Professional Documents
Culture Documents
Zwirglmaier 2016
Zwirglmaier 2016
art ic l e i nf o a b s t r a c t
Article history: Discrete Bayesian networks (BNs) can be effective for risk- and reliability assessments, in which prob-
Received 26 October 2015 ability estimates of (rare) failure events are frequently updated with new information. To solve such
Received in revised form reliability problems accurately in BNs, the discretization of continuous random variables must be per-
7 April 2016
formed carefully. To this end, we develop an efficient discretization scheme, which is based on finding an
Accepted 16 April 2016
Available online 23 April 2016
optimal discretization for the linear approximation of the reliability problem obtained from the First-
Order Reliability Method (FORM). Because the probability estimate should be accurate under all possible
Keywords: future information scenarios, the discretization scheme is optimized with respected to the expected
Bayesian networks posterior error. To simplify application of the method, we establish parametric formulations for efficient
Discretization
discretization of random variables in BNs for reliability problems based on numerical investigations. The
Near-real-time
procedure is implemented into a software prototype. Finally, it is applied to a verification example and an
Structural reliability
Updating application example, the prediction of runway overrun of a landing aircraft.
& 2016 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.ress.2016.04.008
0951-8320/& 2016 Elsevier Ltd. All rights reserved.
K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109 97
Fig. 1. Design point and linear approximation of the limit state surface. Left side: original random variable space; right side: standard normal space (from [27].
98 K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109
The FORM approximation of PrðFÞ is obtained by substituting M i or influencing variables I j . Whenever new evidence on these
the LSF in U-space GðUÞ by a linear function GL ðUÞ, i.e. a first-order variables is available, the BN should be evaluated (in near-real time)
Taylor expansion of GðUÞ. The key idea of FORM is to choose as the utilizing exact BN inference algorithms.
expansion point the so-called design point u , which is the point To enable exact inference algorithms, all continuous random
that minimizes ku k subject to GL ðUÞ r 0. u also known as the variables are discretized. These include the X, and possibly the M i
most likely failure point, as it is the point in the failure domain and I j . In the general case, the computational effort for solving
with the highest probability density. Since all marginal distribu- the BN is a direct function of the CPT size of ‘Component per-
tions of the standard uncorrelated multinormal distribution are formance’. The size of this CPT is 2∏ni¼ 1 ni , where n is the number
standard normal, it can be shown that the FORM probability of of random variables in X, and ni is the number of states used for
failure Pr½GL ðUÞ r 0 is discretizing X i . In this paper we do not describe the discretization
of random variables M i and I j , since it is typically straightforward
Pr½GL ðUÞ r 0 ¼ Φ β FORM ð2Þ
and does not contribute significantly to computational perfor-
where Φ is the standard normal CDF and βFORM is the distance mance. The key parameter for computational efficiency and
from the origin to the design point, i.e. βFORM ¼ ku k. The problem accuracy is the discretization scheme for X, which is described in
thus reduces to finding the design point u . If GðUÞ is linear, the Sections 2.5 and 2.6.
FORM solution of the probability of failure is exact, otherwise it is an
approximation, which however is sufficiently accurate in most
2.4. Simplification of BNs through node removal
practical applications with limited numbers of random variables [22].
The linearized LSF GL ðUÞ can be written as
Removing random variables from a BN is one possibility to
GL ðUÞ ¼ βFORM αT U ð3Þ reduce the computational effort associated with a model. A formal
approach for removing nodes from a BN is described in [30]. In
where α ¼ ½α1 ; …; αn is the vector of FORM importance mea-
order to decide which nodes to remove from the BN the following
sures. These importance measures are defined as
questions should be considered:
ui
αi ¼ ð4Þ
βFORM Which random variables are relevant for prediction? (These
where ui
is the ith component of the design point coordinates. include ‘Component performance’.)
The αi ’s take values between 1 and 1, and it is kαk ¼ 1. αi is 0, if Which random variables can potentially be observed? (These
the uncertainty on U i has no influence on PrðGL ðUÞ r 0Þ, and it is include the measurement variables.)
Which random variables simplify the modeling of dependen-
1 or 1, if U i is the only random variable affecting Pr g L ðUÞ r 0 .
When the original random variables X i are mutually independent, cies? (These are e.g. common influencing factors such as I 2 in
the αi ’s are readily applicable also in the original space, otherwise Fig. 2.)
the αi ’s can be transformed as described in [3].
For which random variables is it desirable to explicitly show
their influence on component performance?
