UQL AB U SER M ANUAL

S TRUCTURAL R ELIABILITY (R ARE E VENT

E STIMATION )
S. Marelli, R. Schöbi, B. Sudret

C HAIR OF R ISK , S AFETY AND U NCERTAINTY Q UANTIFICATION Risk, Safety &

S TEFANO -F RANSCINI -P LATZ 5 Uncertainty Quantification

CH-8093 Z ÜRICH
How to cite UQL AB

S. Marelli, and B. Sudret, UQLab: A framework for uncertainty quantification in Matlab, Proc. 2nd Int. Conf. on
Vulnerability, Risk Analysis and Management (ICVRAM2014), Liverpool, United Kingdom, 2014, 2554-2563.

How to cite this manual

S. Marelli, R. Schöbi, B. Sudret, UQLab user manual – Structural reliability (Rare event estimation), Report
UQLab-V2.0-107, Chair of Risk, Safety and Uncertainty Quantification, ETH Zurich, Switzerland, 2022

BIBTEX entry

@TechReport{UQdoc_20_107,
author = {Marelli, S. and Sch\"obi, R. and Sudret, B.},
title = {{UQLab user manual -- Structural reliability (Rare event estimation)}},
institution = {Chair of Risk, Safety and Uncertainty Quantification, ETH Zurich,
Switzerland},
year = {2022},
note = {Report UQLab-V2.0-107}
}

List of contributors:

P.-R. Wagner Implementation of SSER

C. Lamas Original implementation of FORM, SORM, Monte Carlo and importance sampling methods
Document Data Sheet

Document Ref. UQL AB-V2.0-107

Title: UQL AB user manual – Structural reliability (Rare event estimation)

Authors: S. Marelli, R. Schöbi, B. Sudret

Chair of Risk, Safety and Uncertainty Quantification, ETH Zurich,
Switzerland
Date: 01/02/2022

Doc. Version Date Comments

V2.0 01/02/2022 UQL AB V2.0 release
• added SSER method
V1.4 01/02/2021 UQL AB V1.4 release
V1.3 19/09/2019 UQL AB V1.3 release
V1.2 22/02/2019 UQL AB V1.2 release
V1.1 05/07/2018 UQL AB V1.1 release
• Updated usage and reference list for AK-MCS
V1.0 01/05/2017 UQL AB V1.0 release
• added PC-Kriging to AK-MCS
• bugfixes and display improvements
V0.92 15/02/2016 Initial release
 Abstract

Structural reliability methods aim at the assessment of the probability of failure of complex
systems due to uncertainties associated to their design, manifacturing, environmental and
operating conditions. The name structural reliability comes from the emergence of such
computational methods back in the mid 70’s to evaluate the reliability of civil engineering
structures. As these probabilities are usually small (e.g. 10−2 − 10−8 ), this type of problems
is also known as rare events estimation in the recent statistics literature.
The structural reliability module of UQL AB offers a comprehensive set of techniques for the
efficient estimation of the failure probability of a wide range of systems. Classical (crude
Monte Carlo simulation, FORM/SORM, Subset Simulation) and state-of-the-art algorithms
(AK-MCS, SSER) are available and can be easily deployed in association with other UQL AB
tools, e.g. surrogate modelling or sensitivity analysis.
The structural reliability user manual is divided in three parts:

• A short introduction to the main concepts and techniques used to solve structural reli-
ability problems, with a selection of references to the relevant literature

• A detailed example-based guide, with the explanation of most of the available options
and methods

• A comprehensive reference list detailing all the available functionalities in the UQL AB
structural reliability module.

Keywords: Structural Reliability, FORM, SORM, Importance Sampling, Monte Carlo Simula-
tion, Subset Simulation, AK-MCS, SSER, UQL AB, rare event estimation
Contents

1 Theory 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Problem statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.1 Limit-state function . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2.2 Failure Probability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Strategies for the estimation of Pf . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Approximation methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4.1 First Order Reliability Method (FORM) . . . . . . . . . . . . . . . . . . 4
1.4.2 Second Order Reliability Method (SORM) . . . . . . . . . . . . . . . . 9
1.5 Simulation methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.1 Monte Carlo Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . 11
1.5.2 Importance Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.5.3 Subset Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.6 Metamodel-based methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.6.1 Adaptive Kriging Monte Carlo Simulation . . . . . . . . . . . . . . . . 15
1.6.2 Stochastic spectral embedding-based reliability . . . . . . . . . . . . . 18

2 Usage 25
2.1 Reference problem: R-S . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.2 Problem set-up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3 Reliability analysis with different methods . . . . . . . . . . . . . . . . . . . . 26
2.3.1 First Order Reliability Method (FORM) . . . . . . . . . . . . . . . . . . 26
2.3.2 Second Order Reliability Method (SORM) . . . . . . . . . . . . . . . . 29
2.3.3 Monte Carlo Simulation (MCS) . . . . . . . . . . . . . . . . . . . . . . 30
2.3.4 Importance Sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
2.3.5 Subset Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.3.6 Adaptive Kriging Monte Carlo Simulation (AK-MCS) . . . . . . . . . . 38
2.3.7 Stochastic spectral embedding-based reliability (SSER) . . . . . . . . . 41
2.4 Advanced limit-state function options . . . . . . . . . . . . . . . . . . . . . . . 43
2.4.1 Specify failure threshold and failure criterion . . . . . . . . . . . . . . 43
2.4.2 Vector Output . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2.5 Excluding parameters from the analysis . . . . . . . . . . . . . . . . . . . . . 44
3 Reference List 45
3.1 Create a reliability analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.2 Accessing the results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.2.1 Monte Carlo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.2.2 FORM and SORM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3.2.3 Importance sampling . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.2.4 Subset simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.2.5 AK-MCS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.2.6 SSER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.3 Printing/Visualizing of the results . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.3.1 Printing the results: uq print . . . . . . . . . . . . . . . . . . . . . . . 63
3.3.2 Graphically display the results: uq display . . . . . . . . . . . . . . . . 63
Chapter 1

Theory

1.1 Introduction

A structural system is defined as a structure required to provide specific functionality under

well-defined safety constraints. Such constraints need to be taken into account during the
system design phase in view of the expected environmental/operating loads it will be subject
to.
In the presence of uncertainties in the physical properties of the system (e.g. due to tolerances
in the manufacturing), in the environmental loads (e.g. due to exceptional weather condi-
tions), or in the operating conditions (e.g. traffic), it can occur that the structure operates
outside of its nominal range. In such cases, the system encounters a failure.
Structural reliability analysis deals with the quantitative assessment of the probability of
occurrence of such failures (probability of failure), given a model of the uncertainty in the
structural, environmental and load parameters.
Following the formalism introduced in Sudret (2007), this chapter is intended as a brief the-
oretical introduction and literature review of the available tools in the structural reliability
module of UQL AB. Consistently with the overall design philosophy of UQL AB, all the algo-
rithms presented follow a black-box approach, i.e. they rely on the point-by-point evaluation
of a computational model, without knowledge about its inner structure.

1.2 Problem statement

1.2.1 Limit-state function

A limit state can be defined as a state beyond which a system no longer satisfies some perfor-
mance measure (ISO Norm 2394). Regardless on the choice of the specific criterion, a state
beyond the limit state is classified as a failure of the system.
Consider a system whose state is represented by a random vector of variables X ∈ DX ⊂ RM .
One can define two domains Ds , Df ⊂ DX that correspond to the safe and failure regions of
the state space DX , respectively. In other words, the system is failing if the current state
x ∈ Df and it is operating safely if x ∈ Ds . This classification makes it possible to construct
a limit-state function g(X) (sometimes also referred to as performance function) that assumes

1
UQL AB user manual

Figure 1: Schematic representation of the safe and failure domains Ds and Df and the cor-
responding limit-state surface g(x) = 0.

positive values in the safe domain and negative values in the failure domain:

x ∈ Ds ⇐⇒ g(x) > 0
(1.1)
x ∈ Df ⇐⇒ g(x) ≤ 0

The hypersurface in M dimensions defined by g(x) = 0 is known as the limit-state surface,

and it represents the boundary between safe and failure domains. A graphical representation
of Ds , Df and the corresponding limit-state surface g(x) = 0 is given in Figure 1.

1.2.2 Failure Probability

If the random vector of state variables X is described by a joint probability density function
(PDF) X ∼ fX (x), then one can define the failure probability Pf as:

Pf = P (g(X) ≤ 0) . (1.2)

This is the probability that the system is in a failed state given the uncertainties of the state
parameters. The failure probability Pf is then calculated as follows:
Z Z
Pf = fX (x)dx = fX (x)dx. (1.3)
Df {x: g(x)≤0}

Note that the integration domain in Eq. (1.3) is only implicitly defined by Eq. (1.1), hence
making its direct estimation practically impossible in the general case. This limitation can be
circumvented by introducing the indicator function of the failure domain, a simple classifier
given by: (
1 if g(x) ≤ 0
1Df (x) = , x ∈ DX .
0 if g(x) > 0
In other words, 1Df (x) = 1 when the input parameters x cause the system to fail and

UQL AB-V2.0-107 -2-

 Structural reliability (Rare event estimation)

1Df (x) = 0 otherwise. This function allows one to cast Eq. (1.3) as follows:
Z
 
Pf = 1Df (x)fX (x)dx = E 1Df (X) , (1.4)
DX

where E [·] is the expectation operator with respect to the PDF fX (x). This reduces the
calculation of Pf to the estimation of the expectation value of 1Df (X).

1.3 Strategies for the estimation of Pf

From the definition of 1Df (x) in Section 1.2.2 it is clear that determining whether a certain
state vector x ∈ DX belongs to Ds or Df requires the evaluation of the limit-state function
g(x). In the general case this operation can be computationally expensive, e.g. when it
entails the evaluation of a computational model on the vector x. For a detailed overview
of standard structural reliability methods and applications, see e.g. Ditlevsen and Madsen
(1996); Melchers (1999); Lemaire (2009).
In the following, three strategies are discussed for the evaluation of Pf , namely approxima-
tion, simulation and adaptive surrogate-modelling-based methods.

Approximation methods
Approximation methods are based on approximating the limit-state function locally at a
reference point (e.g. with a linear or quadratic Taylor expansion). This class of methods
can be very efficient (in that only a relatively small number of model evaluations is needed
to calculate Pf ), but it tends to become unreliable in the presence of complex, non-linear
limit-state functions. Two approximation methods are currently available in UQL AB:

• FORM (First Order Reliability Method) – it is based on the combination of an iterative

gradient-based search of the so-called design point and a local linear approximation of
the limit-state function in a suitably transformed probabilistic space.

• SORM (Second Order Reliability Method) – it is a second-order refinement of the solution

of FORM. The computational costs associated to this refinement increase rapidly with
the number of input random variables M .

Simulation methods
Simulation methods are based on sampling the joint distribution of the state variables X
and using sample-based estimates of the integral in Eq. (1.4). At the cost of being compu-
tationally very expensive, they generally have a well-characterized convergence behaviour
that can be exploited to calculate confidence bounds on the resulting Pf estimates. Three
sampling-based algorithms are available in UQL AB:

• Monte Carlo simulation – it is based on the direct sample-based estimation of the expec-
tation value in Eq. (1.4). The total costs increase very rapidly with decreasing values
of the probability Pf to be computed.

UQL AB-V2.0-107 -3-

UQL AB user manual

• Importance Sampling – it is based on improving the efficiency of Monte Carlo simulation

by changing the sampling density so as to favour points in the failure domain Df . The
choice of the importance sampling (a.k.a. instrumental) density generally uses FORM
results.

• Subset Simulation – it is based on iteratively solving and combining a sequence of condi-

tional reliability analyses by means of Markov Chain Monte Carlo (MCMC) simulation.

Metamodel-based adaptive methods

Metamodel-based adaptive methods are based on iteratively building surrogate models
that approximate the limit-state function in the direct vicinity of the limit-state surface. The
metamodels (see e.g. UQL AB User Manual – Polynomial Chaos Expansionsand UQL AB User
Manual – Kriging (Gaussian process modelling)) are adaptively refined by adding limit-state
function evaluations to their experimental designs until a suitable convergence criterion re-
lated to the accuracy of Pf is satisfied. One algorithm is currently available in UQL AB,
namely Adaptive Kriging Monte Carlo Simulation (AK-MCS). It is based on building a Kriging
(aka Gaussian process regression) surrogate model from a small initial sampling of the input
vector X. The surrogate is then iteratively refined close to the currently estimated limit-state
surface so as to evaluate accurately the probability of failure.

In the following, a detailed description of each of the methods is given.

1.4 Approximation methods

1.4.1 First Order Reliability Method (FORM)

The first order reliability method aims at the approximation of the integral in Eq. (1.3) with
a three-step approach:

• An isoprobabilistic transform of the input random vector X ∼ fX (x) into a standard

normal vector U ∼ N (0, IM )

• A search for the most likely failure point in the standard normal space (SNS), known
as the design point U ∗

• A linearization of the limit-state surface at the design point U ∗ and the analytical com-
putation of the resulting approximation of Pf .

1.4.1.1 Isoprobabilistic transform

The first step of the FORM method is to transform the input random vector X ∼ fX into
a standard normal vector U ∼ N (0, IM ). The corresponding isoprobabilistic transform T
reads:
X = T −1 (U ) (1.5)

UQL AB-V2.0-107 -4-

 Structural reliability (Rare event estimation)

Figure 2: Graphical representation of the isoprobabilistic transform from physical to standard

normal space in Eq. (1.5). From Sudret, 2015: Lectures on structural reliability and risk
analysis.

For details about the available isoprobabilistic transforms in UQL AB, please refer to the
UQL AB User Manual – the INPUT module (Section 1.5).

This transform can be used to map the integral in Eq. (1.3) from the physical space of X to
the standard normal space of U :
Z Z
Pf = fX (x)dx = ϕM (u)du (1.6)
Df {u∈RM : G(u)≤0}

where G(u) = g(T −1 (u)) is the limit-state function evaluated in the standard normal space
and ϕM (u) is the standard multivariate normal PDF given by:
 
−M/2 1 2 2
ϕM (u) = (2π) exp − (u1 + · · · + uM ) . (1.7)
2

A graphical illustration of the effects of this transform for a simple 2-dimensional case is
given in Figure 2. The advantage of casting the problem in the standard normal space is that
it is a probability space equipped with the Gaussian probability measure PG :
Z Z
ϕM (u)du = (2π)−M/2 exp u21 + · · · + u2M du.


