involves the inspector assigning a
signed value of confidence to a performance fea-
ture. This represents the likely difference between
the assigned condition and actual condition. A
positive score indicates that the actual value could
lie in the range between the assigned value and the
assigned value plus the confidence value. A neg-
ative score indicates that it lies between the
assigned condition score and the condition score
minus the confidence value. Unsigned values
would indicate that the range of potential values
lies in the range of the assigned condition index
plus or minus the confidence value. For example:
condition
¼
2, confidence
¼þ
1, potential condi-
tion range
¼
2-3.
In those instances where the performance fea-
ture is not inspectable, the confidence score
should not be used to reflect this lack of knowl-
Failure mode likelihood index (FI)
The FI is a score in the same 1-5 range as the
condition scores assigned to each performance
feature. It is related to a specific failure mode for
an asset type and is calculated using the formula
below:
FI
ð
Failure Mode
Þ¼
X
Contribution
ð
i
Þ
*Condition
ð
i
Þ=
i
¼
Performance Feature
X
Contribution
ð
i
Þ
ð
6
:
1
Þ
i
¼
Performance Feature
The nature of the visual inspection process can
be highly uncertain for a variety of reasons such as
poor weather conditions or restrictive access to
assets when inspecting. This will inevitably lead
to instances where certain performance features
cannot be inspectedwith any degree of confidence.
Any such incomplete inspection will therefore
have a set of contributions totalling less than one
for one or more failure modes. In this case the
calculated failure mode indices will reflect this by
the division by total contributions assessed. The
fact that the FI was calculated on a reduced con-
tribution total would be noted and may trigger the
need for a re-inspection of the asset to check the
omitted performance features. This could require
visiting the asset at low tide or some other method
for gaining access to the inaccessible features.