2.3. Treatment of a reliability problem in a BN
If a random variable does not belong to any of these categories,
We combine discrete BNs and structural reliability concepts to the corresponding node in the BN can be removed. Since the
facilitate updating of rare event (failure) probabilities under new computational efficiency of the model is governed by the size of
observations. The general problem setting is illustrated in the BN of the CPT of the 'Component state’ node, the primary interest is in
Fig. 2. We here limit the presentation to component reliability pro- removing basic random variables X i from the BN. As a measure for
blems; system problems are considered later. The binary random the relevance of a basic random variable, importance measures αi
variable ‘Component performance’ is described by the LSF gðXÞ. from a FORM analysis may be used. To better understand the
The basic random variables X of the model are included in the relation between αi and X i ’s relevance for prediction, consider a
BN as parents of ‘Component performance’. The nodes M i repre- linearized LSF GL ðUÞ. Following [3], the variance of GL ðUÞ can be
sent measurements of individual random variables X i , and nodes I j decomposed as
represent factors influencing the basic random variables. Depen-
σ 2GL ¼ ∇G2 α21 þ α22 þ ⋯ þ α2n ð5Þ
dence between the variables in X is modeled either directly by
links among them (here X 2 -X 1 and X 4 -X 5 ) or through common where ∇G denotes the gradient vector of the non-linearized LSF
influencing factors (here I 2 -X 3 and I 2 -X 4 ). The component GðUÞ. From Eq. (5) it is seen that a random variable X i with cor-
performance node can have (multiple) child nodes, which, how- responding αi accounts for α2i U100% of the variance σ 2GL . Therefore,
ever, does not impact the discretization of the reliability problem. observing a random variable X i with αi ¼ 0:1 will reduce the var-
Ultimately, the goal is to predict the component performance, i.e. iance σ 2GL by 1%, whereas observing X j with αj ¼ 0:5 will reduce σ 2GL
PrðFÞ, conditional on other model parameters, such as measurements by 25%.
2
In a BN, basic random variables X are independent if they are not connected
through links, if they have no common (unknown) ancestors and if no evidence is
Fig. 2. A general BN including a component reliability problem. available on any of their joint descendants.
K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109 99
2.5.1. Independent basic random variables the computation of the conditional failure probability following
The situation is illustrated in Fig. 3. The performance of the Eq. (8) is based on the prior probability PrðF jY 1 ¼ 3Þ, the dis-
component depends on n statistically independent random variables cretization introduces an approximation in this case. The error
and is described by a LSF g ðXÞ ¼ gðX 1 ; …; X n Þ. For all basic random occurs only in the intervals that are cut by the limit state surface.
variables X i , corresponding measurements M i can be performed. To In the simple one-dimensional case of Fig. 4, an optimal dis-
obtain an equivalent discrete BN, the continuous X i are replaced by cretization approach would be to discretize the whole outcome
the discrete random variables Y i , and the LSF is replaced by the CPT space in two intervals, one capturing the survival and one the
of component performance conditional on Y ¼ ½Y 1 ;…;Y n . For each failure domain. This discretization would have zero approximation
discrete random variable Y i with ni states 1; 2; …; ni , we define a error. However, already in a two-dimensional case, such a solution
discretization scheme Di ¼ d0 ; d1 ; …; dni 1 ; dni consisting of ni þ 1 is not possible. This is illustrated in Fig. 5, where the cells cut by
interval boundaries. The first and the last interval boundaries are the limit state surface are indicated in gray. The failure probability
given by the boundaries of X i 0 s outcome space. conditional on measurements calculated according to Eq. (8) will
Since here the X i , and thus the Y i , have no parents, the PMF of necessarily be an approximation. The approximation error will be
Y i is defined as small, if the contribution of the cells cut by the limit state surface
(the gray cells in Fig. 5) to the total failure probability is small. An
pY i ðjÞ ¼ F X i dj F X i dj 1 ; with jϵ½1; …; ni ð6Þ
efficient discretization will thus limit this contribution with as few
where F X i denotes the cumulative distribution function (CDF) of intervals as possible.
X i . The probability of failure corresponding to the discrete BN in
Fig. 3b can be calculated as 2.5.2. Dependent basic random variables
X
n1 X
nn
Eqs. (6)–(8) must be adjusted when dependence among the X i ’s
PrðF Þ ¼ … pY 1 y1 …pYn yn U Pr F Y 1 ¼ y1 \ … \ Y n ¼ yn is present, in accordance with the case-specific BN structure.
y1 ¼ 1 yn ¼ 1 However, the principles outlined above for independent X 1 ; …; X n
ð7Þ hold equally for dependent basic random variables: The dis-
Note that the discretization does not introduce any approx- cretization error is a function of the cells cut by the limit state
function.
imation error here, as long as the conditional Pr F Y 1 ¼ y1 \ … \
Y n ¼ yn Þ is computed exactly. When determining an optimal discretization, we propose in the
Once measurements from the nodes M ¼ ½M 1 ;…;M n are avail- following to find the FORM approximation of the reliability
able, the conditional failure probability can be calculated as
1 Xn1 Xnn pY 1 ðy1 ÞU pM1 jY 1 ðm1 jy1 Þ … pY n ðyn ÞU pMn jY n ðmn jyn ÞU
PrðFjM ¼ mÞ … continuous
pM ðmÞ y ¼ 1 y ¼ 1 PrðFjY 1 ¼ y1 \ … \ Y n ¼ yn Þ
1 n
ð8Þ
where Pr F Y 1 ¼ y1 ; …; Y n ¼ yn is the conditional probability of
component failure given y1 ; …; yn . If no measurements are avail-
able for some of the basic random variables, the corresponding
likelihood terms p
are simply omitted in Eq. (8).