PG (U ∈ A) = (1.8)
A A

This probability measure is spherically symmetric: ϕM (u) only depends on kuk2 and it de-
cays exponentially as ϕM (u) ∼ exp −kuk2 /2 . Therefore, when evaluating the integral in

Eq. (1.6) in the standard normal space, most of the contributions are given by the region clos-
est to the origin. The FORM method capitalizes on this property by linearly approximating
the limit-state surface in the region closest to the origin of the standard normal space.

UQL AB-V2.0-107 -5-

UQL AB user manual

Figure 3: Graphical representation of the linearization of the limit-state function around the
design point at the basis of the FORM estimation of Pf . From Sudret, 2015: Lectures on
structural reliability and risk analysis.

1.4.1.2 Search for the design point

The design point U ∗ is defined as the point in the failure domain closest to the origin of the
standard normal space:
U ∗ = argmin {kuk, G(u) ≤ 0)} . (1.9)
u∈RM

Due to the probability measure in Eq. (1.8), U ∗ can be interpreted as the most likely failure
point in the standard normal space. The norm of the design point kU ∗ k is an important
quantity in structural reliability known as the Hasofer-Lind reliability index (Hasofer and Lind,
1974):
βHL = kU ∗ k. (1.10)

An important property of the βHL index is that it is directly related to the exact failure prob-
ability Pf in the case of linear limit-state function in the standard normal space:

Pf = Φ(−βHL ), (1.11)

where Φ is the standard normal cumulative density function. The estimation of Pf in the
FORM algorithm is based on approximating the limit-state function as the hyperplane tan-
gent to the limit-state function at the design point. Figure 3 illustrates this approximation
graphically for the two-dimensional case.

In the general non-linear case, Eq. (1.9) may be cast as a constrained optimization problem
with Lagrangian:
1
L(u, λ) = kuk2 + λG(u) (1.12)
2

UQL AB-V2.0-107 -6-

 Structural reliability (Rare event estimation)

where λ is the Lagrange multiplier. The related optimality conditions read:

∇u L(U ∗ , λ∗ ) = 0,
∂L ∗ ∗ (1.13)
(U , λ ) = 0,
∂λ
which can be explicitly written as:

G(U ∗ ) = 0,
(1.14)
U ∗ + λ∗ ∇G(U ∗ ) = 0.

The first condition in Eq. (1.14) guarantees that the design point belongs to the limit-state
surface. The second condition guarantees that the vector U ∗ is colinear to the limit-state
surface normal vector at U ∗ , i.e. ∇G(U ∗ ). The standard iterative approach to solve this non-
linear constrained optimization problem is given by the Rackwitz-Fiessler algorithm (Rackwitz
and Fiessler, 1978).

Hasofer-Lind - Rackwitz-Fiessler algorithm (HL-RF)

The rationale behind the Rackwitz-Fiessler algorithm is to iteratively solve a linearized
problem around the current point. Normally, the algorithm is started with U0 = 0.
At each iteration, the limit-state function is approximated as:

G(U ) ≈ G(Uk ) + ∇G|Uk · (U − Uk ) (1.15)

The two optimality conditions in Eq. (1.14) read for each iteration k:

∇G|Uk · (Uk+1 − Uk ) + G(Uk ) = 0

(1.16)
Uk+1 = λ∇G|Uk ,

which after some basic algebra reduce to:

∇G|Uk · Uk − G(Uk )
Uk+1 = ∇G|Uk . (1.17)
k∇G|Uk k2

By introducing the unit vector:

∇G|Uk
αk = − , (1.18)
k∇G|Uk k
one finally obtains:  
G(Uk )
Uk+1 = αk · Uk + αk . (1.19)
k∇G|Uk k
The associated estimate of the reliability index βk associated to the k-th iteration is then:

G(Uk )
β k = αk · Uk + . (1.20)
k∇G|Uk k

Perfect convergence of the algorithm is obtained when G(U ∗ ) = 0, yielding βHL = α∗ · U ∗ .

However, in practice the algorithm is iterated until some stopping criteria are satisfied, i.e.,

UQL AB-V2.0-107 -7-

UQL AB user manual

Table 1: Common stopping criteria for the FORM algorithm and associated description.

Criterion Typical value Description

|βk+1 − βk | ≤ β 10−3 Stability of β between iterations
kUk+1 − Uk k ≤ U 10−3 Stability of U between iterations
|G(Uk+1 )/G(U0 )| ≤ G 10−6 Closeness to the limit-state surface

until one or more convergence conditions are verified. The standard stopping criteria used
in FORM are reported in Table 1.
Note: In UQL AB the gradients ∇G(Uk ) in Eqs. (1.13) to (1.20) are calculated numeri-
cally in the standard normal space and not in the physical space.

Improved HL-RF algorithm (iHL-RF)

The Rackwitz-Fiessler algorithm is a particular case of a wide class of iterative algorithms
generically denoted as descent direction algorithms, of the form:

Uk+1 = Uk + λk dk , (1.21)

where λk is the step size at the k-th iteration and dk is the corresponding descent direction
given by:
∇G|Uk · Uk − G(Uk )
dk = ∇G|Uk − Uk . (1.22)
k∇G|Uk k2
In the original HL-RF algorithm, λk = 1 ∀k. Zhang and Der Kiureghian (1995) proposed
an “improved” version of the same algorithm that takes advantage of a more sophisticated
step-size calculation based on the assumption that G(U ) is differentiable everywhere. They
introduced the merit function m(U ):
1
m(U ) = kU k + c|G(U )|, (1.23)
2
kU k
where c > is a real penalty parameter. This function has its global minimum in the
k∇G(U )k
same location as the original Eq. (1.9), as well as the same descent direction d. In addition,
it allows one to use the Armijo rule (Zhang and Der Kiureghian, 1995) to determine the best
step length λk at each iteration as:

λk = max{bs | m(Uk + bs dk ) − m(Uk ) ≤ −abs ∇m(Uk ) · dk }, (1.24)

s

where a, b ∈ (0, 1) are pre-selected parameters, and s ∈ N.

1.4.1.3 FORM results

Once the design point U ∗ is identified, it can be used to extract additional important infor-
mation. According to Eq. (1.20), after the convergence of FORM the Hasofer-Lind index βHL

UQL AB-V2.0-107 -8-

 Structural reliability (Rare event estimation)

is given by:
βHL = α∗ · U ∗ , (1.25)

with associated failure probability:

Pf,FORM = Φ−1 (βHL ) (1.26)

The local sensitivity indices Si are defined as the fraction of the variance of the safety margin
g(X) = G(U ) due to the component of the design vector Ui . It can be demonstrated that
they are given by:  2

∂G 
Si = /k∇G(U ∗ )k2 . (1.27)
∂ui U ∗
From Eq. (1.18) it follows that:
Si = αi2 . (1.28)

If the input variables are independent, then each coordinate in the SNS Ui corresponds to a
single input variable in the physical space Xi . Therefore, the importance factor of each Xi is
identified with αi2 .

1.4.2 Second Order Reliability Method (SORM)

The second-order reliability method (SORM) is a second-order refinement of the FORM Pf

estimate. After the design point U ∗ is identified by FORM, the failure probability is approxi-
mated by a tangent hyperparaboloid defined by the second order Taylor expansion of G(U ∗ )
given by:
1
G(U ) ≈ ∇GT|U ∗ · (U − U ∗ ) + (U − U ∗ )T H(U − U ∗ ), (1.29)
2
where H is the Hessian matrix of the second derivatives of G(U ) evaluated at U ∗ .

The failure probability in the SORM approximation can be written as a correction factor of
the FORM estimate that depends on the curvatures of the hyper-hyperboloid in Eq. (1.29).
To estimate the curvatures, the hyperparaboloid in Eq. (1.29) is first cast in canonical form
by rotating the coordinates system such that one of its axes is the α vector. Usually the last
coordinate is chosen arbitrarily for this purpose. A rotation matrix Q can be built by setting
α as its last row and by using the Gram-Schmidt procedure to orthogonalize the remaining
components of the basis. Q is a square matrix such that QT Q = I. The resulting vector V
satisfies:
U = QV . (1.30)

In the new coordinates system and after some basic algebra (see e.g. Breitung (1989) and
Cai and Elishakoff (1994)), one can rewrite Eq. (1.29) as:
1
G(V ) ≈ k∇G(U ∗ )k(β − VM ) + (V − V ∗ )QHQT (V − V ∗ ) (1.31)
2
def
where β is the Hasofer-Lind reliability index calculated by FORM, VM = αT (QT V ) and

UQL AB-V2.0-107 -9-

UQL AB user manual

Figure 4: Comparison between FORM and SORM approximations of the failure domain for
a simple 2-dimensional case. From Sudret, 2015: Lectures on structural reliability and risk
analysis.

V ∗ = {0, · · · , β}T is the design point in the new coordinates system. By dividing Eq. (1.31)
def
by the gradient norm k∇G(U ∗ )k and introducing the matrix A = QHQT /k∇G(U ∗ )k, one
obtains:
1
G̃(V ) ≈ β − VM + (V − V ∗ )A(V − V ∗ ), (1.32)
2
where G̃(V ) = G(V )/k∇G(U ∗ )k. After neglecting second-order terms in VM and diagonal-
izating the A matrix via eigenvalue decomposition one can rewrite Eq. (1.32) explicitly in
terms of the curvatures κi of an hyper-paraboloid with axis α:
M −1
1 X
G̃(V ) ≈ β − VM + κi Vi . (1.33)
2
1

For small curvatures κi < 1, the failure probability Pf can be approximated by the Breitung
formula (Breitung, 1989):
M −1
1
(1 + βHL κi )− 2
Y
B
Pf,SORM = Φ(−βHL ) κi < 1 (1.34)
i=1

Note that for small curvatures the Breitung formula approaches the FORM linear limit. The
accuracy of Eq. (1.34) decreases for larger values of κi , sometimes even if κi < 1 (Cai and
Elishakoff, 1994). A more accurate formula is given by the Hohenbichler formula (Hohen-
bichler et al., 1987):
M −1  − 1
H
Y ϕ(βHL ) 2
Pf,SORM = Φ(−βHL ) 1+ κi . (1.35)
Φ(−βHL )
i=1

Additional methods are available in the literature for the exact computation of the failure
probability, e.g.Tvedt (1990). They are, however, outside the scope of this manual.

UQL AB-V2.0-107 - 10 -
 Structural reliability (Rare event estimation)

1.5 Simulation methods

1.5.1 Monte Carlo Simulation

Monte Carlo (MC) simulation is used to directly compute the integral in Eq. (1.4) by sampling
the probabilistic input model. Given a sample of size N of the input random vector X,
X = x(1) , . . . x(N ) , the unbiased MCS estimator of the expectation value in Eq. (1.4) is


given by:
N
def 1 X Nf ail
Pf,MC = Pf =
b 1Df (x(k) ) = , (1.36)
N N
k=1

where Nf ail is the number of samples such that g(x) ≤ 0. In other words, the Monte Carlo
estimate of the failure probability is the fraction of samples that belong to the failure domain
over the total number of samples. An advantage of Monte Carlo simulation is that it provides
an error estimate for Eq. (1.36). Indeed the indicator function 1Df (x) follows by construction
a Bernoulli distribution with mean µ1Df = Pf and variance σ12D = Pf (1 − Pf ). For large
f
enough N it can be approximated by the normal distribution:
 
Pbf ∼ N µ b1Df ,
b1Df , σ (1.37)
q
where µ b1Df = Pbf and σ b1Df = Pbf (1 − Pbf ). Hence, the estimator of Pf has a normal
distribution with mean Pbf and variance given by:

σ12D Pbf (1 − Pbf )

bP2 f
σ = f
= . (1.38)
N N

Confidence intervals on Pbf can therefore be given as follows (Rubinstein, 1981):

h i
def def
Pbf ∈ Pbf− = Pbf + σbPf Φ−1 (α/2), Pbf+ = Pbf + σ
bPf Φ−1 (1 − α/2) , (1.39)

where Φ(x) is the standard normal CDF and α ∈ [0, 1] is a scalar such that the calculated
bounds correspond to a confidence level of 1 − α. An important measure for assessing the
convergence of a MCS estimator is given by the coefficient of variation CoV defined as:
v
σPbf
u
u 1 − Pbf
CoV = =t . (1.40)
Pbf N Pbf

The coefficient of variation of the MCS estimate of a failure probability therefore decreases
√
with N and increases with decreasing Pf . To give an example, to estimate a Pf = 10−3
with 10% accuracy N = 105 samples are needed. The CoV is often used as a convergence
criterion to adaptively increase the MC sample size until some desired CoV is reached.
An associated generalized reliability index βM CS can be defined as:

βMCS = −Φ−1 (Pbf ). (1.41)

UQL AB-V2.0-107 - 11 -
UQL AB user manual

In analogy, upper and lower confidence bounds on βMCS can be directly inferred from the
confidence bounds on Pbf in Eq. (1.39):

±
βMCS = −Φ−1 (Pbf± ). (1.42)

The MCS method is powerful, when applicable, due to its statistically sound formulation and
global convergence. However, its main drawback is the relatively slow converge rate that
depends strongly on the probability of failure.

