See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/370251760

On Reliability: A Mathematical Fault Tree

Article · March 2023

DOI: 10.24412/1932-2321-2023-172-78-85

CITATIONS READS

0 38

3 authors:

Mohamed Fahmy Ahmed Ibrahim Ahmed

October University for Modern Sciences and Arts October University for Modern Sciences and Arts
4 PUBLICATIONS 0 CITATIONS 5 PUBLICATIONS 0 CITATIONS

SEE PROFILE SEE PROFILE

Mohamed Khalil
October University for Modern Sciences and Arts
45 PUBLICATIONS 576 CITATIONS

SEE PROFILE

All content following this page was uploaded by Ahmed Ibrahim Ahmed on 13 July 2023.

The user has requested enhancement of the downloaded file.

M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

On Reliability: A Mathematical Fault Tree

M. S. Fahmy, A. I. Ahmed, M. Khalil
•
Faculty of Engineering, October University for Modern Sciences and Arts (MSA), Egypt.
[email protected]; [email protected]; [email protected]

Abstract

Fault tree analysis (FTA) is a top down approach that was initially used and developed in Bell laboratories
in the year 1962 by H Watson and A Mearns for the intercontinental ballistic missile (ICBM) system for
the US air force called the Minuteman System. Since then, the technique has been adopted and adapted
by many companies who are interested in reliability engineering and dangerous technology. Today FTA is
widely used in system safety and reliability engineering, aerospace, nuclear power, chemical and process,
pharmaceutical, petrochemical and other high-hazard industries; but is also used in fields as diverse as risk
factor identification relating to social service system failure and in software engineering for debugging
purposes and is closely related to cause-elimination technique used to detect bugs. Now FTA is considered
as one of the most important system reliability and safety analysis techniques. Fault tree analysis has
proved to be a useful analytical tool to analyze the potential for system or machine failure by graphically
and mathematically representing the system itself. It is a top-down approach that reverse-engineers the
root causes of a potential failure through the root cause analysis process. Our main contribution is to
develop a mathematical theory of fault tree analysis using some statistical concepts relating to probability
of series and parallel systems to set up a mathematical model that represent any hierarchical control
system to calculate its reliability for both homogeneous and nonhomogeneous structures. A Fault Tree is
a hierarchical model used to analyze the probability that an event will occur. Fault Tree provides all the
tools needed to build graphic representations of large-scale problems gracefully so we can use it to set up
a mathematical model that represent any hierarchical control system and evaluate its reliability using our
general mathematical formula that represent the structure in its two cases. The graphical representation
(fault tree diagram) for a hierarchical controlled system enabled us to set up a mathematical general
formula that help us to evaluate the reliability of the system in general case (nonhomogeneous structure)
and another derived formula for the special case (homogeneous structure). This analysis may help to
understand how one or more small failure events lead to a catastrophic failure.

Keywords: Reliability, Mathematical Modeling, Analysis of Fault Tree, Serial and Parallel Systems

1. Introduction
Reliability as a state can be a factor of of many parameters; such as, Mean Time To Failure,
Reliability, Availability and few others. These terms have been developing over the last six decade.
Its evident that such a concept will be portrayed in a structure of system or systems, Barlow in
1973 and Fussel in 1974, discussed the fault tree construction and concept respectively [1 and 2].
In some articles it has been observed that many of the quantities computed by fault tree analysis
can also be computed using the concepts and techniques of reliability theory. Henceforth, this
paper aims to build and construct an intuitive rigor of understanding reliability in the realm of
Mathematical Statistics, in form of a fault tree mathematical model, whereas we aim to introduce
the concepts in a matter of detail, building the theory upon prior establishments, and presenting
a a general model by the end, incorporating the aforementioned usage of the prior sections. It
serves, to add, that general assumptions will be accounted for and discussed, also mentioned

78
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

when done otherwise. The Introduction section will serve as a block of terms, dissected into
multi-topics that adds up to the required definition this paper aims to deliver. For in Section
1 we get to understand what reliability as a concept with their basic definitions. In Section 2
we discuss how system connections influence their reliability, and least in Section 3; We present
and elaborate the homogeneity of a system of fault tree representation, giving a derived general
model by the end.