M j Y j mj yj
Fig. 3. Representation of a basic reliability problem with n independent basic random variables in a BN. Left: original problem with continuous basic random variables X i ,
right: discrete BN, in which X i 's are substituted with discrete nodes Y i .
100 K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109
h
i
~ :
expected preposterior error EM~ e d; M
h
i Z
opt
d ¼ arg minEM~ e d; M ~ ~ f M~ ðmÞd
¼ arg min e d; m ~ m ~ ð11Þ
d d ~
M
2.6. Efficient discretization These parameters are illustrated in Fig. 6. For a problem with n
basic random variables, the full set of optimization parameters is
2.6.1. Optimal discretization of linear problems in standard normal d ¼ ½w1 ; …; wn ; n1 ; …; nn 1 ; v1 ; …; vn .
space Clearly, the discretization error reduces with increasing ni .
To find an efficient discretization of X, we consider the FORM Because the computational efficiency of the final BN is a direct
solution to the reliability problem. Evaluating the linearized FORM function of the size of the CPT associated with component per-
LSF GL ðUÞ is computationally inexpensive once the design point u formance, which is ∏ni¼ 1 ni , we constrain its size. To this end,
is known. Therefore, it is feasible to find a discretization of U that
we define cup as the maximum allowed number of parameters of
is optimal for the event GL ðUÞ r 0 through optimization. If GðUÞ the CPT of the component state node. This puts a constraint on the
is not strongly non-linear, this solution will be an efficient dis-
optimization of Eq. (11):
cretization for GðUÞr0 and, after a transformation to the ori- n
ginal space, also for g ðXÞ r 0 . ∏ ni r cup ð12Þ
As discussed in Section 2.5.1, the approximation error of the i¼1
discretization is associated with the change from the prior to the The optimization is implemented through a two-level
posterior distribution of the basic random variables. A measure of approach. The optimization of the continuous parameters width
optimality must thus consider possible measurements of X or U. wi and position of the discretization frame vi for all i ¼ 1; …; n is
We consider hypothetical measurements M ~ i (this notation is used
carried out using unconstrained nonlinear optimization for fixed
to distinguish hypothetical measurements from actual measure-
ments M i ¼ mi ) of all standard normal random variables U i with
additive measurement error ϵi N 0; σ ϵi . Since both the prior
distribution and the measurement error are normal distributed,
the likelihood is also normal distributed:
M~ i fU i ¼ ui g N ui ; σ ε ð9Þ
i
values of ni . The optimization of the discrete ni is performed the optimal discretization frame coincides with the design point,
through a local search algorithm. Note that the optimization is i.e. vopt
i ¼ 0. The optimal discretization widths wopt i vary sig-
performed offline, i.e. prior to running the BN, hence it does not nificantly with the importance measures αi , as shown in Section
affect the goal of near-real time performance of the BN. Further- 3.1.2. However, the optimal number of intervals nopti is approxi-
more, in Section 3 a heuristic is derived that can replace the time- mately the same for all random variables in all investigated cases,
independent of the αi values, i.e. nopt
1=n
consuming solution of the optimization problem. i ¼ cup .
2.6.2. Efficient discretization of the original random variables X 3.1.1. On why the optimal number of intervals nopt
i is independent of
Since the nodes in the BN represent random variables X in their αi
original outcome space, the discretization schemes, which are To better understand why the number of intervals does not
derived for the corresponding standard normal random variables depend on jαi j (for jαi j that are significantly larger than 0), recall
U, need to be transformed to the X-space. In the case of mutually that an efficient discretization scheme should focus on the area
independent random variables X i , any point on the ith interval around the limit state surface. More precisely, the discretization
boundary in U-space – if transformed – will result in the same error in the posterior case is induced by the cells that are cut by
corresponding ith interval boundary in X-space. This is not the the limit state surface. Exemplarily, Fig. 8 shows a linear problem
case for dependent random variables X i , where a mapping of in standard normal space with two basic random variables U 1 and
the interval boundaries in U-space to X-space will not lead to an U 2 , where α1 ¼ 0:45 and α2 ¼ 0:89. While Fig. 8a shows an optimal
orthogonal discretization scheme in X-space. To preserve ortho- discretization with 5 intervals per dimension, Fig. 8b shows a
gonality throughout the transformation, we propose to represent discretization scheme, where the more important random variable
each interval boundary through a characteristic point and deter- U 2 is discretized with 6 intervals and U 1 with 4 intervals. In both
cases, the discretization frame is centered at the design point. It is
mine the boundary in X-space through a transformation of this
observed that the probability mass of the (gray) cells, associated
point. For transforming the interval boundary of X i , the char-
with the discretization error in the posterior case, is higher in
acteristic point is selected as the design point u , where the ith
Fig. 8b than for the optimal discretion scheme in Fig. 8a. In the
element is substituted by the coordinate of the interval boundary.
example shown here it is 1:4∙10 3 compared to 2:5∙10 4 .