1.5.2 Importance Sampling

Importance sampling (IS) is an extension of the FORM and MCS methods that combines the
fast convergence of FORM with the robustness of MC. The basic idea is to recast Eq. (1.4) as:
Z  
fX (x) fX (X)
Pf = 1Df (x) Ψ(x)dx = EΨ 1Df (X) , (1.43)
DX Ψ(x) Ψ(X)

where Ψ(X) is an M −dimensional sampling distribution (also referred to as importance dis-

tribution) and EΨ denotes the expectation value with respect to the same distribution. The
estimate of Pf given a sample X = x(1) , . . . , x(N ) drawn from Ψ is therefore given by:


N
1 X fX (x(k) )
Pf,IS = 1Df (x(k) ) . (1.44)
N
k=1
Ψ(x(k) )

In the standard normal space, Eq. (1.43) can be rewritten as:

 
−1 ϕM (U )
Pf = EΨ 1Df (T (U )) . (1.45)
Ψ(U )

When the results from a previous FORM analysis are available, a particularly efficient sam-
pling distribution in the standard normal space is given by (Melchers, 1999):

Ψ(u) = ϕM (u − U ∗ ) (1.46)

where U ∗ is the estimated design point. Given a sample U = u(1) , . . . , u(N ) of Ψ(u), the


estimate of Pf becomes:

N
1
X    
Pf,IS = 2
exp −βHL /2 1Df T −1 (u(k) ) exp −u(k) · U ∗ (1.47)
N
k=1

with corresponding variance:

N
!2
1 1 X   ϕ(u(k) )
bP2 f,IS
σ = 1Df T −1 (u(k) ) − Pf,IS . (1.48)
N N −1
k=1
Ψ(u(k) )

±
The coefficient of variation and the confidence bounds Pf,IS can be calculated analogously to
Eqs. (1.40) and (1.39), respectively, and can be used as a convergence criterion to adaptively

UQL AB-V2.0-107 - 12 -
 Structural reliability (Rare event estimation)

improve the estimation of Pf,IS . The corresponding generalized reliability index reads:

βIS = −Φ−1 (Pbf,IS ), (1.49)

with upper and lower bounds:

± ±
βIS = −Φ−1 (Pbf,IS ). (1.50)

Note that exact convergence of FORM is not necessary to obtain accurate results, even an
approximate sampling distribution can significantly improve the convergence rate compared
to standard MC sampling.

1.5.3 Subset Simulation

Monte Carlo simulation may require a large number of limit-state function evaluations to
converge with an acceptable level of accuracy when Pf is small (see Eq. (1.40)). Subset sim-
ulation is a technique introduced by Au and Beck (2001) that aims at offsetting this limitation
by solving a series of simpler reliability problems with intermediate failure thresholds.
Consider a sequence of failure domains D1 ⊃ D2 ⊃ · · · ⊃ Dm = Df such that Df = m
T
k=1 Dk .
With the conventional definition of limit-state function in Eq. (1.1), such sequence can be
built with a series of decreasing failure thresholds t1 > · · · > tm = 0 and the corresponding
intermediate failure domains Dk = {x : g(x) ≤ tk }. One can then combine the probability
mass of each intermediate failure region by means of conditional probability. By introducing
the notation P (DX ) = P (x ∈ DX ) one can write (Au and Beck, 2001):
m m−1
!
\ Y
Pf = P (Dm ) = P P (Dk ) = P (D1 ) P (Di+1 |Di ) . (1.51)
k=1 i=1

With an appropriate choice of the intermediate thresholds t1 , . . . , tm , Eq. (1.51) can be eval-
uated as a series of structural reliability problems with relatively high probabilities of failure
that are then solved with MC simulation. In practice the intermediate probability thresholds
ti are chosen on-the-fly such that they correspond to intermediate values P (Dk ) ≈ 0.1. The
convergence of each intermediate estimation is therefore much faster than the direct search
for Pf given in Eq. (1.36).

1.5.3.1 Sampling

To estimate Pf from Eq. (1.51) one thus needs to estimate the intermediate probabilities
P (D1 ) and conditional probabilities {P (Di+1 |Di ) , i = 1, . . . , m − 1}. Given an initial thresh-
old t1 , P (D1 ) can be readily estimated from a sample of size NS of the input distribution
X = x(1) , . . . , x(N ) with Eq. (1.36):


NS
1 X
P (D1 ) ≈ Pb1 = 1D1 (x(k) ). (1.52)
NS
k=1

The remaining conditional probabilities can be estimated similarly, but an efficient sampling

UQL AB-V2.0-107 - 13 -
UQL AB user manual

algorithm is needed for the underlying conditional distributions. The latter can be efficiently
accomplished by using the modified Metropolis-Hastings Markov Chain Monte Carlo (MCMC)
simulation introduced by Au and Beck (2001).

1.5.3.2 Intermediate failure thresholds

The efficiency of the subset-simulation method depends on the choice of the intermediate
failure thresholds tk . If the thresholds are too large the MCS convergence in each subset
would be very good, but the number of subsets needed would increase. Vice-versa, too small
intermediate thresholds would correspond to fewer subsets with inaccurate estimates of the
underying P (Dk ). A strategy to deal with this problem comes by sampling each subset Dk and
determining each threshold tk as the empirical quantile that correspond to a predetermined
failure probability, typically P (Dk ) ≈ P0 = 0.1. Note that for practical reasons, P0 is normally
limited to 0 < P0 ≤ 0.5. For each subset, the samples falling below the calculated threshold
are used as MCS seeds for the next subset (Au and Beck, 2001).

1.5.3.3 Subset simulation algorithm

The subset simulation algorithm can be summarized in the following steps:

1. Sample the original space with standard MC sampling (see UQL AB User Manual – the
INPUT module for efficient sampling strategies)

2. Calculate the empirical quantile tk in the current subset such that Pbk ≈ P0

3. Using the samples below the identified quantile as the seeds of parallel MCMC chains,
sample Dk+1 |Dk until a predetermined number of samples is available

4. Repeat Steps 2 and 3 until the identified quantile tm < 0

5. Calculate the failure probability of the last subset Pbm by setting tm = 0

6. Combine the intermediate calculated failure probabilities into the final estimate of Pbf .
The last step of the algorithm consists simply in evaluating Eq. (1.51) with the current esti-
mates of the conditional probabilities Pi :
m
Y
Pbf,SS = Pbi = P0m−1 Pbm . (1.53)
i=1

1.5.3.4 Error estimation

Due to the intrinsic correlation of the samples drawn from each subset resulting from the
MCMC sampling strategy, the estimation of a CoV for the Pf estimate in Eq. (1.53) is non-
trivial. Au and Beck (2001) and Papaioannou et al. (2015) derived an estimate for the CoV
of Pbf :
Xm
CoVf ≈ δi2 , (1.54)
i=1

UQL AB-V2.0-107 - 14 -
 Structural reliability (Rare event estimation)

where m is the number of subsets and δi is defined as:

r
1 − Pi
δi = (1 + γi ), (1.55)
N Pi
with
N/Ns  
X kNs
γi = 2 1− ρi (k), (1.56)
N
k=1

where NS is the number of seeds, Pi is the conditional failure probability of the i−th subset
and ρi (k) is the average k-lag auto-correlation coefficient of the Markov Chain samples in
±
the i−th subset. By assuming normally distributed errors, confidence bounds Pf,SS can be
given on Pf,SS based on the calculated CoVf in analogy with Eq. (1.39). The corresponding
generalized reliability index reads:

βSS = −Φ−1 (Pbf,SS ), (1.57)

with upper and lower bounds:

± ±
βSS = −Φ−1 (Pbf,SS ). (1.58)

1.6 Metamodel-based methods

1.6.1 Adaptive Kriging Monte Carlo Simulation

Adaptive Kriging Monte Carlo Simulation (AK-MCS) combines Monte Carlo simulation with
adaptively built Kriging (a.k.a. Gaussian process modelling) metamodels. In cases where
the evaluation of the limit-state function is costly, Monte Carlo simulation and its variants
may become intractable due to the large number of limit-state function evaluations they
require. In AK-MCS, a Kriging metamodel surrogates the limit-state function to reduce the
total computational costs of the Monte Carlo simulation.
Kriging metamodels (see UQL AB User Manual – Kriging (Gaussian process modelling)) pre-
dict the value of the limit-state function most accurately in the vicinity of the experimental
design samples X = X (1) , . . . , X (n) . These samples, however, are generally not optimal


to estimate the failure probability. Thus, an adaptive experimental design algorithm is intro-
duced to increase the accuracy of the surrogate model in the vicinity the limit-state function.
This is achieved by adding carefully selected samples to the experimental design of the Krig-
ing metamodel based on the current estimate of the limit-state surface (g(x) = 0).
The adaptive experimental design algorithm is summarized as follows (Echard et al., 2011;
Schöbi et al., 2016):

1. Generate a small initial experimental design X = x(1) , . . . , x(N0 ) and evaluate the


corresponding limit-state function responses Y = y (1) , . . . , y (N0 ) = g(x(1) ), . . . , g(x(N0 ) )

 

2. Train a Kriging metamodel gb based on the experimental design {X , Y}

UQL AB-V2.0-107 - 15 -
UQL AB user manual

3. Generate a large set of NM C candidate samples S = s(1) , . . . , s(NM C ) and predict the


corresponding metamodel responses gb(s(1) ), . . . , gb(s(NM C ) )



4. Choose the best next sample s∗ to be added to the experimental design X based on an
appropriate learning function

5. Check whether some convergence criterion is met. If it is, skip to Step 7, otherwise
continue with Step 6

6. Add s∗ and the corresponding limit-state function response y ∗ = g(s∗ ) to the experi-
mental design of the metamodel. Return to Step 2

7. Estimate the failure probability through Monte Carlo simulation with the final limit-
state function surrogate gb(x).

1.6.1.1 Selection of the best next candidate sample

A learning function is a measure of the attractiveness of a candidate sample X with respect to

improving the estimate of the failure probability when it is added to the experimental design
X . A variety of learning functions are available in the literature (Bichon et al., 2008; Dani
et al., 2008; Echard et al., 2011; Srinivas et al., 2012; Ginsbourger et al., 2013; Dubourg,
2011), amongst which the U -function (Echard et al., 2011). The U -function is based on the
concept of misclassification and is defined for a Gaussian process as follows:
 
µgb(X)
U (X) = , (1.59)
σgb(X)

where µgb(X) and σgb(X) are the prediction mean and standard deviation of gb. A misclassifica-
tion happens when the sign of the surrogate model and the sign of the underlying limit-state
function do not match. The corresponding probability of misclassification is then:

Pm (X) = Φ (−U (X)) ,

where Φ is the CDF of a standard Gaussian variable.

The next candidate sample from the set S = s(1) , . . . , s(NM C ) is chosen as the one that


maximizes the probability of misclassification or, in other words, as the one most likely to
have been misclassified as safe/failed by the surrogate limit-state function gb(x):

s∗ = arg min U (s) ≡ arg max Pm (s). (1.60)

s∈S s∈S

Another popular learning function is the expected feasibility function (EFF) (Bichon et al.,

UQL AB-V2.0-107 - 16 -
 Structural reliability (Rare event estimation)

2008):

−µgb(x) − − µgb(x)  − µgb(x)

      
EF F (x) = µgb(x) 2Φ −Φ −Φ
σgb(x) σgb(x) σgb(x)
−µgb(x) − − µgb(x)  − µgb(x)
      
− σgb(x) 2ϕ −ϕ −ϕ
σgb(x) σgb(x) σgb(x)
 − µgb(x) − − µgb(x)
    
+ Φ −Φ , (1.61)
σgb(x) σgb(x)

where  = 2σgb(x) and ϕ is the PDF value of a standard normal Gaussian variable. The next
candidate sample is then chosen by:

s∗ = arg max EF F (s). (1.62)

s∈S

1.6.1.2 Convergence criteria

The convergence criterion terminates the addition of samples to the experimental design of
the Kriging metamodel and thus terminates the improvement in the accuracy of the fail-
ure probability estimate. The standard convergence criterion related to the U -function is
defined as follows (Echard et al., 2011): the iterations stop when mini U (s(i) ) > 2 where
i = 1, . . . , NM C . Schöbi et al. (2016) demonstrated that this criterion is very conservative
and that an alternative stopping criterion, related to the uncertainty in the estimate of the
failure probability itself, is often more efficient in the context of structural reliability. It is
given by the following condition:
Pbf+ − Pbf−
≤ Pbf , (1.63)
Pb0
f

where Pbf = 10% and the three failure probabilities are defined as:

Pbf0 = P µgb(X) ≤ 0 ,


(1.64)

Pbf± = P µgb(X) ∓ kσgb(X) ≤ 0 ,

(1.65)

where k = Φ−1 (1 − α/2) sets the confidence level (1 − α), typically k = Φ−1 (97.5%) = 1.96.
A similar convergence criterion can be defined for the reliability index:

βb+ − βb−
≤ βb, (1.66)
βb0

where the threshold βb = 5% and the three reliability indices correspond to the aforemen-
f
tioned failure probabilities:
βb0 = −Φ−1 (Pbf0 ), (1.67)

βb± = −Φ−1 (Pbf∓ ). (1.68)

UQL AB-V2.0-107 - 17 -
UQL AB user manual

[Quantile space] [Physical space]

Figure 5: Stochastic spectral embedding: Illustration of SSE representation in Eq. (1.69) in

the quantile and physical space. Red points denote the experimental design and orange
areas denote terminal or unsplit
S domains, that partition the input space into a set of disjoint
subdomains such that DX = k∈T DX k .

1.6.1.3 AK-MCS with a PC-Kriging metamodel

As originally proposed by Echard et al. (2011), AK-MCS uses an ordinary Kriging model
for approximating the limit-state function. As demonstrated by Schöbi et al. (2016) replac-
ing the ordinary Kriging metamodel with a Polynomial-Chaos-Kriging (PC-Kriging) one (see
also UQL AB User Manual – PC-Kriging) can significantly improve the convergence of the al-
gorithm. The corresponding reliability methods is called Adaptive PC-Kriging Monte Carlo
Simulation (APCK-MCS) and is available in UQL AB as an advanced option of AK-MCS (see
Section 2.3.6.2).

1.6.2 Stochastic spectral embedding-based reliability

Stochastic spectral embedding-based reliability (SSER, see Wagner et al. (2021)) is an active-
learning reliability method that uses the stochastic spectral embedding (SSE, see Marelli et al.
(2021) and UQL AB User Manual – Stochastic spectral embedding) metamodelling technique
to approximate the limit-state function with
X X
g(X) ≈ gSSE (X) = b k (X) +
1Dk (X) R b k (X),
1Dk (X) R (1.69)
X
S X
S
k∈K\T k∈T

where T is the set of indices denoting the terminal domains, i.e.the unsplit domains among
all the the card(K) domains used in the SSE representation (see Figure ??).
Because the set of terminal domains T constitutes a complete partition of the input domain,
k , the failure probability from Eq. (1.2) can be rewritten as a weighted
S
i.e.DX = k∈T DX

UQL AB-V2.0-107 - 18 -
 Structural reliability (Rare event estimation)

sum Z
X
k
Pf = V Pfk , with k
V = fX (x) dx, (1.70)
k∈T k
DX

where V k is the domain-wise probability mass, and Pfk is a conditional failure probability given
by h i h i
def k
Pfk = P g(X) ≤ 0|X ∈ DX k
≈ P gSSE (X) ≤ 0|X ∈ DX . (1.71)

These conditional failure probabilities can then be estimated with suitable probabilistic sim-
ulation methods utilizing the SSE representation of the limit-state function.
SSER shares many similarities with other active learning reliability methods (see UQL AB User
Manual – Active learning reliability). However, SSER does not strictly fit in the general active-
learning reliability framework because its construction relies on the SSE-specific sequential
partitioning algorithm.

