Assessment of risk mitigation measures effectiveness: application to natural risks in mountains
Abstract
Mitigation actions against natural risks in mountains are based
both on non-structural measures such as land-use control and
on structural measures such as civil engineering protection
works. To prevent or, at least, limit risks, protection works play
an essential part in reducing both the causes and effects of
phenomena. The analysis of their effectiveness is an essential
stage of the risk management process. Decision support tools
are needed to take into account their structural state and
functional effects on phenomena (stopping, braking, guiding,
diverting etc.) but also their economic efficiency, comparing
their construction and maintenance costs to direct and indirect
economic losses reduction. The concepts of dependability
analysis, functional analysis and failure modes and effects
analysis (FMEA), already used in civil engineering especially for
hydraulic dams, are extended to assess and evaluate the
structural and functional effectiveness of protection works. The
methodology elicits the expert reasoning process and evaluates
the danger level of the protection works. It uses of both
dependability analysis, an emerging technique for natural
hazards domain and uncertainty theories such as fuzzy logics,
possibility theory to take into account the imperfection of
information. Fuzzy intervals and possibility distributions are
used to assess the indicators that describe the possible failures
and the performance level of disposals. Indicators are defined
as information formalised so as to make its use repeatable and
reproducible. They provide indications on a function
performance. Their implementation is based first on a system
description and on the proposal of indicators related to
structural and functional failure modes. The concept of risk, the
conditions and limits of cost-benefit analysis are discussed with
regard to imperfect of knowledge about frequency, physical
effects on phenomena, vulnerability curves and economic
valuations of losses.