In Fig. 7 this is shown for an example with n ¼ 2 random variables.
For input random variables X i with a value of jαi j close to zero,
the above observations do not hold. Following Section 2.4, these
variables should be removed from the BN prior to discretization.
3. Development of an efficient discretization procedure
3.1.2. Dependence of the optimal discretization width on αi
3.1. Optimization of the FORM approximation In the optimization, it is found that the optimal discretization
width wopt
i varies strongly with the random variable’s importance,
We present the optimal discretization for the FORM approx- expressed through αi . It is reminded that the width wi describes
imation GL ðUÞ for n ¼ 2 and 3 dimensions. Extension to higher the domain in which a fine discretization mesh is applied (Fig. 6).
numbers of random variables is discussed. Because the linear LSF In general, wopt
i decreases with increasing αi . This effect can be
employed in FORM is described only by the reliability index β FORM observed in Fig. 8a, where U 2 is the more important input random
and the vector α of FORM sensitives (Eq. (3)), it facilitates para- variable and it is wopt opt
2 ow1 .
metric studies. A clear relation between wopti and αi can be observed by plot-
Initially, we consider a reliability index β FORM ¼ 4:26, corre- ting the probability mass enclosed by wopt against αi , as shown in
i
sponding to a probability of failure of 10 5 . The standard deviation Fig. 9. The results of Fig. 9 indicate that the probability mass
of the additive measurement error is set to either σ ε ¼ 0:5 or contained within this interval should be a direct function of αi . The
σ ε ¼ 1:0. Different combinations of FORM sensitivity values αi are more important the variable, the finer the discretization around
selected, to investigate their effect on the optimal discretization. In the design point should become. The observed relationship
all investigated cases, we find that the position of the midpoint of between this probability mass and αi follows a clear trend, and a
Fig. 7. Transformation of a discretization scheme from U-space to X-space. To preserve orthogonality each interval boundary in U-space is represented by a characteristic
point. The random variables X 1 and X 2 are Weibull distributed with scale and shape parameter 1 and their correlation is 0.5.
102 K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109
Fig. 8. Linear problem in standard normal space, with two random variables U 1 and U 2 , where α1 ¼ 0:45 and 0.89. The intervals cut by the limit state surface, i.e. those which
potentially lead to a posterior discretization error are marked in gray.
function can be fitted (Fig. 9). Neither the dimensionality of the As evident from Figs. 10 and 11, there is a clear dependence of
problem nor the standard deviation of the measurement error the probability mass enclosed by the optimal discretization frame
appear to have an influence on this relation. However, as shown in (with width wi ) on the FORM sensitivity values jαi j. The following
the following section, it is found that the relation does depend on parametric function captures this dependence:
the prior failure probability of the problem (i.e. on β FORM ) and on log ðΦðubi Þ Φðlbi ÞÞ ¼ a U expðb U jαi jÞ ð13Þ
the number of intervals ni used to discretize the domain.
To facilitate the application in practice and extending the where ubi is the upper and lbi the lower interval boundary in
results to larger numbers of random variables, in Section 3.2 dimension i, such that wi ¼ ubi lbi . a and b are the parameters of
parametric functions are fitted to the optimization results to the exponential function. This function is depicted in Figs. 10 and 11.
capture the dependency between the optimal discretization width Table 1 shows the parameter values a and b for the different
i and the FORM importance measures αi .
wopt combinations of the prior reliability index β and number of inter-
vals per dimension
3.1.3. Dependence of the optimal discretization on the reliability From the left sides of Figs. 10 and 11, it can be observed that the
index β and the number of discretization cells cup relation between α2i and the optimal wi is fairly diffuse for random
The influence of the prior failure probability and the maximum variables with jαi j o 0:6. Here, the parametric relationship of Eq.
size of the CPT, cup , on the optimal discretization is investigated (13) is less accurate. However, these random variables by defini-
through 10 problems with n ¼ 2 random variables in standard tion have lower importance on the reliability estimate. Hence, the
normal space. The FORM importance measures of the random inaccuracy of Eq. (13) for random variables with jαi j o 0:6 is not
variables are selected between 0:1 and 0:995 and the standard critical, as is confirmed by the numerical investigations performed
deviation of the measurement error is fixed to σ ε ¼ 1:0. We find in the remainder of the paper.