1.6.2.1 Sequential partitioning algorithm

The algorithm to construct SSEs implemented in UQL AB is called sequential partitioning al-
gorithm. It sequentially partitions selected refinement domains and constructs local expan-
sions of the residual to ultimately produce a limit-state approximation of the form shown in
Eq. (1.69). This algorithm is explained in detail in the UQL AB User Manual – Stochastic spec-
tral embedding. In Wagner et al. (2021), modifications to this algorithm were proposed that
turned the SSE construction into an efficient active-learning reliability algorithm. These mod-
ifications pertain to the introduction of bootstrap SSE, reliability-specific refinement domain
selection and sample enrichment strategies as well as a reliability-specific stopping criterion,
and a way to estimate the conditional failure probabilities.

Bootstrap SSE To take into account the point-wise prediction accuracy of SSE, we use the
local error estimation capabilities of bootstrap SSE (Efron, 1979; Marelli and Sudret, 2018;
Wagner et al., 2021). Using B resampled experimental designs, we construct B residual
expansions in every terminal domain to obtain B bootstrap predictions
(b) b k,(b) (X),
X X
gSSE (X) = b k (X) +
1Dk (X) R 1Dk (X) R for b ∈ {1, · · · , B}. (1.72)
X
S X S
k∈K\T k∈T

By analogy with Marelli and Sudret (2018), we can then define an estimate of the point-wise
misclassification probability as:
B
def 1 X  (b)

P̂m (x) = 1Df,SSE (x) − 1Df,SSE (x) , (1.73)

B
b=1

(b)
where 1Df,SSE and 1Df,SSE are evaluated with the mean and bootstrap predictors of SSE re-
spectively.
The bootstrap replications can also be used to obtain statistics of the conditional failure

UQL AB-V2.0-107 - 19 -
UQL AB user manual

probability estimates, such as variance or confidence bounds. As an example, the conditional

failure probability variance is given by
h i hn oi
k,(1) k,(B)
Var P̂fk = Var P̂f , · · · , P̂f , (1.74)

with the failure probability of the b-th replication given by

h i
k,(b) def (b) k
P̂f = P gSSE (X) ≤ 0|X ∈ DX . (1.75)

By means of Eq. (1.70) this conditional failure probability variance can be used to write an
expression for the total failure probability variance as
h i X  2 h i
Var P̂f = V k Var P̂fk . (1.76)
k∈T

Similarly, the bootstrap replications can be used to define confidence bounds on the total
(b) def P k,(b)
failure probability. Let P̂f = k∈T V k P̂f be the total failure probability estimated with
n o
the b-th replication from all terminal domains. The equal tail confidence bounds P̂fk , P̂fk
are then defined by
h n oi
(1) (B)
P P̂fk ≤ P̂f , · · · , P̂f ≈ α,
hn o i (1.77)
(1) (B)
P P̂f , · · · , P̂f ≤ P̂fk ≈ 1 − α

def
with α ∈ [0, 0.5] and γ = 1 − 2α is called the symmetrical confidence level.

Refinement domain selection At every refinement step, the sequential partitioning algo-
rithm chooses a refinement domain based on a refinement score (see Section 1.3.1 in UQL AB
User Manual – Stochastic spectral embedding). For SSER, this refinement score is given by:
( k
V , h i if Eq. (1.79) is met,
Ek = 2 (1.78)
V k Var P̂fk , otherwise,

which depends on the following intermediate re-prioritization criterion

h i
(i−2) (i−1) (i)
Var P̂f , P̂f , P̂f
< εP̂f , (1.79)
(i) 2
 
P̂f

(i)
where P̂f is the total failure probability estimator at the i-th iteration of the algorithm and
the threshold is heuristically chosen to εP̂f = 0.1%. This criterion is triggered when the
failure probability estimator does not change significantly in three successive iterations.

Partitioning strategy The partitioning strategy determines how a selected refinement do-
main is split. SSER separates regions that predict failure/safety correctly from those that do
so incorrectly (see Figure 6). A measure of prediction accuracy in this respect is given by the

UQL AB-V2.0-107 - 20 -
 Structural reliability (Rare event estimation)

misclassification probability Pm (Echard et al., 2011) defined for bootstrap SSE in Eq. (1.73).
To this end, we define two auxiliary conditional random vectors in the quantile space DU as
def def
Z 0 = U |P̂m = 0, and Z + = U |P̂m > 0. (1.80)

These vectors identify regions of zero (Z 0 ) and non-zero (Z + ) misclassification probability

in a given refinement domain.
Based on these random vectors, we choose a set of splitting locations in each dimension
υi ∈ DUi , i = 1, · · · , M that confine a maximum of Z + ’s probability mass to one side of
the split, while confining a maximum of Z 0 ’s probability mass to the other side of the split.
Denoting by Zi+ and Zi0 the marginals of the auxiliary random vectors in the i-th dimension,
we pick the splitting location
( 
P Zi+ ≤ υi + P Zi0 > υi ,
  
def
υ
bi = arg max Li (υi ), with Li (υi ) = −1 + max (1.81)
P Zi+ > υi + P Zi0 ≤ υi .
   
υi ∈DXi

To ultimately choose a splitting direction d ∈ {1, · · · , M }, we compare the values of the

objective functions Li (b
υi ) and split along the dimension that achieves the best split, i.e.

d = arg max Li (b
υi ). (1.82)
i∈{1,··· ,M }

In practice, a sufficiently large sample of the auxiliary random vectors Z 0 and Z + is used to
conduct the computations of this section.
It might occur, at the first step of the algorithm or after the intermediate re-prioritization
criterion Eq. (1.79) has been triggered, that the algorithm does not detect misclassified sam-
ple points. When this happens, the auxiliary random vector Z + is not defined. To let the
algorithm proceed in this case, the definition of the auxiliary random vectors is updated in
Eq. (1.80) where Pm is replaced with
B
def 1 X (b)
Pt (x) = 1Dt (x), (1.83)
B
b=1

(b)
where 1Dt is defined with an empirical quantile t such that P [|gSSE (X)| ≤ t] = εt as
( (b)
(b) def 1, if |gSSE (x)| ≤ t,
1Dt (x) = (b) (1.84)
0, if |gSSE (x)| > t.

The value εt = 0.01 has proven to be a good choice in practice.

Sample enrichment SSER uses a sequentially sampled experimental design (see Sec-
tion 1.3.3 in UQL AB User Manual – Stochastic spectral embedding). To exploit the fact that
the SSE for reliability applications needs to be more accurate near the limit-state surface, we

UQL AB-V2.0-107 - 21 -
UQL AB user manual

Figure 6: Partitioning strategy exemplified on a one dimensional function showing the boot-
strap replications and the resulting support of Z + and Z 0 as well as the objective function L.
The splitting location υ is highlighted as a vertical, dashed red line. The plots are shown in
the quantile space DU .

split the sample budget and randomly place

(
Nenr Ncurr
2 − 2
Nuni = max (1.85)
0

points uniformly in the refinement domain quantile space, where Ncurr is the number of
existing points in the refinement domain. The remainder of the sample budget is used to
randomly place sample points in the subset of the refinement domain with non-zero mis-
classification probability Pm . This corresponds to sampling the auxiliary random vector Z +
defined in Eq. (1.80). We do this to ensure stability of the residual expansions that are known
to deteriorate in accuracy when the experimental design clusters in a small subspace.

Conditional failure probability estimation To estimate the conditional failure probabili-

ties in Eq. (1.70), simulation methods can be used at a low computational cost. We therefore
use the subset simulation method (see Section 1.5.3.3) that is known to effectively estimate
even extremely small failure probabilities (Moustapha et al., 2021).
The process of estimating the conditional failure probabilities is implemented as a post-
expansion operation (see Section 1.3.5 in UQL AB User Manual – Stochastic spectral embed-
ding).

Stopping criterion The algorithm is terminated when the reliability index bounds, also
known as beta bounds, do not change significantly over a set of iterations. The generalized

UQL AB-V2.0-107 - 22 -
 Structural reliability (Rare event estimation)

reliability index can be computed from the failure probability as β = −Φ−1 (Pf ), where Φ−1
is the inverse normal CDF (Ditlevsen, 1979). Using the 95% confidence intervals on β̂ from
the bootstrap replications as defined for the failure probability in Eq. (1.77), the stopping
criterion is given by
β̂ − β̂
< εβ̂ , (1.86)
β̂
where the failure threshold is heuristically set to εβ̂ = 3%. For stability, we abort the algo-
rithm only if this criterion is fulfilled in three consecutive iterations.

UQL AB-V2.0-107 - 23 -
Chapter 2

Usage

In this section, a reference problem will be set up to showcase how each of the techniques in
Chapter 1 can be deployed in UQL AB.

2.1 Reference problem: R-S

The benchmark of choice to showcase the methods described in Section 1.3 is a basic problem
in structural reliability, namely the R-S case. It is one of the simplest possible abstract setting
consisting of only two input state variables: a resistance R and a stress S. The system fails
when the stress is higher than the resistance, leading to the following limit-state function:

X = {R, S} g(X) = R − S; (2.1)

The two-dimensional probabilistic input model consists of independent variables distributed

according to Table 2.

Table 2: Distributions of the input parameters of the R − S model in Eq. (2.1).

Name Distributions Parameters Description

R Gaussian [5, 0.8] Resistance of the system
S Gaussian [2, 0.6] Stress applied to the system

An example UQL AB script that showcases how to deploy all of the algorithms available in the
structural reliability module can be found in the example file:
Examples/Reliability/uq_Examples_Reliability_01_RS.m

2.2 Problem set-up

Solving a structural reliability problem in UQL AB requires the definition of three basic com-
ponents:

• a MODEL object that describes the limit-state function

• an INPUT object that describes the probabilistic model of the random vector X

25
UQL AB user manual

• a reliability ANALYSIS object.

The UQL AB framework is first initialized with the following command:

uqlab

The model in Eq. (2.1) can be added as a MODEL object directly with a M ATLAB vectorized
string as follows:
MOpts.mString = 'X(:,1) - X(:,2)'; % R−S
MOpts.isVectorized = 1;
myModel = uq_createModel(MOpts);

For more details on the available options to create a model object in UQL AB, please refer to
the UQL AB User Manual – the MODEL module.
Correspondingly, an INPUT object with independent Gaussian variables as specified in Table 2
can be created as:
IOpts.Marginals(1).Name = 'R'; % Resistance
IOpts.Marginals(1).Type = 'Gaussian';
IOpts.Marginals(1).Moments = [5, 0.8];
IOpts.Marginals(2).Name = 'S'; % Stress
IOpts.Marginals(2).Type = 'Gaussian';
IOpts.Marginals(2).Moments = [2, 0.6];

myInput = uq_createInput(IOpts);

For more details about the configuration options available for an INPUT object, please refer
to the UQL AB User Manual – the INPUT module.

2.3 Reliability analysis with different methods

This section showcases how all the methods introduced in Section 1.3, can be deployed in
UQL AB. In addition, visualization and advanced options are also described in detail. The
following methods are showcased in this section:

• FORM: Section 2.3.1

• SORM: Section 2.3.2

• Monte Carlo Simulation: Section 2.3.3

• Importance Sampling: Section 2.3.4

• Subset Simulation: Section 2.3.5

• AK-MCS: Section 2.3.6

2.3.1 First Order Reliability Method (FORM)

Running a FORM analysis on the specified UQL AB MODEL and INPUT objects does not require
any specific configuration. The following minimum syntax is required:

UQL AB-V2.0-107 - 26 -
 Structural reliability (Rare event estimation)

FORMOpts.Type = 'Reliability';
FORMOpts.Method = 'FORM';
FORMAnalysis = uq_createAnalysis(FORMOpts);

Once the analysis is performed, a report with the FORM results can be printed on screen by:
uq_print(FORMAnalysis)

which produces the following:

-------------------------------------------
FORM
-------------------------------------------
Pf 1.3499e-03
BetaHL 3.0000
ModelEvaluations 8
-------------------------------------------
Variables R S
Ustar -2.400000 1.800000
Xstar 3.08e+00 3.08e+00
Importance 0.640000 0.360000
-------------------------------------------

The results can be visualized graphically as follows:

uq_display(FORMAnalysis)

which produces the images in Figure 7. Note that the graphical representation of the FORM
iterations (right panel of Figure 7) is only produced for the 2-dimensional case.

FORM - Convergence FORM - Design point, failure plane

3.5
3.5 Iterations
3 FORM limit state surface
3
2.5 2.5
2 2
βHL

u2

1.5 1.5
1 1

0.5 0.5
0
0
1 2 -5 -4 -3 -2 -1 0
number of iterations u1

Figure 7: Graphical visualization of the results of the FORM analysis in Section 2.3.1.

UQL AB-V2.0-107 - 27 -
UQL AB user manual

Note: In the preceding example no specifications are provided. If not further specified
the FORMruns with the following defaults:

– Algorithm to find the design point: 'iHLRF';

– Starting point for the Rackwitz-Fiessler (RW) algorithm: (0, . . . , 0);
– Tolerance value for the RW algorithm on the design point: 10−4 ;
– Tolerance value for the RW algorithm on the limit-state function: 10−4 ;
– Maximum number of iterations for the RW algorithm: 100;
– Failure is defined for: limit-state g(x) ≤ 0.

Since FORM is a gradient-based method, the gradient of the limit-state function

needs to be computed. This is done using finite differences with the following
defaults:

– Type of finite difference scheme: 'forward';

– Value of the difference scheme: 10−3 .

