loss (or cost) as the two primary determinants. Hazard levels are readily
compared or ranked in relation to either risk or loss provided that either is
defined as the decisive parameter. In UK practice a simple hazard categori-
zation is embedded in the procedure for the determination of the design
flood and spillway size. (See Section 4.2 and Table 4.1).
Several options have been developed for reservoir hazard analysis,
ranging from simplistic expressions involving the reservoir capacity and
height of the dam to sophisticated techniques of probabilistic risk assessment
(PRA). The cumbersome methodologies for PRA were derived from those
developed for high-risk or sensitive activities, e.g. in the nuclear and chemical
industries. Options in hazard analysis are reviewed in Moffat (1988, 1995),
with proposals for a quantitative reservoir hazard rating (RHR). The RHR
was conceived as a semi-rigorous but relatively simplistic expression,
intended primarily for the ranking of hazard levels and priorities within a
population of dams. It was designed principally for general application in
first-level screening and for the identification of high-hazard reservoirs.
Alternative methodologies have been comprehensively discussed
and reviewed (Binnie and Partners, 1992). The application of relatively
simplistic rapid assessment techniques to hazard ranking a population of
dams is discussed in Thompson and Clark (1994).
A further approach to risk assessment in the context of surveillance
and reservoir management, developed for use in the UK, is presented in
Hugheset al.(2000). This approach is based on the semi-empirical but sys-
tematic and logical concept of ‘Failure Modes, Effects and Criticality
Analysis’ (FMECA). The procedure offers a balance between the extremes
of relying solely upon qualitative and subjective engineering judgement on
the one hand and the rigour of an expensive and difficult probabilistic risk
analysis based on limited statistical evidence on the other.
The FMECA-based risk assessment procedure has three stages:
Stage 1: potential failure impact assessment;
Stage 2: from Stage 1, selection of one of three options, i.e. detailed
FMECA, standard FMECA, or no further assessment necessary;
Stage 3: the FMECA assessment, to the degree selected in Stage 2.
Stage 1 requires application of a rapid method of determining potential
flood water levels downstream of the dam, together with a semi-
quantitative assessment of flood impact in relation to lives, infrastructure
and property.
In Stage 2, the magnitude of the potential impact as assessed in Stage
1 is used as the basis of selection. Where the assessed impact score is con-
sidered to be minimal, no further work is necessary. For medium and high
impacts, FMECA assessment is required with, in the case of high impact
scores, the consideration of alternative failure scenarios and consequences.
Stage 3, the FMECA assessment, provides for review of the several
310 DAM SAFETY: INSTRUMENTATION AND SURVEILLANCE