that the optimal discretization frame is generally centered at the The parameter values of Table 1 are derived from two-
design point, i.e. vopt
i ¼ 0, and that the intervals are distributed dimensional problems. In Fig. 9 it is shown that there are no
uniformly among the dimensions. notable differences between two and three dimensions. On this
Firstly, we vary the maximum CPT size cup , i.e. the total number basis, it is hypothesized that the heuristics are applicable also to
of discretization cells. The reliability index is β FORM ¼ 5:2. Fig. 10 problems with higher dimensions. This assumption is furthermore
shows the influence of cup on the resulting width of the dis- supported by the verification examples to be presented in Section 4,
cretization frame wi . Three cases are considered: cup ¼ 25, 100 and where the heuristics are applied also to four-dimensional problems
400. These choices correspond to 5, 10 and 20 intervals for each without any notable deterioration in the results.
K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109 103
Fig. 10. Optimization results for 10 two-dimensional, linear problems in standard normal space, which are discretized with 5, 10 and 20 intervals per dimension. In all cases
the prior failure probability is 10 7 (β ¼ 5:2). The crosses represent the optimization results. The solid lines are the fitted parametric functions (Eq. (13)). The left-hand side
shows the relation between the width of a discretization frame wi and α2i and the right-hand side shows the relation between the probability mass enclosed by the
discretization frame with width wi and α2i .
3.3. Summary of the proposed procedure c. the width of the discretization frame follows Eq. (13)
6. Transform the discretization scheme to X-space
The steps of the proposed procedure are: 7. Compute the CPTs of the component state node and the basic
random variables using Monte Carlo simulation or Latin
1. Formulate the reliability problem hypercube sampling
2. Set up the corresponding BN
3. Perform a FORM analysis for the reliability problem A MATLAB-based software tool performing these steps is
4. Simplify the BN by removing nodes based on: available for download under www.era.bgu.tum.de/software.
a. their importance for prediction
b. their observability
c. whether or not a node simplifies modeling of dependencies
d. whether or not it is desired to explicitly show a node in the 4. Applications
BN for communication purposes
5. Find the discretization scheme in U-space based on the pro- 4.1. Verification example I
posed heuristics i.e.:
a. the discretization scheme is centered at the design point from For verification purposes, we apply the proposed methodology to
the FORM analysis the discretization of a general limit state with non-normal depen-
b. the same number of intervals is used for each random dent random variables. The approximation error made by this dis-
variable cretization is investigated for different measurement outcomes.
104 K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109
Fig. 11. Optimization results for 10 two-dimensional, linear problems in standard normal space, which are discretized with 10 intervals per dimension. The prior failure
probabilities are 10 3 (β ¼ 3:1), 10 5 (β ¼ 4:3) and 10 7 (β ¼ 5:2). The crosses represent the optimization results. The solid lines are the fitted parametric functions (Eq. (13)).
The left-hand side shows the relation between the width of a discretization frame wi and α2i and the right-hand side shows the relation between the probability mass
enclosed by the discretization frame with width wi and α2i .
Table 2
Evaluation of the discretization error for different measurement outcomes m, for problems with n ¼ 3 random variables. a is the constant in the LSF, Eq. (14); ρij is the
correlation coefficient between X i and X j for all i a j; P F and P F jM denote the analytically calculated prior and posterior failure probabilities; P^ F jM is the conditional failure
probability calculated with the discrete BN.
Table 3
Evaluation of the discretization error for different measurement outcomes m. The number of random variables n ¼ 4; a is the constant in the LSF, Eq. (14); ρij is the
correlation coefficient between X i and X j for all i a j; P F and P F jM denote the analytically calculated prior and posterior failure probabilities; P^ FjM is the conditional failure
probability calculated with the discrete BN.
from the error measure defined in Eq. (10), which balances the
relative with the absolute error. In addition, the results do not
display any apparent effect of correlation on the accuracy.
To assess the effect of the choice of the number of discretiza-
tion intervals, the failure probability P^ FjM was calculated for a
discretization scheme with up to 20 intervals per RV for the fourth
measurement case in Table 2. The estimated failure probabilities
P^ FjM are plotted together with the exact solution in Fig. 12.
X
n
g ðx Þ ¼ a Xi ð15Þ warning system is developed with the proposed discretization
i¼1 procedure. It provides RWO probabilities conditional on observa-
tions of the landing-weight, the headwind and the approach speed
Again the basic random variables X 1 –X n are distributed as X 1
for different aircraft types and different airports. For a detailed
LN ð0; 0:5Þ and X 2 ; …; X n LN ð1; 0:3Þ. Different cases with n ¼ 2,
description of how this problem can be treated in BN framework
3 and 4 random variables are investigated. Measurements M i ¼ mi
we refer to [35].
are available for all basic random variables; associated to these
RWO is the event of the operational landing distance exceeding
measurement are multiplicative measurement errors ϵi LN
the available runway length (Fig. 13). Correspondingly, a LSF for
ð0; 0:71Þ. For independent random variables X i it is possible to
runway overrun can be defined as:
determine posterior distributions f X i jMi ðxi jmi Þ analytically. The
posterior failure probabilities P FjM , which are used as reference g ðXÞ ¼ Runway n length Operational n landing
solutions, are calculated through importance sampling with 107 n distanceðXÞ ð16Þ
samples. The results are presented in Table 4.