2.3.1.1 Accessing the results

The analysis results can be accessed in the FORMAnalysis.Results structure:

FORMAnalysis.Results
ans =
BetaHL: 3
Pf: 0.0013
ModelEvaluations: 8
Ustar: [-2.4000 1.8000]
Xstar: [3.0800 3.0800]
Importance: [0.6400 0.3600]
Iterations: 2
History: [1x1 struct]

In the Results structure, Pf is the estimate of Pf according to Eq. (1.26), BetaHL the
corresponding Hasofer-Lindt reliability index, Ustar the design point U ∗ in the standard
normal space, Xstar the correspondingly transformed design point in the original space
X ∗ = T −1 (U ∗ ), Importance the importance factors Si in Eq. (1.28), Iterations the num-
ber of FORM iterations needed to converge, ModelEvaluations the total number of eval-
uations of the limit-state function, and History a set of additional information about the
convergence behaviour of the algorithm.

2.3.1.2 Advanced options

Several advanced options are available for the FORM method to tweak which algorithm is
used to calculate the solution. They can be specified by adding a FORM field to the FORM
options structure FORMOpts. In the following, the most common advanced options for FORM
are specified:

UQL AB-V2.0-107 - 28 -
 Structural reliability (Rare event estimation)

• Specify the FORM algorihm: by default, the iHL-RF algorithm is used (see page 8).
The original HL-RF algorithm (see page 7) can be enforced by adding:
FORMOpts.FORM.Algorithm = 'HLRF';

• Specify a starting point: by default the search of the design point is started in the
SNS at U0 = 0. It is possible to specify an alternative starting point (useful, e.g., when
multiple design points are expected) as:
FORMOpts.FORM.StartingPoint = [u1,...,uM];

where [u1,...,uM] are the desired coordinates of the starting point in the SNS.

• Numerical calculation of the gradient: advanced options related to the numerical

calculation of the gradient can be specified by using the FORMOpts.Gradient structure.
As an example, to specify a gradient relative step-size h = 0.001 one can write:
FORMOpts.Gradient.h = 0.001;

Details on the gradient computation options are given in Table 7, page 51.

For a comprehensive list of the advanced options available to the FORM method, please see
Table 6, page 50.

2.3.2 Second Order Reliability Method (SORM)

A SORM analysis is set up very similarly to its FORM counterpart:

SORMOpts.Type = 'Reliability';
SORMOpts.Method = 'SORM';
SORMAnalysis = uq_createAnalysis(SORMOpts);

Once the analysis is performed, a report with the FORM+SORM results can be printed by:
uq_print(SORMAnalysis)

which produces the following:

-------------------------------------------
FORM/SORM
-------------------------------------------
Pf 1.3499e-03
BetaHL 3.0000
PfFORM 1.3499e-03
PfSORM 1.3499e-03
PfSORMBreitung 1.3499e-03
ModelEvaluations 20
-------------------------------------------
Variables R S
Ustar -2.400000 1.800000
Xstar 3.08e+00 3.08e+00
Importance 0.640000 0.360000
-------------------------------------------

UQL AB-V2.0-107 - 29 -
UQL AB user manual

The results can be visualized graphically as follows:

uq_display(SORMAnalysis)

which produces the same images as in FORM (Figure 7), as SORM is only a refinement of
the final FORM Pf estimate.

Note: In the preceding example no specifications are provided. If not further specified
the SORM runs with the same defaults as FORM.

2.3.2.1 Accessing the results

The analysis results can be accessed in the SORMAnalysis.Results structure:

SORMAnalysis.Results
ans =
Pf: 0.0013
BetaHL: 3
ModelEvaluations: 20
PfFORM: 0.0013
PfSORM: 0.0013
PfSORMBreitung: 0.0013
Ustar: [-2.4000 1.8000]
Xstar: [3.0800 3.0800]
Importance: [0.6400 0.3600]
Iterations: 2
History: [1x1 struct]

The Results structure contains the same fields as FORM (see Section 2.3.1) (when necessary
with a FORM or SORM suffix for clarity). In addition, the two PfSORMBreitung and PfSORM
H
fields provide the Pf,SORM and Pf,SORM given in Eqs. (1.34) and (1.35), respectively.

2.3.2.2 Advanced options

The SORM method shares the same advanced options as the FORM method, described in
Section 2.3.1.2.

2.3.3 Monte Carlo Simulation (MCS)

The Monte Carlo simulation (MCS) algorithm only requires the user to specify the maximum
number of limit-state function evaluations, corresponding to N in Eq. (1.36), if different from
the default value N = 105 . As an example, to run a reliability analysis with N = 106 samples
one can write:
MCOpts.Type = 'Reliability';
MCOpts.Method = 'MCS';
MCOpts.Simulation.MaxSampleSize = 1e6;
MCAnalysis = uq_createAnalysis(MCOpts);

Once the analysis is performed, a report with the Monte Carlo simulation results can be
printed on screen by:

UQL AB-V2.0-107 - 30 -
 Structural reliability (Rare event estimation)

uq_print(MCAnalysis)

which produces the following:

-------------------------------------------
Monte Carlo simulation
-------------------------------------------
Pf 1.3110e-03
Beta 3.0089
CoV 0.0276
ModelEvaluations 1000000
PfCI [1.2401e-03 1.381919e-03]
BetaCI [2.9929e+00 3.025751e+00]
-------------------------------------------

The results can be visualized graphically as follows:

uq_display(MCAnalysis)

which produces the convergence plots in Figure 8 and the plot of Monte Carlo sample points
in Figure 9. Note that in uq_display, the maximum number of samples plotted is n = 105
to limit the size of the figure.

Figure 8: Graphical visualization of the convergence of the Monte Carlo simulation analysis
in Section 2.3.3.

Note: In the preceding example only the maximum sample size for the analysis is pro-
vided. If not further specified the Monte Carlo simulation runs with the following
defaults:

– Confidence level: 0.05;

– Number of samples evaluated per batch: 104 ;
– Failure is defined for: limit-state g(x) ≤ 0.

UQL AB-V2.0-107 - 31 -
UQL AB user manual

MCS - Samples
5

x2
2

0 g(X) ≤ 0
g(X) > 0
-1
2 4 6 8
x1

Figure 9: Graphical visualization of the samples of the Monte Carlo simulation analysis in
Section 2.3.3.

2.3.3.1 Accessing the results

The analysis results can be accessed in the MCAnalysis.Results structure:

MCAnalysis.Results
ans =
Pf: 0.0013
Beta: 3.0013
CoV: 0.0273
ModelEvaluations: 1000000
PfCI: [0.0013 0.0014]
BetaCI: [2.9855 3.0180]
History: [1x1 struct]

The Results structure contains the following fields: Pf, the estimated Pbf,MC as in Eq. (1.36);
Beta, the corresponding generalized reliability index in Eq. (1.41); CoV, the coefficient of
variation calculated with Eq. (1.40); ModelEvaluations, the total number of limit-state
function evaluations; PfCI, the confidence intervals calculated with Eq. (1.39); BetaCI, the
corresponding confidence intervals on βHL calculated with Eq. (1.42); History, a structure
containing the convergence of Pf , CoV and the corresponding confidence intervals calcu-
lated at preset sample batches (by default once every 104 samples). The content of the
History structure is used to produce the convergence plot shown in Figure 8.

2.3.3.2 Advanced options

The advanced options available in Monte Carlo simulation are related to the convergence
criterion of the algorithm and to the definition of the confidence bounds reported in the
MCAnalysis.Results structure. In the following, a list of the most commonly used param-
eters for a MC analysis are given:

• Specify a target CoV and a corresponding batch size: in addition to specifying the

UQL AB-V2.0-107 - 32 -
 Structural reliability (Rare event estimation)

MaxSampleSize option, one can specify a target CoV . The algorithm will sequentially
add batches of points to the current sample and stop as soon as the current CoV is be-
low the specify threshold. To specify a target CoV = 0.01 and batches of size NB = 104 ,
one can write:
MCOpts.Simulation.TargetCoV = 0.01;
MCOpts.Simulation.BatchSize = 1e4;

Note that the two options are independent from each other. The BatchSize option is
also used to set the breakpoints for the MCAnalysis.Results.History structure.

• Specify α for the confidence intervals: the α in Eq. (1.39) can also be specified. To
set α = 0.1, one can write:
MCOpts.Simulation.Alpha = 0.1;

For a comprehensive list of the advanced options available for Monte Carlo simulation, please
refer to Table 5, page 49.

2.3.4 Importance Sampling

Importance sampling shares configuration options from both the FORM and the Monte Carlo
simulation methods. A basic IS analysis can be setup with the following code:
ISOpts.Type = 'Reliability';
ISOpts.Method = 'IS';
ISAnalysis = uq_createAnalysis(ISOpts);

Using this minimal setup the analysis will run FORM first with the default options as in Sec-
tion 2.3.1 to determine the design point U ∗ . Then the standard normal importance density
centred at the obtained design point is used. Sampling is carried out with the following
default options: N = 1000, batch size NB = 100.
Once the analysis is performed, a report with the importance sampling results can be printed
on screen by:
uq_print(ISAnalysis)

which produces the following:

-------------------------------------------
Importance Sampling
-------------------------------------------
Pf 1.3132e-03
Beta 3.0084
CoV 0.0599
ModelEvaluations 1008
PfCI [1.1591e-03 1.467421e-03]
BetaCI [2.9745e+00 3.046121e+00]
-------------------------------------------

The results can be visualized graphically as follows:

UQL AB-V2.0-107 - 33 -
UQL AB user manual

uq_display(ISAnalysis)

which produces the convergence image in Figure 10. As with FORM, the second panel in
Figure 10 is produced only in the 2D case.

IS - FORM Design point and failure plane

5 FORM iterations
FORM limit state surface
4 g(X) ≤ 0
g(X) > 0
3

u2
2

-1
-6 -4 -2 0
u1

Figure 10: Graphical visualization of the convergence of the importance sampling analysis in
Section 2.3.4.

Note: In the preceding example no specifications are provided. The Importance Sam-
pling shares the same defaults values as FORM and MCS. The exceptions are:

– Maximum number of evaluated samples: 103 ;

– Number of samples evaluated per batch: 102 .

2.3.4.1 Accessing the results

The results of the importance sampling analysis can be accessed with the ISAnalysis.Results
structure:
ISAnalysis.Results
ans =
Pf: 0.0013
Beta: 3.0095
CoV: 0.0593
ModelEvaluations: 100008
PfCI: [0.0012 0.0015]
BetaCI: [2.9759 3.0468]
History: [1x1 struct]
FORM: [1x1 struct]

The basic structure of Results closely resembles that of Monte Carlo simulation (see Sec-
tion 2.3.3.1). However, an additional structure Results.FORM is available:
ISAnalysis.Results.FORM
ans =
BetaHL: 3

UQL AB-V2.0-107 - 34 -
 Structural reliability (Rare event estimation)

Pf: 0.0013
ModelEvaluations: 8
Ustar: [-2.4000 1.8000]
Xstar: [3.0800 3.0800]
Importance: [0.6400 0.3600]
Iterations: 2
History: [1x1 struct]

This structure is identical to the FORM results given in Section 2.3.1.1.

2.3.4.2 Advanced options

The importance sampling algorithm accepts all of the options specific to both FORM and
Monte Carlo simulation, described in Section 2.3.1.2 and Section 2.3.3.2. In addition, two
additional options can be specified for importance sampling:

• Specify existing FORM results: by default, importance sampling first runs FORM to
determine the design point, followed by sampling around this design point to calculate
Pf,IS . If the results of a previous FORM analysis are already available, they can be
specified so as to avoid running FORM again. If the results are stored in a FORMResults
structure with the same format as described in Section 2.3.1.1, one can write:
ISOpts.IS.FORM = FORMResults;

Alternatively, one can also directly specify a pre-existing UQL AB FORM or SORM anal-
ysis, say FORMAnalysis:
ISOpts.IS.FORM = FORMAnalysis;

• Specify a custom sampling distribution: alternatively, one can directly specify a cus-
tom sampling distribution. This can be achieved by providing the marginals and copula
structure of the desired distribution, e.g.IOpts in Section 2.2, as follows:
ISOpts.IS.Instrumental = IOpts;

Alternatively, a pre-existent UQL AB INPUT object, say myISInput, can also be specified:
ISOpts.IS.Instrumental = myISInput;

In case the model has multiple outputs Nout , it might be desirable to specify a custom
sampling distribution for each one of them. This can be done by either providing the
IOpts as a 1 × Nout structure or the pre-existing inputs myISInputs as a 1 × Nout
uq_input object. Please note, that custom distributions should be specified either for
all outputs or for none.

For a complete overview of the available options specific to the importance sampling algo-
rithm, see Table 8.

UQL AB-V2.0-107 - 35 -
UQL AB user manual

2.3.5 Subset Simulation

The subset simulation algorithm can be used with the default options P0 = 0.1 and NS = 103
by specifying:
SSOpts.Type = 'Reliability';
SSOpts.Method = 'Subset';
SSimAnalysis = uq_createAnalysis(SSOpts);

Once the analysis is performed, a report with the subset-simulation results can be printed on
screen by:
uq_print(SSimAnalysis)

which produces the following:

-------------------------------------------
Subset simulation
-------------------------------------------
Pf 1.0600e-03
Beta 3.0729
CoV 0.2433
ModelEvaluations 2680
PfCI [5.5463e-04 1.565373e-03]
BetaCI [2.9546e+00 3.261242e+00]
-------------------------------------------

The results can be visualized graphically as follows:

uq_display(SSimAnalysis)

which illustrates the samples of each subset in Figure 11 (applicable only for one and two-
dimensional problems).

SubsetSim - Samples in each subset

5

3
x2

0
0 2 4 6 8
x1
Figure 11: Graphical visualization of the convergence of the subset simulation analysis in
Section 2.3.5.

UQL AB-V2.0-107 - 36 -
 Structural reliability (Rare event estimation)

Note: In the preceding example no specifications are provided. The Importance Sam-
pling shares the same defaults values as FORM and MCS. Additionally, there are
the following default values:

– Target conditional failure probability of auxiliary limit-staates: 0.1;

– Maximum number of subsets: 20;
– Type of the proposal distribution in the Markov Chain: 'uniform';
– Parameter (standard deviation / halfwidth) of the proposal distribution: 1.