1.1. Basic Definitions and Concepts

In this section we introduce some basic terms as reliability and its components, as well as its rigor
definitions.

1.1.1 Reliability: What we quantify as reliable

Reliability purposefulness can be interpreted as; it serves as a developmental method for Engi-
neering. In which it helps in: cost, effort, and time efficiency. It is essential to grasp that reliability
is not a property, but a characteristic of an item, in which it is expressed by the probability that
the said system will function as expected for a stated time interval [3]. It is usually denoted by R.
One can view reliability in other terms, namely, a quantitative approach, as from said approach,
the one thing we care about the most, is how long a system stays operational with no interruption.
However, this does not imply that redundant parts might not fail, as they can fail and be repaired
without causing an operational interruption at system level. Thus, one may conclude that, the
concept of reliability can be applied to both repairable and non-repairable systems.
Definition 1.1 (Reliability). Given n statistical identical systems, which starts into operations at
t = 0, v̄ < n is to accomplish them successfully, where v̄ ∈ Rn .Then, We can write that the ratio
v̄
n is a mere random variable which converges to the true value of the reliability as n increases.
Namely, limn→∞ nv̄ = R: | nv̄ − R| < ϵ
Definition 1.2 (Reliability). It is the ability of an object (or process or service) to function as
expected to fulfil the demanded tasks under given conditions. In other accurate words reliability
is the probability a component or system will perform as designed. In which the value would
range from [0,1]. reliability is related to failure rate by a simple exponential function: R = e−λt
Where R is the reliability, e is the Euler constant, λ is the failure rate, and t is the time.

1.1.2 Essential Metrics: Basic Definitions

Few Metrics need to be defined in order to be able to calculate the reliability of a system [3, 4,
and 5].
Definition 1.3 (Failure rate (λ)). The frequency of a component failure per unit time, it is an
essential metric that is used to calculating either reliability or availability.
1 1
λ= MTBF ≡ MTTF

Definition 1.4 (Repair rate (µ)). The frequency of successful repair operations performed on a
failed component per unit time.
1
µ= MTTR

Definition 1.5 (Mean time to failure (MTTF)). The average time duration before a non-repairable
system component fails.
∑ Hours o f Operation 1
MTTF = ∑ Units
≡ λ

Definition 1.6 (Mean time between failure (MTBF)). The average time duration between inherent
failures of a repairable system component.
∑ Hours o f Operation 1
MTBF = ∑ Failurs
≡ λ ≡ MTTF+ MTTR

79
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

2. System Types: Analysis

What we are concerned about mathematically, when we are assessing the reliability of a system is
its own basic sub-systems. The purpose of this paper, is to detail the explanation to the process,
given a system or a network of systems that are adjoined together. The physical layout or rather
the connection of the system is of a cruciality to both its functioning, and its reliability.

2.1. Series Systems

A network with N systems or blocks in a series link (Figure 1 [3]) where the failure of any one
item causes the entire system to fail [3]. As a result, for a series system to perform properly, all of
its components must function properly within the time range specified t.

Figure 1: Series System

Definition 2.1. (Reliability of a Series System) The reliability of a system in a series connection is
the probability that all N items succeed during its intended interval of time t.

N
R s ( t ) = R1 ( t ) · R2 ( t ) . . . R N ( t ) = ∏ Ri ( t )
i =1

A practical conclusion is that the reliability of a series system is always lower than the reliability
of any of its components.

We are also concerned with the instantaneous failure rate, one can conclude such an outcome
by recalling the definition of λ(t):

−d ln ∏iN=1 Ri (t) N
−d ln Ri (t) N
λs (t) = ≡∑ ≡ ∑ λi ( t )
dt i =1
dt i =1

2.2. Parallel Systems

In a parallel arrangement or link, a system failure is caused by the failure of all components, not
just one. As a result, the performance of only one unit will be enough to ensure the system’s
overall success.