The results in Table 4 do not differ substantially from with X representing the basic random variables of the problem.
Tables 2 and 3. This indicates that the (weak) non-linearity of the Drees and Holzapfel [6] proposed a model for the operational
LSF function describing failure does not affect the accuracy landing distance required by a landing aircraft, which is applied
significantly. here. The model, as well as the basic random variables X, are
presented in [34], which also includes a detailed description of the
4.3. Runway overrun reliability and sensitivity analysis.
We consider two different airports (AP I and AP II) and two dif-
Runway overrun (RWO) of a landing aircraft is one of the most ferent aircraft types (AC A and AC B). While the aircraft type affects
critical accidents types in civil aviation [11]. A conceptual RWO the landing-weight, the airport affects both the headwind and the
106 K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109
Table 4
Evaluation of the discretization error for different measurement outcomes m. The problems have n ¼ 2; 3 or 4 random variables; a is the constant in the LSF, Eq. (15); ρij is the
correlation coefficient between X i and X j for all ia j; P F and P F jM denote the prior respectively posterior failure probabilities, which are calculated through importance
sampling with 107 samples; P^ F jM is the conditional failure probability calculated with the discrete BN. Since for correlated basic random variables there is no analytical
solution for the posterior distribution, in these cases the updating of the basic random variables was performed through rejection sampling with 4 5E7 accepted samples.
n¼2
12 102 0 1.3E 5 ½2:8; 4:5 1.4E 5 1.2E 6 2E 6 15
12 102 0 1.3E 5 ½2:3; 2:4 3.3E 6 3.5E 6 2E 7 6
10 122 0 1.7E 4 ½4:0; 3:2 4.0E 4 3.7E 4 3E 5 7
12 102 0:5 1.7E 4 ½2:3; 2:4 4.8E 5 5.0E 5 2E 6 4
n¼3
15 103 0 3.7E 5 ½2:1; 5:6; 5:0 4.8E 5 4.5E 5 4E 6 7
15 103 0 3.7E 5 ½1:1; 3:7; 3:4 1.5E 5 1.8E 5 3E 6 20
13 123 0 5.0E 4 ½3:0; 3:0; 3:0 5.4E 4 5.4E 5 3E 6 1
16 103 0:5 9.1E 4 ½3:0; 6:0; 5:0 1.8E 3 1.9E 3 3E 5 2
n¼4
20 84 0 7.4E 6 ½2:0; 4:0; 3:4; 3:0 4.9E 6 5.6E 6 6E 7 13
17 84 0 3.1E 4 ½1:0; 1:4; 1:2; 2:0 4.5E 5 5.2E 5 7E 6 15
17 124 0 3.1E 4 ½3:1; 2:0; 3:3; 2:4 2.5E 4 2.5E 4 7E 6 3
24 84 0:5 3.7E 4 ½1:0; 1:4; 1:2; 2:0 1.1E 5 1.1E 5 9E 7 8
Table 6
Distribution models for head wind conditional on the airport.
Table 7
Distribution models for approach speed deviation conditional on the airport.
Fig. 13. Runway definitions.
Airport Approach speed deviation [kts]
Table 5
Distribution Mean Std. deviation
Distribution models for landing weight conditional on the aircraft.
approach speed. The distribution models for landing-weight, head- Random variable αi Annotation
wind and approach speed deviation at the different airports and
with the different aircraft types are given in Tables 5–7. All other (I/A) (I/B) (II/A) (II/B)
basic random variables of the problem are not affected by the airport
Landing weight [t] 0:09 0:10 0:11 0:09 Modeled
and aircraft type and are as in [34]. Headwind [kts] 0:65 0:61 0:67 0:60 Modeled
Table 8 summarizes the FORM importance measures of all Temperature [°C] 0:03 0:00 0:03 0:03 Not important
random variables X computed for the four combinations of air- Air pressure [hPa] 0:01 0:01 0:01 0:00 Not important
crafts and airports. Touchdown point [m] 0:20 0:16 0:18 0:20 Modeled
Approach speed deviation 0:20 0:21 0:20 0:24 Not observable
[kts]
4.3.1. Selection of relevant random variables Time of spoiler deployment 0:00 0:00 0:01 0:01 Not important
The applied RWO model includes 10 basic random variables. [s]
However, it is sufficient to include only a selection of these Time of breaking initiation 0:70 0:74 0:68 0:73 Not observable
explicitly in the BN. Random variables that are not relevant for the [s]
Time of reverser deploy- 0:03 0:04 0:06 0:05 Not important
prediction of RWO in the considered scenarios can be excluded. ment [s]
This is the case for random variables with a low FORM importance, Time of breaking end [s] 0:02 0:01 0:01 0:02 Not important
whose value does not depend significantly on airport and aircraft
type. Here, all random variables, whose absolute value of the
FORM importance measure jαi j's is smaller than 0:1, are excluded One can additionally exclude random variables that cannot be
(see Table 8). The one exception is landing weight, since its mean measured before the decision on whether to land or not is made.