2.3.5.1 Accessing the results

The results of subset simulation are stored in the SSimAnalysis.Results structure:

SSimAnalysis.Results
ans =
Pf: 0.0016
Beta: 2.9402
CoV: 0.2426
ModelEvaluations: 2597
PfCI: [8.4859e-04 0.0024]
BetaCI: [2.8160 3.1387]
NumberSubsets: 3
History: [1x1 struct]

The fields in the Results structure have the same meaning as their counterparts in impor-
tance sampling and Monte Carlo simulation. Further, the field NumberSubsets denotes the
number of subsets. Note that the ModelEvaluations field does not contain exactly the ex-
pected N = NS ∗ m ∗ (1 − P0 ) = 27000 limit-state function evaluations, but a slightly smaller
N = 26095. This discrepancy is due to the modified Metropolis-Hastings MCMC acceptance
criterion described in Au and Beck (2001), which in some uncommon cases can reject sam-
ples without the need of evaluating the limit-state function.

2.3.5.2 Advanced options

Subset simulation uses the same advanced options as Monte Carlo simulation described in
Section 2.3.3.2, as well as some additional options. The most important are summarized in
the following:

• Specify P0 : the value of P0 in Eq. (1.53) can be specified in 0 < P0 ≤ 0.5. One can set
e.g. P0 = 0.2 as follows:
SSOpts.SubsetSim.p0 = 0.2;

UQL AB-V2.0-107 - 37 -
UQL AB user manual

• Specify the number of samples in each subset: the number of samples in each subset
NS can be specified by using the .Simulation.BatchSize field. To set it to NS = 1000
one can write:
SSOpts.Simulation.BatchSize = 1000;

For a comprehensive overview of the available options specific to subset simulation see Ta-
ble 9, page 52.

2.3.6 Adaptive Kriging Monte Carlo Simulation (AK-MCS)

Adaptive Kriging Monte Carlo simulation method with default values (see Table 11 and the
related linked tables for details on the defaults) can be deployed in UQL AB with the following
code:
AKOpts.Type = 'Reliability';
AKOpts.Method = 'AKMCS';
AKAnalysis = uq_createAnalysis(AKOpts);

Once the analysis is complete, a report with the AK-MCS results can be printed on screen by:
uq_print(AKAnalysis)

which produces the following:

-------------------------------------------
AK-MCS
-------------------------------------------
Pf 1.5200e-03
Beta 2.9637
CoV 0.0810
ModelEvaluations 18
PfCI [1.2785e-03 1.761457e-03]
BetaCI [3.0165e+00 2.917994e+00]
PfMinus/Plus [1.5200e-03 1.520000e-03]
-------------------------------------------

The results can be visualized graphically as follows:

uq_display(AKAnalysis)

which produces the images in Figure 12. Note that the plot on the right of Figure 12 is
only available when the input is two-dimensional. Additionally, if the verbosity is set to
.Display ≥ 5 the AK-MCS analysis will plot the convergence of Pf and β while the analysis
is running.

UQL AB-V2.0-107 - 38 -
 Structural reliability (Rare event estimation)

AK-MCS - Experimental design

4

3.5

2.5

1.5
2 3 4 5 6

Figure 12: Graphical visualization of the convergence of the AK-MCS analysis in Sec-
tion 2.3.6.

Note: In the preceding example no specifications are provided. If not further specified
the Monte Carlo simulation runs with the following defaults:

– Confidence level: 0.05;

– Maximum number of evaluated samples: 105 ;
– Number of samples evaluated per batch: 104 ;
– Failure is defined for: limit-state g(x) ≤ 0;
– Type of metamodel: 'Kriging';
– Learning function to determine the best next sample(s): 'U';
– Convergence criterion for the adaptive ED algorithm: 'stopU';
– Number of samples added to the ED for the metamodel: 103 ;
– Number of samples in the initial ED: Nini = max(10, 2M )
– Initial ED sampling strategy: 'LHS'.

2.3.6.1 Accessing the results

The results from the AK-MCS algorithm are stored in the AKAnalysis.Results structure:
AKAnalysis.Results
ans =
Pf: 0.0015
Beta: 2.9637
CoV: 0.0810
ModelEvaluations: 25
PfCI: [0.0013 0.0018]
BetaCI: [2.9180 3.0165]
Kriging: [1x1 uq_model]
History: [1x1 struct]

UQL AB-V2.0-107 - 39 -
UQL AB user manual

The fields in the Results structure have the same meaning as their counterparts in Monte
Carlo simulation. Further, the field Kriging contains the final Kriging metamodel used to
estimate the failure probability. This metamodel can be reused within UQL AB for any other
purpose, see UQL AB User Manual – Kriging (Gaussian process modelling) for details.

2.3.6.2 Advanced options

AK-MCS uses the same advanced options as Monte Carlo simulation described in Sec-
tion 2.3.3.2, as well as some additional options. The most important are summarized in
the following:

• Learning function: The learning function can be set to a custom function handle,
which points to an existing M ATLAB function file. For instance, for a learning function
called EFF, a function called uq_LF_EFF must be available in the M ATLAB path. It is
then used in AK-MCS as:
AKOpts.AKMCS.LearningFunction = 'EFF';

• Convergence criterion: There are three different convergence criteria mentioned in

Section 1.6.1.2. They are all available and can be specified e.g.criterion on failure
probability:
AKOpts.AKMCS.Convergence = 'stopPf';

• Specify the Kriging metamodel: The specifications for the Kriging metamodel (see
also UQL AB User Manual – Kriging (Gaussian process modelling)) can be set in the
field .AKMCS.Kriging, e.g.for an ordinary Kriging model:
AKOpts.AKMCS.Kriging.Trend = 'ordinary';

• Specify the initial experimental design: Apart from specifying a number of points
and a sampling strategy, the initial experimental design can be specified by providing
X = x(1) , . . . , x(N0 ) in the matrix X and the corresponding limit-state function values


g(x(1) , . . . , x(N0 ) in G:


AKOpts.AKMCS.IExpDesign.X = X;
AKOpts.AKMCS.IExpDesign.G = G;

• Specify the number of added experimental design points: The maximum number of
samples added to the experimental design of the Kriging metamodel can be specified
to e.g.100:
AKOpts.AKMCS.MaxAddedED = 100;

Note that the total number of runs of the limit-state function then is at most the initial
ED size plus the above number.

UQL AB-V2.0-107 - 40 -
 Structural reliability (Rare event estimation)

• Use a PC-Kriging metamodel: Instead of Kriging, a PC-Kriging model (see also UQL AB
User Manual – PC-Kriging) can be used as a surrogate model in AK-MCS.
AKOpts.AKMCS.MetaModel = 'PCK';

Specific options of the PCK model can be added in the field .AKMCS.PCK. As an exam-
ple, a Gaussian correlation function in PC-Kriging is set as:
AKOpts.AKMCS.PCK.Kriging.Corr.Family = 'Gaussian';

For an overview of the advanced options available for the AK-MCS method, refer to Table 11,
page 53.

2.3.7 Stochastic spectral embedding-based reliability (SSER)

The stochastic spectral embedding-based reliability method with default values (see Table 13
and the related linked tables for details on the defaults) can be deployed in UQL AB with the
following code:
SSEROpts.Type = 'Reliability';
SSEROpts.Method = 'SSER';
SSERAnalysis = uq_createAnalysis(SSEROpts);

Once the analysis is complete, a report with the AK-MCS results can be printed on screen by:
uq_print(SSERAnalysis)

which produces the following:

---------------------------------------------------
Stochastic spectral embedding-based reliability
---------------------------------------------------
Pf 1.3686e-03
Beta 2.9958
CoV 0.0009
ModelEvaluations 80
PfCI [1.3686e-03 1.3712e-03]
BetaCI [2.9952e+00 2.9958e+00]
---------------------------------------------------

The results can be visualized graphically as follows:

uq_display(SSERAnalysis)

which produces the images in Figure 13. Note that the plot on the right of Figure 13 is only
available when the input is two-dimensional.

UQL AB-V2.0-107 - 41 -
UQL AB user manual

Figure 13: Graphical visualization of the convergence of the SSER analysis in Section 2.3.7.

Note: In the preceding example no specifications are provided. If not further specified
the SSER analysis runs with the following defaults:

– Number of enrichment samples Nenr : 10;

– Maximum number of evaluated samples: 105 ;
– Failure is defined for: limit-state g(x) ≤ 0;
– Partitioning strategy: implemented in uq_SSE_partitioning_misclass
(see Section 1.6.2.1)
– Refinement score: implemented in uq_SSE_refineScore_Pf (see Sec-
tion 1.6.2.1)
– Post-expansion operation to estimate the conditional failure probability: im-
plemented in uq_SSE_postExpansion_computePf (see Section 1.6.2.1)
– Stopping criterion: implemented in uq_SSE_stopping_Beta (see Sec-
tion 1.6.2.1)

2.3.7.1 Accessing the results

The results from the SSER algorithm are stored in the SSERAnalysis.Results structure:
SSERAnalysis.Results

ans =

Pf: 0.0015
Beta: 2.9637
CoV: 0.0810
ModelEvaluations: 25
PfCI: [0.0013 0.0018]
BetaCI: [2.9180 3.0165]
Kriging: [1x1 uq_model]
History: [1x1 struct]

The fields in the Results structure have the same meaning as their counterparts in Monte

UQL AB-V2.0-107 - 42 -
 Structural reliability (Rare event estimation)

Carlo simulation. Further, the field SSER contains the final SSE metamodel used to estimate
the failure probability. This metamodel can be reused within UQL AB for any other purpose,
see UQL AB User Manual – Stochastic spectral embedding for details.
For an overview of the advanced options available for the SSER method, refer to Table 13,
page 54.

2.4 Advanced limit-state function options

2.4.1 Specify failure threshold and failure criterion

While it is normally good practice to define the limit-state function directly as a UQL AB
MODEL object as in Section 2.2, in some cases it can be useful to be able to create one from
small modifications of existing MODEL objects. A typical scenario where this is apparent is
when the same objective function needs to be tested against a set of different failure thresh-
olds, e.g. for a parametric study. In this case, the limit-state specifications can be modified.
As an example, when g(x) ≤ T = 5 defines the failure criterion, one can use the following
syntax:
MCOpts.LimitState.Threshold = 5;
MCOpts.LimitState.CompOp = '<=';

UQL AB offers several possibilities to create simple (or arbitrarily complex) objective functions
from existing MODEL objects (see also UQL AB User Manual – the INPUT module).
For an overview of the advanced options for the limit-state function, refer to Table 4, page
49.

2.4.2 Vector Output

In case the limit-state function g(X) results in a vector rather than a scalar value, the struc-
tural reliability module estimates the failure probability for each component independently.

Note: There is no system-type reasoning implemented to combine the failure probabil-

ities of each component.

However, the implemented methods make use of evaluations of the limit-state function if
available, as follows:

• Monte Carlo simulation: The enrichment of the sample size is increased until the
convergence criteria are fulfilled for all components.

• Subset Simulation: The first batch of samples (MCS) is reused for every output com-
ponent limit-state.

• AK-MCS: The initial experimental design for the Kriging model of output component i
consists of the final experimental design of component i − 1.

UQL AB-V2.0-107 - 43 -
UQL AB user manual

2.5 Excluding parameters from the analysis

In various usage scenarios (e.g. parametric studies) one or more input variables may be set
to fixed constant values. This can have important consequences for many of the methods
available in UQL AB e.g. FORM/SORM and AK-MCS, whose costs increase significantly with
the number of input variables. Whenever applicable, UQL AB will appropriately account for
the set of constant input parameters and exclude them from the analysis so as to avoid
unnecessary costs. This process is transparent to the user as the analysis results will still
show the excluded variables, but they will not be included in the calculations.
To set a parameter to constant, the following command can be used when the probabilistic
input is defined (See UQL AB User Manual – the INPUT module):
inputOpts.Marginals.Type = 'Constant' ;
inputOpts.Marginals.Parameters = value;

Furthermore, when the standard deviation of a parameter equals zero, UQL AB treats it as a
Constant. For example, the following uniformly distributed variable whose upper and lower
bounds are identical is automatically set to a constant with value 1:
inputOpts.Marginals.Type = 'Uniform' ;
inputOpts.Marginals.Parameters = [1 1];

UQL AB-V2.0-107 - 44 -
Chapter 3

Reference List

How to read the reference list

Structures play an important role throughout the UQL AB syntax. They offer a natural way
to semantically group configuration options and output quantities. Due to the complexity of
the algorithms implemented, it is not uncommon to employ nested structures to fine-tune
the inputs and outputs. Throughout this reference guide, a table-based description of the
configuration structures is adopted.
The simplest case is given when a field of the structure is a simple value or array of values:

Table X: Input
.Name String A description of the field is put here

which corresponds to the following syntax:

Input.Name = 'My Input';

The columns, from left to right, correspond to the name, the data type and a brief description
of each field. At the beginning of each row a symbol is given to inform as to whether the
corresponding field is mandatory, optional, mutually exclusive, etc. The comprehensive list
of symbols is given in the following table:

Mandatory
 Optional
⊕ Mandatory, mutually exclusive (only one of
the fields can be set)

Optional, mutually exclusive (one of them
can be set, if at least one of the group is set,
otherwise none is necessary)

When one of the fields of a structure is a nested structure, a link to a table that describes the
available options is provided, as in the case of the Options field in the following example:

45
UQL AB user manual

Table X: Input
.Name String Description
 .Options Table Y Description of the Options
structure

Table Y: Input.Options
.Field1 String Description of Field1
 .Field2 Double Description of Field2

In some cases, an option value gives the possibility to define further options related to that
value. The general syntax would be:
Input.Option1 = 'VALUE1' ;
Input.VALUE1.Val1Opt1 = ...;
Input.VALUE1.Val1Opt2 = ...;

This is illustrated as follows:

Table X: Input
.Option1 String Short description
'VALUE1' Description of 'VALUE1'
'VALUE2' Description of 'VALUE2'

.VALUE1 Table Y Options for 'VALUE1'

.VALUE2 Table Z Options for 'VALUE2'

Table Y: Input.VALUE1
 .Val1Opt1 String Description
 .Val1Opt2 Double Description

Table Z: Input.VALUE2
 .Val2Opt1 String Description
 .Val2Opt2 Double Description

Note: In the sequel, double and doubles mean a real number represented in double
precision and a set of such real numbers, respectively.

UQL AB-V2.0-107 - 46 -
 Structural reliability (Rare event estimation)

3.1 Create a reliability analysis

Syntax

myAnalysis = uq_createAnalysis(ROpts)

Input

All the parameters required to determine the analysis are to be given as fields of the struc-
ture ROpts. Each method has its own options, that will be reviewed in different tables. The
options described in Table 3 are common to all methods.

