1823
Slope stability with partial safety factor method
Stabilité des pentes à l’aide de la méthode de sécurité partielle
Länsivaara T., Poutanen T.
Tampere University of Technology, Tampere , Finland
ABSTRACT: In most European countries Eurocode 7 (EN 1997) has been taken into use in geotechnical design. Although there are
several design approaches in EN 1997, slope stability is commonly addressed by applying partial safety factors to material properties
and variable loads. The outcome of such procedure has been evaluated with emphasis on uncertainty and consequences of failure. An
attempt to improve the EN-1997 partial safety method by the introduction of consequence classes into stability analysis is then
presented. Therein the partial safety factors for soil strength will be connected into consequence classes, i.e. the consequences of
possible failure. Eurocode 1990 defines the minimum reliability index values for different Reliability Classes, which are then
associated with the consequences classes. The result by the alternative method is then compared with results from the original
approach.
RÉSUMÉ : Dans la plupart des pays européens, l’ Eurocode 7 (EN 1997) a été mis en service dans la conception géotechnique. Bien
qu'il existe plusieurs approches de conception de la norme EN 1997, la stabilité des pentes est généralement abordée en appliquant des
facteurs de sécurité partiels pour les propriétés des matériaux et des charges variables. Le résultat d'une telle procédure a été évaluée
en mettant l'accent sur l'incertitude et les conséquences de l'échec. Une tentative pour améliorer la méthode de sécurité EN-1997
partielle par l'introduction de classes de conséquence dans l'analyse de la stabilité est ensuite présenté. D’apès celle-ci, les facteurs
partiels de sécurité pour la résistance du sol seront connecté en classes de conséquence, c'est à dire les conséquences d'un échec
possible. L’Eurocode 1990 définit les valeurs minimales indices de fiabilité pour les différentes classes de fiabilité, qui sont ensuite
associées à des classes de conséquences. Le résultat par la méthode alternative est finalement comparé avec les résultats de l'approche
originale.
KEYWORDS: slope stability, partial safety method, reliability, consequence class
1 INTRODUCTION
In most European countries, Eurocodes have been taken into use
as the main design standard. Eurocodes are based on the
principles of limit states design in conjunction with the partial
factor of safety method. EN 1990 describes the principles and
requirements for safety, serviceability and durability of
structures for all materials, whilst EN 1997 sets the rules for
geotechnical design. For ultimate limit states design, Eurocode
requires the verification of various ultimate limit states by
applying partial safety factors to actions or the effect of actions,
material properties or resistances. For geotechnical design, it
has so far been impossible to find a single way of combining
factors between actions, ground properties and resistances, and
thus three different design approaches (DA) are permitted in
EN1997. However, for slope stability there seems to be a rather
large consensus on how safety should be applied. Most
countries have chosen either design approach 1 (DA1) or 3
(DA3) for slope stability. DA1 consists of two combinations of
sets of partial factors, of which combination 2 is the one
relevant for slope stability and analogous to DA3. As there is
such a large agreement on applying safety for slope stability it
could be considered that this part of the geotechnical design is
well formulated in the Eurocodes. The intention of this article is
to critically review the applied partial factor of safety method
for the slope stability with respect to reliability and
consequences of failure.
2 SHORT OVERVIEW ON THE EUROCODES
2.1
Safety in Eurocodes
According to EN-1990 the partial safety factors should account
for the possible unfavourable deviation of the property from its
characteristic value and the uncertainties in the model used in
calculations. The consequences of the ultimate limit states are
further considered based on three consequences classes. Therein
the consequence of failure is accounted by multiplying the
factors for actions by a separate factor depending on the
consequence. In principle the partial factors of safety can be
determined in two ways. One is the conventional method where
the factors are calibrated to past experience. The other is to use
probabilistic methods and calibrate the factors against a target
reliability index value.
For the most common design situations, corresponding to
the reliability class RC2, the recommended reliability index
β
for a 50 years reference period is 3.8. This corresponds to
nominal probability for ultimate limit states of approximately
1/15 000. Whilst some of the partial factors for actions have
been determined based on probabilistic methods, it is the
understanding of the authors that material and resistance factors
for geotechnical design has mainly been determined based on
calibration to old codes.
2.2
Slope stability in EN-1997 according to DA1 and DA3
As previously discussed, most European countries have chosen
either DA1 combination 2 or DA3 to be used in slope stability.