value is substantially influenced by the aircraft type. This holds for Touchdown point and the time at which the pilot
K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109 107
Table 9
RWO probabilities for the different airports and aircrafts calculated with the dis-
crete BN pBN , together with solutions calculated by importance sampling around
the design point pDS . The latter have a sampling error with coefficient of variation in
the order of 10%.
Table 10
Probabilities of RWO and corresponding decision on landing, computed with the
BN for different sets of observations.
variables. As demonstrated by the verification examples, the The size of the sub-region of the outcome space of a random
heuristic procedure leads to accurate results. variable X i can be reduced significantly for random variables
This paper is restricted to static discretization. Application of whose corresponding uncertainty is dominating the reliability
the proposed procedure within dynamic discretization (e.g. [21]) problem.
should be investigated. The results of the procedure can serve as The number of intervals used for discretization should be
an initial discretization scheme, which is iteratively adjusted approximately equal for all basic random variables.
within dynamic discretization. This might strongly enhance the
convergence performance of these algorithms. On this basis, we propose a heuristic that can be used to find an
The gain in computational efficiency resulting from the pro- efficient discretization scheme. In verification examples, this
posed procedure over alternative static discretization approaches heuristic is found to give good accuracy and efficiency.
is problem specific. Some insights can be gained from Fig. 8. A
discretization with intervals of equal width centered at the origin
would require between approximately 2 and 5 times more inter-
vals per random variable to achieve the same accuracy. Because of References
the exponential increase of computational effort with number of
random variables, this leads to a considerable increase in effi- [1] Agena. 2005. Available: 〈http://www.agenarisk.com〉.
ciency. The gain compared to discretization with equal-frequency [2] Au S-K, Beck JL. Estimation of small failure probabilities in high dimensions by
subset simulation. Probab Eng Mech 2001;16:263–77.
intervals is expected to be even higher, since equal frequency [3] Der Kiureghian A. First- and second-order reliability methods. In: Nikolaidis E,
intervals focuses the fine intervals on the region of high prob- Ghiocel DM, Singhal S, editors. Engineering Design Reliability: Handbook. CRC
ability density rather than on the tails of the distribution. Press Inc.; 2005.
[4] Ditlevsen O, Madsen HO. Structural Reliability Methods. John Wiley & Sons;
This paper focuses on component reliability problems, which 2007.
are characterized by a single design point. Nevertheless the [5] Dougherty J, Kohavi R, Mehran S. Supervised and unsupervised discretization
heuristics derived can also be applied to system reliability pro- of continuous features. In: Prieditis A, Russel S, editors. Machine learning:
Proceedings of the twelfth international conference. San Francisco; 1995.
blems. System reliability problems can in general be treated as [6] Drees L, Holzapfel F. Determining and quantifying hazard chains and their
combinations of component reliability problems. Parallel and contribution to incident probabilities in flight operation. In: AIAA modeling
series systems are to be distinguished. For parallel systems dis- and simulation technologies conference. American Institute of Aeronautics
and Astronautics; 2012.
cretization should be performed based on the joint design point of [7] Friis-Hansen A. Bayesian networks as a decision support tool in marine
the problem. For series systems, following the same line of applications. Technical University of Denmark; 2000.
thought as in the runway overrun example, discretization can be [8] Hanea A, Napoles OM, Ababei D. Non-parametric Bayesian networks: improving
theory and reviewing applications. Reliab Eng Syst Saf 2015;144:265–84.
performed separately for each component problem (corresponds [9] Hanea AM, Kurowicka D, Cooke RM. Hybrid method for quantifying and
to the discrete cases i.e. airport–aircraft combinations in the RWO analyzing Bayesian belief nets. Qual Reliab Eng Int 2006;22:709–29.
example). In a second step the discretization schemes can be [10] Hohenbichler M, Rackwitz R. Nonnormal dependent vectors in structural
safety. J Eng Mech 1981;107:1227–38.
merged. In the same way it is possible to apply the heuristic to [11] Iata. Safety report 2012. International Air Transport Association (IATA); 2013.
multi-state components. One can treat each limit state surface [12] Jensen FV, Lauritzen SL, Olesen KG. Bayesian updating in causal probabilistic
(LSF) defining the boundary between two states separately and networks by local computations. Comput Stat Q 1990;4:269–82.