Table 3: ROpts
.Type 'uq_reliability' Identifier of the module. The options
corresponding to other types are in
the corresponding guides.
.Method String Type of structural reliability method.
The available options are listed
below:
'MCS' Monte Carlo simulation.

'FORM', First order reliability method.

'SORM', Second order reliability method.

'IS' Importance sampling.

'Subset' Subset simulation.
'AKMCS' Adaptive Kriging Monte Carlo
Simulation (AK-MCS).
'SSER' Stochastic spectral embedding-based
reliability (SSER).
 .Name String Name of the module. If not set by
the user, a unique string is
automatically assigned to it.

 .Input INPUT object INPUT object used in the analysis. If

not specified, the currently selected
one is used.

 .Model MODEL object MODEL object used in the analysis. If

not specified, the currently selected
one is used.

UQL AB-V2.0-107 - 47 -
UQL AB user manual

 .LimitState See Table 4 Specification of the limit-state

function.

 .Display String Level of information displayed by the

default: 'standard' methods.
'quiet' Minimum display level, displays
nothing or very few information.
'standard' Default display level, shows the most
important information.
'verbose' Maximum display level, shows all the
information on runtime, like updates
on iterations, etc.

 .Simulation See Table 5 Options field for the simulation

methods. Only applies when
ROpts.Method is 'MCS', 'IS',
'SS', or 'AKMCS'.

 .FORM See Table 6 Options field for the FORM

algorithm methods. Only applies
when ROpts.Method is 'FORM',
'SORM', or 'IS'.

 .Gradient See Table 7 Options field for computing the

gradient. It applies to the methods
that use FORM, namely, when
ROpts.Method is 'FORM',
'SORM', or 'IS'.

 .IS See Table 8 Options field for importance

sampling. It applies only when
ROpts.Method is 'IS'.

 .Subset See Table 9 Options field for subset simulation. It

applies only when ROpts.Method is
'Subset'.
 .AKMCS See Table 11 Options field for the adaptive
experimental design algorithm in
AK-MCS. This applies when
ROpts.Method is 'AKMCS'.

 .SSER See Table 13 Options field for the SSER algorithm.

This applies when ROpts.Method is
'SSER'.

 .SaveEvaluations Logical Storage or not of performed

default: true evaluations of the limit-state
function.

UQL AB-V2.0-107 - 48 -
 Structural reliability (Rare event estimation)

true Store the evaluations.

false Do not store the evaluations.

In order to perform a structural reliability analysis, the limit-state function g(x) is compared
to a threshold value T (by default T = 0). In analogy with Eq. (1.1), failure is defined as
g(x) ≤ T . Alternatively, failure can be specified as g(x) ≥ T by adjustment of the field
ROpts.LimitState.CompOp to '>='. The relevant options are summarized in Table 4:

Table 4: ROpts.LimitState
 .Threshold Double Threshold T , compared to the
default: 0 limit-state function g(x).

 .CompOp String Comparison operator for the

default: '<=' limit-state function.

'<', '<=' Failure is defined by g(x) < T .

'>', '>=' Failure is defined by g(x) > T .

The available methods to perform structural reliability analysis are Monte Carlo simulation,
importance sampling, subset simulation, AK-MCS, FORM, and SORM. The first four methods
share the simulation options. FORM and SORM are gradient-based, so they allow the user to
specify the finite difference options as well as the algorithm options.
In Table 5, the options for the simulation methods (Monte Carlo, importance sampling, subset
simulation and AK-MCS) are shown:

Table 5: ROpts.Simulation
 .Alpha Double Confidence level α. For the Monte
default: 0.05 Carlo estimators, a confidence
interval is constructed with
confidence level 1 − α.

UQL AB-V2.0-107 - 49 -
UQL AB user manual

 .MaxSampleSize Integer Maximum number of samples to be

default: 103 for 'IS'; evaluated. If there is no target
105 otherwise coefficient of variation (CoV), this is
the total number of samples to be
evaluated. If the target CoV is
present, the method will run until
TargetCoV or MaxSampleSize is
reached. In this case, the default
value of MaxSampleSize, if not
specified in the options, is Inf,
i.e.the method will run until the
target CoV is achieved.

 .TargetCoV Double Target coefficient of variation. If

present, the method will run until
the estimate of the CoV (Eq. (1.40))
is below TargetCoV or until
MaxSampleSize function
evaluations are performed. The
value of the coefficient of variation of
the estimator is checked after each
BatchSize evaluations. By default
this option is disabled. Note: this
option has no effect in
Method = 'Subset' and
'AKMCS'.

 .BatchSize Integer Number of samples that will be

default: 104 for 'MCS' evaluated at once. Note that this
and 'AKMCS'; option has no effect in
103 for 'Subset'; Method = 'AKMCS'.
102 for 'IS'

Note: In order to use importance sampling after an already computed FORM analysis,
one can provide these results to the analysis options in order to avoid repeating
FORM. If FORMResults is a structure containing the results of a FORM analysis,
the syntax reads:
ISOpts.Type = 'Reliability';
ISOpts.Method = 'IS';
ISOpts.FORM = FORMResults;
ISAnalysis = uq_createAnalysis(ISOpts);

The FORM algorithm has special parameters that can be tuned in the subfield FORM of the
options. These parameters also affect the methods that depend on FORM, namely importance
sampling and SORM. These are listed in Table 6.

Table 6: ROpts.FORM
 .Algorithm String Algorithm used to find the design
default: 'iHLRF' point.

UQL AB-V2.0-107 - 50 -
 Structural reliability (Rare event estimation)

'iHLRF' Improved HLRF.

'HLRF' HLRF.

 .StartingPoint 1 × M Double Starting point for the

default: zeros(1,M) Rackwitz-Fiessler algorithm.

 .StopU Double Tolerance value for the

default: 10−4 Rackwitz-Fiessler algorithm on the
design point. The algorithm will stop
when |Uk+1 − Uk | < StopU .

 .StopG Double Tolerance value for the

default: 10−6 Rackwitz-Fiessler algorithm on the
limit-state function value. The
algorithm will stop when
| G(Uk)
G(U0 ) | < StopG.

 .MaxIterations Integer Maximum number of iterations

default: 100 allowed in the Rackwitz-Fiessler
algorithm. If this property should be
ignored, it can be set to Inf.

Since FORM is a gradient-based method, the gradient of the limit-state function needs to be
computed. This is done using finite differences. The options for the differentiation are listed
in Table 7.

Table 7: ROpts.Gradient
 .h Double Value of the difference for the
default: 10−3 scheme.

 .Method String Specifies the type of finite differences

default: 'forward' scheme to be used.

∂g
'forward' Forward finite differences. ∂x i
is
approximated using g(x) and
g(x + hei ).

∂g
'backward' ∂xi is approximated using g(x) and
g(x − hei ).

∂g
'centered' ∂xiis approximated using g(x + hei )
and g(x − hei ). (More accurate and
more costly.)

UQL AB-V2.0-107 - 51 -
UQL AB user manual

The options specifically set for the importance sampling are presented in Table 8. Note that
the options of .Simulation and .FORM are also processed in the case of importance sampling
due to the nature of the MCS, FORM and IS.

Table 8: ROpts.IS
 .Instrumental 1 × Nout INPUT object or Instrumental distribution defined as
Struct either a structure of input marginals
and copula or an INPUT object (refer
to UQL AB User Manual – the INPUT
module for details).

 .FORM FORM ANALYSIS object or FORM results computed previously.

FORMAnalysis.Results See Section 2.3.4.2 for details.
Struct

The options specifically set for subset simulation are presented in Table 9. Note that the
options of .Simulation are also processed in the case of subset simulation due to the similar
nature of Monte Carlo simulation and subset simulation.

Table 9: ROpts.Subset
 .p0 Double Target conditional failure probability
default: 0.1 of auxiliary limit-states
(0 <p0≤ 0.5).

 .Proposal See Table 10 Description of the proposal

distribution in the Markov Chain.

 .MaxSubsets Integer Maximum number of subsets. In the

default: subset simulation algorithm, the
MaxSampleSize maximum number of subsets is set to
BatchSize·(1−p0)
the minimum of MaxSubsets and
MaxSampleSize
BatchSize·(1−p0) .

The settings of the Markov Chain Monte Carlo simulation in subset simulation are summa-
rized in Table 10. Note that the default values are taken from Au and Beck (2001).

Table 10: ROpts.Subset.Proposal (Proposal distributions)

 .Type String Type of proposal distribution (in the
default: 'Uniform' standard normal space).

'Gaussian' Gaussian distribution.

'Uniform' Uniform distribution.

UQL AB-V2.0-107 - 52 -
 Structural reliability (Rare event estimation)

 .Parameters Double Parameter of the proposal

default: 1 distribution. Corresponds to the
standard deviation for a Gaussian
distribution and the half-width for
the uniform distribution.

AK-MCS is a combination of Kriging metamodels and Monte Carlo simulation. The options
for the Kriging metamodel and the adaptive experimental design algorithm are listed here.

Table 11: ROpts.AKMCS

 .MetaModel String Choice of metamodel in AK-MCS.
default: 'Kriging'

.Kriging Structure Kriging options when field
MetaModel = 'Kriging'. If
none is set, then the default Kriging
options are used (refer to UQL AB
User Manual – Kriging (Gaussian
process modelling)). Note that a
small nugget of 10−10 is added by
default to the correlation options to
improve numerical stability.

.PCK Structure PC-Kriging options when field

MetaModel = 'PCK'. If none is
set, then the default PC-Kriging
options are used (refer to UQL AB
User Manual – PC-Kriging). Note
that a small nugget of 10−10 is added
by default to the correlation options
to improve numerical stability.

 String Learning function to determine the

.LearningFunction default: 'U' best next sample(s) to be added to
the experimental design.

'U' U -function (see Eq. (1.59)).

'EFF' Expected feasibility function (see

Eq. (1.61)).

 .Convergence String Convergence criterion for the

default: 'stopU' adaptive experimental design
algorithm.

'stopU' Convergence when min U (x) ≥ 2

(see Echard et al. (2011)) on the
candidate set.

UQL AB-V2.0-107 - 53 -
UQL AB user manual

'stopPf' Convergence criterion based on the

convergence of the failure
probability estimate (see Eq. (1.63)).
'stopBeta' Convergence criterion based on the
convergence of the reliability index
estimate (see Eq. (1.66)).
 .MaxAddedED Integer Number of samples added to the
default: 1000 experimental design of the Kriging
metamodel.

 .IExpDesign See Table 12 Specification of the initial

experimental design of the
metamodel.

The initial experimental design of AK-MCS can either be given by a number of samples and
a sampling method or by a matrix containing the set of input samples and the corresponding
values of the limit-state function.

Table 12: ROpts.AKMCS.IExpDesign

 .N Integer Number of samples in the initial
default: 10 experimental design.

 .Sampling String Sampling techniques of the initial

default: 'LHS' experimental design. See UQL AB
User Manual – the INPUT modulefor
more sampling techniques.

 .X N × M Double Matrix containing the initial

experimental design.

 .G N × Nout Double Vector containing the responses of

the limit-state function
corresponding to the initial
experimental design, corrected by
the threshold k (see also Table 4):
(g(x) − T ) for Criterion = '<=',
(T − g(x)) for Criterion = '>='.

SSER is an active-learning reliability method that uses a highly customized sequential parti-
tioning algorithm to construct an SSE of the limit state function. The options for the SSE and
the sequential partitioning algorithm are listed here.

Table 13: ROpts.SSER

 .Partitioning Function handle Partitioning strategy according to
default: Table 8 of UQL AB User Manual –
uq_[...]_misclass Stochastic spectral embedding.

UQL AB-V2.0-107 - 54 -
 Structural reliability (Rare event estimation)

 .Refine Structure Refinement options according to

default: Table 11 of UQL AB User Manual –
.Score=uq_[...]_Pf Stochastic spectral embedding.
 .PostExpansion Function handle Post-expansion operation according
default: to Table 7 of UQL AB User Manual –
uq_[...]_computePf Stochastic spectral embedding.
 .Stopping Structure Stopping criterion according to
default: Table 12 of UQL AB User Manual –
Stochastic spectral embedding.
.Criterion=...
uq_[...]_Beta
 .ExpDesign Structure Properties of the experimental design
default: according to Table 18 of the UQL AB
User Manual – Stochastic spectral
embedding.

.Type = 'Sequential'
.Enrichment=...
uq_[...]_LSS
.NEnrich=10
.NSamples=2e6
 .ExpOptions Structure Properties of the residual expansions
default: according to UQL AB User Manual –
.Boots.Repl = 200 Polynomial Chaos Expansions.
 .[...] Additional options can be selected
according to Table 1 of the UQL AB
User Manual – Stochastic spectral
embedding.

UQL AB-V2.0-107 - 55 -
UQL AB user manual

3.2 Accessing the results

Syntax

myAnalysis = uq_createAnalysis(ROpts)

Output

The information stored in the myAnalysis.Results structure depends on which kind of

analysis is performed. In the sequel, the results for each of the methods are reviewed.

• Monte Carlo - Table 14

• FORM - Table 16

• SORM - Table 18

• Importance sampling - Table 19

• Subset simulation - Table 20

• AK-MCS - Table 22

• SSER - Table 24

3.2.1 Monte Carlo

The results are summarized in Table 14.

Table 14: myAnalysis.Results

.Pf Double Estimator of the failure probability, Pf,MCS .

.Beta Double Associated reliability index,

βMCS = −Φ−1 (Pf,MCS ).

.CoV Double Coefficient of variation.

.ModelEvaluations Integer Total number of model evaluations performed

during the analysis.

.PfCI 1 × 2 Double Confidence interval of the failure probability.

.BetaCI 1 × 2 Double Confidence interval of the associated reliability

index.

.History See Table 15 If the simulation is carried out using batches of

points, History(i) contains results obtained
after the i-th batch.

UQL AB-V2.0-107 - 56 -
 Structural reliability (Rare event estimation)

If the simulation has been carried out by using various batches of points, the information on
the convergence in each step is stored in the structure History. Its contents are described in
Table 15.

Table 15: myAnalysis.Results.History

.Pf Double Failure probability estimate after each batch.
.CoV Double Coefficient of variation after each batch.

.Conf Double Confidence interval after each batch.

.X N × M Double Matrix containing the input vectors in the

original space evaluated by the limit-state
function.

.U N × M Double Matrix containing the input vectors in the

standard normal space evaluated by the
limit-state function.

.G N × Nout Values of the limit-state function along the

Double FORM iterations i.e.
g(xk ) − T for Criterion = '<=',
T − g(xk ) for Criterion = '>='.

3.2.2 FORM and SORM

FORM and SORM methods are very close in terms of calculations. Indeed, SORM can be
understood as a correction of the FORM estimation of the probability. Therefore, the results
structures are very similar. The results of FORM are shown in Table 16. When executing
SORM, some fields will be added to the FORM Results structure, shown in Table 18.

Table 16: myAnalysis.Results

.BetaHL Double Hasofer-Lind reliability index, βHL .

.Pf Double Estimator of the failure probability,

Pf,FORM = Φ(−βHL ).

.ModelEvaluations Integer Total number of model evaluations performed

to solve the analysis.

.Ustar 1 × M Double Design point U ∗ in the standard normal space.

.Xstar 1 × M Double Design point X ∗ the original space.
.Importance 1 × M Double Importance factors Si .

UQL AB-V2.0-107 - 57 -
UQL AB user manual

.Iterations Integer Number of iterations carried out by the

optimization algorithm.

.Method String Method used to solve the analysis.

.History See Table 17 Structure with information about the algorithm

steps and runtime information.

The field History of the results contains more detailed information extracted from the algo-
rithm steps.

Table 17: myAnalysis.Results.History

.ExitFlag String Reason why the algorithm stopped.

.BetaHL Double Values of the reliability index along the FORM

iterations.

.OriginValue Double Value of the limit-state function G(u = 0).

.G N × Nout Values of the limit-state function along the

Double FORM iterations i.e.
g(xk ) − T for Criterion = '<=',
T − g(xk ) for Criterion = '>='.

.Gradient Double Values of the gradient of the limit-state

function along the FORM iterations.

.StopU Double Values of the stopping criterion on the design

point convergence, |Uk+1 − Uk |, along the
FORM iterations.

.StopG Double Values of the stopping criteria on the limit-state

function, | G(Uk)
G(U0 ) |, along the FORM iterations.

.U N × M Double Coordinates of the points Uk in the standard

normal space.

.X N × M Double Coordinates of the points Xk in the original

space.

If SORM is also performed, two fields are added to the Results, and two fields are added to
Results.History, as shown in Table 18.

UQL AB-V2.0-107 - 58 -
 Structural reliability (Rare event estimation)

Table 18: myAnalysis.Results

.PfSORM Double SORM estimator of the failure probability,
Pf,SORM , using Hohenbichler’s formula.

.PfSORMBreitung Double SORM estimator of the failure probability,

Pf,SORM , using Breitung’s formula.

.BetaSORM Double Associated reliability index

βSORM = −Φ−1 (Pf,SORM ), using Hohenbichler’s
formula.

.BetaSORMBreitung Double Associated reliability index

βSORM = −Φ−1 (Pf,SORM ), using Breitung’s
formula.

.History.FORMEvals Integer Number of model evaluations carried out to

perform the FORM analysis.

.History.Hessian M × M Double Hessian matrix of the limit-state function at the

design point, U ∗ .

3.2.3 Importance sampling

Since importance sampling is a simulation method, the structure of the results is similar to
the one of Monte Carlo simulation. The results are listed in Table 19.

Table 19: myAnalysis.Results

.Pf Double Estimator of the failure probability, Pf,IS .
.Beta Double Associated reliability index, βIS = −Φ−1 (Pf,IS ).
.CoV Double Coefficient of variation.

.ModelEvaluations Integer Total number of model evaluations performed

during the analysis.
.PfCI 1 × 2 Double Confidence interval of the failure probability.

.BetaCI 1 × 2 Double Confidence interval of the associated reliability

index.

.FORM See Table 16 Results of the FORM analysis used to find the
design point.

UQL AB-V2.0-107 - 59 -
UQL AB user manual

.History See Table 15 If importance sampling is carried out using

batches of points, History(i) contains
results obtained after the i-th batch.

3.2.4 Subset simulation

Since subset simulation is a simulation method, the structure of the results is similar to the
one of Monte Carlo simulation. The results are listed in Table 20 and Table 21.

Table 20: myAnalysis.Results

.Pf Double Estimator of the failure probability, Pf,SS .

.Beta Double Associated reliability index, βSS = −Φ−1 (Pf,SS ).

.CoV Double Coefficient of variation.

.ModelEvaluations Integer Number of model evaluations during the

analysis.

.PfCI 1 × 2 Double Confidence interval of the failure probability.

.BetaCI 1 × 2 Double Confidence interval of the associated reliability

index.

.NumberSubsets Integer Number of auxiliary subsets during the

analysis.
.History See Table 21 Data related to each subset.

The field History of the results contains more detailed information extracted from the algo-
rithm steps.

Table 21: myAnalysis.Results.History

.delta2 Double δj2 for estimating the coefficient of variation
(see Eq. (1.55)).

.q Double Intermediate limit-state thresholds.

.X Cell Samples in the input space of each subset.

.U Cell Samples in the standard normal space of each
subset.

UQL AB-V2.0-107 - 60 -
 Structural reliability (Rare event estimation)

.G Cell Values of the limit-state function of each

sample in each subset corrected by the
threshold T (see also Table 4):
(g(x) − T ) for Criterion = '<=',
(T − g(x)) for Criterion = '>='.

.Pfcond Double Conditional failure probability estimates

P (Di+1 |Di ).

.gamma Double γj for computing the coefficient of variation

(see Eq. (1.56)).

3.2.5 AK-MCS

Since AK-MCS relies upon Monte Carlo simulation, the structure of the results is similar to
the one of Monte Carlo simulation. The results are listed in Table 22 and Table 23.

Table 22: myAnalysis.Results

.Pf Double Estimator of the failure probability, Pf,AK-MCS .

.Beta Double Associated reliability index,

βAK-MCS = −Φ−1 (Pf,AK-MCS ).

.CoV Double Coefficient of variation.

.ModelEvaluations Integer Total number of model evaluations performed
during the analysis.
.PfCI 1 × 2 Double Confidence interval of the failure probability.

.BetaCI 1 × 2 Double Confidence interval of the associated reliability

index.

.Kriging Struct Final Kriging metamodel.

.PCK Struct Final PC-Kriging metamodel.

.History See Table 23 Contains intermediate results along the

AK-MCS iterations.

UQL AB-V2.0-107 - 61 -
UQL AB user manual

Table 23: myAnalysis.Results.History

.Pf Double History of the estimate failure probability.
.PfLower Double History of the estimated lower bound of the
failure probability Pf− .
.PfUpper Double History of the estimated upper bound of the
failure probability Pf+ .
.NSamples Integer Number of samples added to the experimental
design at each iteration.
.NInit Integer Number of samples in the initial experimental
design.
.X Cell Samples in the input space of each subset.
.G Cell Values of the limit-state function of each
sample in each subset corrected by the
threshold T (see also Table 4):
(g(x) − T ) for Criterion = '<=',
(T − g(x)) for Criterion = '>='.

3.2.6 SSER

The results of an SSER are the constructed SSE and the failure probability estimate. The
results are listed in Table 24 and Table 25.

Table 24: myAnalysis.Results

.Pf Double Estimator of the failure probability, Pf,SSER .

.Beta Double Associated reliability index,

βSSER = −Φ−1 (Pf,SSER ).

.CoV Double Coefficient of variation.

.ModelEvaluations Integer Total number of model evaluations performed
during the analysis.
.PfCI 1 × 2 Double Confidence interval of the failure probability.

.BetaCI 1 × 2 Double Confidence interval of the associated reliability

index.

.SSER Struct Final SSE metamodel of the limit-state

function.

.History See Table 25 Contains intermediate results along the SSER

iterations.

UQL AB-V2.0-107 - 62 -
 Structural reliability (Rare event estimation)

Table 25: myAnalysis.Results.History

.Pf Double History of the estimate failure probability and
upper and lower bounds.
.Beta Double History of the associated reliability index and
upper and lower bounds.
.CoV Double History of the coefficient of variation.
.X Cell Samples in the input space that were added at
a certain iteration.
.G Cell Associated values of the limit-state function.

3.3 Printing/Visualizing of the results

UQL AB offers two commands to conveniently print reports containing contextually relevant
information for a given result object:

3.3.1 Printing the results: uq print

Syntax

uq_print(myAnalysis);
uq_print(myAnalysis, outidx);

Description

uq_print(myAnalysis) prints a report on the results of the reliability analysis stored in

the object myAnalysis. If the model has multiple outputs, only the results for the first
output variable are printed.

uq_print(myAnalysis, outidx) prints a report on the results of the reliability analysis

stored in the object myAnalysis for the output variables specified in the array outidx.

Examples:
uq_print(myAnalysis, [1 3]) prints the reliability analysis results for output variables 1
and 3.

3.3.2 Graphically display the results: uq display

Syntax
uq_display(myAnalysis);
uq_display(myAnalysis, outidx);

UQL AB-V2.0-107 - 63 -
UQL AB user manual

Description

uq_display(myAnalysis) creates a visualization of the results of the reliability analysis

stored in the object myAnalysis, if possible. If the model has multiple outputs, only
the results for the first output variable are visualized.

uq_display(myAnalysis, outidx) creates a visualization of the results of the reliability

analysis stored in the object myAnalysis for the output variables specified in the array
outidx.

Examples:
uq_display(myAnalysis, [1 3]) will display the reliability analysis results for output
variables 1 and 3.

UQL AB-V2.0-107 - 64 -
References

Au, S. K. and J. L. Beck (2001). Estimation of small failure probabilities in high dimensions
by subset simulation. Probabilistic Engineering Mechanics 16(4), 263–277. 13, 14, 37, 52

Bichon, B. J., M. S. Eldred, L. Swiler, S. Mahadevan, and J. McFarland (2008). Efficient global
reliability analysis for nonlinear implicit performance functions. AIAA Journal 46(10),
2459–2468. 16

Breitung, K. (1989). Asymptotic approximations for probability integrals. Probabilistic Engi-

neering Mechanics 4(4), 187–190. 9, 10

Cai, G. Q. and I. Elishakoff (1994). Refined second-order reliability analysis. Struct.

Saf. 14(4), 267–276. 9, 10

Dani, V., T. P. Hayes, and S. M. Kakade (2008). Stochastic linear optimization under bandit
feedback. In The 21st Annual Conference on Learning Theory (COLT 2008). 16

Ditlevsen, O. (1979, jan). Generalized second moment reliability index. Journal of Structural
Mechanics 7(4), 435–451. 23

Ditlevsen, O. and H. Madsen (1996). Structural reliability methods. J. Wiley and Sons,
Chichester. 3

Dubourg, V. (2011). Adaptive surrogate models for reliability analysis and reliability-based
design optimization. Ph. D. thesis, Université Blaise Pascal, Clermont-Ferrand, France. 16

Echard, B., N. Gayton, and M. Lemaire (2011). AK-MCS: an active learning reliability method
combining Kriging and Monte Carlo simulation. Struct. Saf. 33(2), 145–154. 15, 16, 17,
18, 21, 53

Efron, B. (1979). Bootstrap methods: another look at the Jackknife. Annals of Statistics 7(1),
1–26. 19

Ginsbourger, D., B. Rosspopoff, G. Pirot, N. Durrande, and R. Renard (2013). Distance-

based Kriging relying on proxy simulation for inverse conditioning. Advances in Water
Resources 52, 275–291. 16

Hasofer, A.-M. and N.-C. Lind (1974). Exact and invariant second moment code format. J.
Eng. Mech. 100(1), 111–121. 6

65
UQL AB user manual

Hohenbichler, M., S. Gollwitzer, W. Kruse, and R. Rackwitz (1987). New light on first- and
second order reliability methods. Struct. Saf. 4, 267–284. 10

Lemaire, M. (2009). Structural reliability. Wiley. 3

Marelli, S. and B. Sudret (2018). An active-learning algorithm that combines sparse poly-
nomial chaos expansions and bootstrap for structural reliability analysis. Struct. Saf. 75,
67–74. 19

Marelli, S., P.-R. Wagner, C. Lataniotis, and B. Sudret (2021). Stochastic spectral embedding.
11(2), 25–47. 18

Melchers, R.-E. (1999). Structural reliability analysis and prediction. John Wiley & Sons. 3,
12

Moustapha, M., S. Marelli, and B. Sudret (2021). A generalized framework for active learning
reliability: survey and benchmark. Reliability Engineering & System Safety. (submitted). 22

Papaioannou, I., W. Betz, K. Zwirglmaier, and D. Straub (2015). MCMC algorithms for subset
simulation. Probabilistic Engineering Mechanics 41, 89–103. 14

Rackwitz, R. and B. Fiessler (1978). Structural reliability under combined load sequences.
Computers & Structures 9, 489–494. 7

Rubinstein, R.-Y. (1981). Simulation and the Monte Carlo methods. John Wiley & Sons. 11

Schöbi, R., B. Sudret, and S. Marelli (2016). Rare event estimation using Polynomial-Chaos-
Kriging. ASCE-ASME Journal of Risk and Uncertainty in Engineering Systems, Part A: Civil
Engineering, D4016002. 15, 17, 18

Srinivas, N., A. Krause, S. Kakade, and M. Seeger (2012). Information-theoretic regret

bounds for Gaussian process optimization in the bandit setting. IEEE Transactions on Infor-
mation Theory 58(5), 3250–3265. 16

Sudret, B. (2007). Uncertainty propagation and sensitivity analysis in mechanical models -

Contributions to structural reliability and stochastic spectral methods. Habilitation thesis,
Université Blaise Pascal, Clermont-Ferrand, France. 1

Tvedt, L. (1990). Distribution of quadratic forms in normal space – Applications to structural

reliability. J. Eng. Mech. 116(6), 1183–1197. 10

Wagner, P.-R., S. Marelli, I. Papaioannou, D. Straub, and B. Sudret (2021, June). Rare event
estimation using stochastic spectral embedding. 18, 19

Zhang, Y. and A. Der Kiureghian (1995). Two improved algorithms for reliability analysis.
In R. Rackwitz, G. Augusti, and A. Bori (Eds.), Proc. 6th IFIP WG7.5 on Reliability and
Optimization of Structural systems, Assisi, Italy. Chapman & Hall, London. 8

UQL AB-V2.0-107 - 66 -