[13] Jensen FV, Nielsen TD. Bayesian networks and decision graphs. New York [u.
merge the discretization schemes afterwards. a.]: Springer; 2007.
The number of basic random variables in a single LSF that can [14] Kjaerulff UB, Madsen AL. Bayesian networks and influence diagrams: a guide
be modeled explicitly in a BN is limited to around 5–8. This is due to construction and analysis. Springer Publishing Company, Incorporated;
2013.
to the exponential growth of the target nodes CPT with increasing [15] Kotsiantis S, Kanellopoulos D. Discretization techniques: a recent survey.
number of parents and is independent of the discretization GESTS Int Trans Comput Sci Eng 2006;32:47–58.
method. Despite this limitation, BNs are applicable to many [16] Kozlov AV, Koller D. Nonuniform dynamic discretization in hybrid networks.
In: Proceedings of the 13th annual conference on Uncertainty in AI (UAI).
practical problems – particularly if one considers that usually not Providence, RI; 1997.
all basic random variables need to be modeled explicitly as nodes, [17] Langseth H, Nielsen TD, Rumi R, Salmerón A. Mixtures of truncated basis
as demonstrated in the presented example. functions. Int J Approx Reason 2012;53:212–27.
[18] Langseth H, Nielsen TD, Rumí R, Salmerón A. Inference in hybrid Bayesian
While in this paper the focus was on the discretization of the networks. Reliab Eng Syst Saf 2009;94:1499–509.
basic random variables, it is straightforward to incorporate the BNs [19] Lauritzen SL, Spiegelhalter DJ. Local computations with probabilities on gra-
discussed into larger models. phical structures and their application to expert systems. J R Stat Soc Ser
B (Methodol) 1988:157–224.
[20] Lerner UN. Hybrid Bayesian networks for reasoning about complex systems.
Stanford University; 2002.
[21] Neil M, Tailor M, Marquez D, Fenton N, Hearty P. Modelling dependable sys-
6. Conclusion
tems using hybrid Bayesian networks. Reliab Eng Syst Saf 2008;93:933–9.
[22] Rackwitz R. Reliability analysis—a review and some perspectives. Struct Saf
We investigate discretization of continuous reliability problems 2001;23:365–95.
[23] Raiffa H, Schlaifer R. Applied statistical decision theory. Boston: Division of
such that they can be treated in a discrete Bayesian network fra-
Research, Graduate School of Business Administration, Harvard University;
mework. Reliability problems with linear LSF in standard normal 1961.
space are considered. These can be seen as FORM approximations [24] Sindel R, Rackwitz R. Problems and Solution Strategies in Reliability Updating.
J Offshore Mech Arct Eng 1998;120:109–14.
of reliability problems. For these linear LSFs, optimal discretization
[25] Straub D. Stochastic modeling of deterioration processes through dynamic
schemes are found, which are optimal with respect to an error Bayesian networks. J Eng Mech 2009;135:1089–99.
measure calculated through a preposterior analysis. Since FORM is [26] Straub D. Reliability updating with equality information. Probab Eng Mech
2011;26:254–8.
known to give good approximations also for most non-linear
[27] Straub D. Engineering risk assessment. Risk – a multidisciplinary introduction.
reliability problem, the resulting discretization schemes are effi- Springer; 2014.
cient also for non-linear LSFs. The main findings presented in this [28] Straub D. Value of information analysis with structural reliability methods.
Struct Saf 2014;49:75–85.
paper are:
[29] Straub D, Der Kiureghian A. Bayesian network enhanced with structural
reliability methods: application. J Eng Mech 2010;136:1259–70.
An optimal discretization scheme should discretize finely the [30] Straub D, Der Kiureghian A. Bayesian network enhanced with structural
area around the FORM design point. reliability methods: methodology. J Eng Mech 2010;136:1248–58.
K. Zwirglmaier, D. Straub / Reliability Engineering and System Safety 153 (2016) 96–109 109
[31] Straub D, Papaioannou I, Betz W. Bayesian analysis of rare events. J Comput [35] Zwirglmaier K, Straub D. 2015. Discretization of structural reliability pro-
Phys 2016;314:538–56. http://dx.doi.org/10.1016/j.jcp.2016.03.018. blems: an application to runway overrun. In: Proceedings of the 12th Inter-
[32] Zhang NL, Poole D. A simple approach to Bayesian network computations. In: national Conference on Applications of Statistics and Probability in Civil
Proceedings of the tenth Canadian conference on artificial intelligence; 1994. Engineering, ICASP12. Vancouver, Canada.
[33] Zhu J, Collette M. A dynamic discretization method for reliability inference in
dynamic Bayesian networks. Reliab Eng Syst Saf 2015;138:242–52.
[34] Zwirglmaier K, Drees L, Holzapfel F, Straub D. Reliability analysis for runway
overrun using subset simulation. In: Proceedings of ESREL 2014. Wroclaw,
Poland; 2014.