Figure 2: Parallel System

80
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

Definition 2.2 (Reliability of a Parallel System). For a set of N independent items connected in
parallel (Figure 2 [3]), their failure rate is be given by:

N
Fs (t) = F1 (t) · F2 (t) . . . FN (t) = ∏ Fi (t)
i =1

Since Ri (t) = 1 − Fi (t)

N
Rs (t) = 1 − Fs (t) = 1 − ∏[1 − Ri (t)]
i =1

The instantaneous failure rate is still an essential metric, however in parallel configuration,
it is not as trivial to come up with one. one can start with the definition that the failure rate is
−d ln R(t)
h(t) = dt , yet it will lead to a complicated formula, for instance, let a system of two units
with constant failure rate be connected in parallel, their failure rate can be given by;

λ1 exp (−λ1 t) + λ2 exp (−λ2 t) − (λ1 + λ2 ) exp (−(λ1 + λ2 )t)

λs (t) =
exp (−λ1 t) + exp (−λ2 t) − exp (−(λ1 + λ2 )t)

For the same instance, or case; of an N identical units in parallel with a constant failure rate, their
reliability can be put to as:
Rs (t) = 1 − [1 − exp (−λt)] N

2.3. General Structure

Consider the K-out-of-N system, which is a more general structure of series and parallel systems
(Figure 3 [3]). If any combination of K units out of N independent units works in this type of
system, the system is guaranteed to succeed. Assume that all units are identical for the sake of
simplicity. The likelihood that the system will work is represented by the binomial distribution
[7]:
N   K −1  
N N
Rs (t) = ∑ [ R(t)]r [1 − R(t)] N −r = 1 − ∑ [ R(t)]r [1 − R(t)] N −r
r =K
r r =0
r
The majority of practical systems are neither parallel nor series, but rather a blend of the two.
Parallel-series systems are a common name for these systems. A complex system that is neither
series nor parallel alone, nor parallel-series, is another sort of complex system. Which is referred
to as nonparallel-series system (Figure 4 [3]).

Figure 3: A parallel-series system Figure 4: A nonparallel-series system

3. Homogeneity of a System
In this section we discuss the homogeneity of a system in form of a fault tree, where their nodes
branching, and rate of distribution affect their reliability. We start by introducing what a fault
tree is, then how they are analyzed.

81
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

3.0.1 Fault Trees and Reliability Block Diagrams

It is important to understand the essential difference between RBDs and fault tree diagrams;
RBDs work in success space while, FTDs work in failure space; the FTDs addresses the failure
combinations while the RBDs address the success combinations. Also FTDs are traditionally used
in analyzing fixed probabilities, while RBDs may include time-varying distributions for the blocks’
failure or success, in addition to other properties like restoration or repair distributions [5 and 6].

3.0.2 Fault Tree Analysis

Fault Tree Analysis can be explained simply as an analytical technique that describes an undesired
state of the system. normally that is in a critical state from a safety standpoint. The system
is inspected in the environmental context and operation to extract all credible ways for the
undesired event to occur [3,4, and 5]. It is also important to point out that a fault tree is a graphic
model of the different sequential and parallel combinations of faults that will happen in the
investigated model. In fact, using the model of fault tree is more convenient to deal with, because
it is qualitative model that enable us to evaluate it quantitatively and do not change the qualitative
nature of the model itself.

3.1. Homogeneous and Non-homogeneous systems: Analysis

The homogeneity of a system influences the reliability of the whole system [3] and that can be
seen evident in a fault tree, where some demonstrate a homogeneous structure, and some do
not. We represent in Figure 5 the model of general fault tree, and define how said figure can be
interpreted to be homogeneous or non-homogeneous.

Figure 5: Generic Tree

Definition 3.1 (Homogeneity of a tree). A tree is said to be homogeneous if and only if, the
number of its sub nodes is equal to the number of every other sub node on any level from the
root. Mainly, n1k = n2m = nij

Definition 3.2 (Non-Homogeneity of a tree). A tree is said to be non-homogeneous if and only if,
the number of its sub nodes is not equal to at least one other sub node on any level from the root.
Mainly, n1k ̸= n2m ̸= nij

82
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

3.2. General Case

Figure 6: φ − Tree

In Figure 6, let P0 be the reliability of the root; q1 , . . . , qm are to be the reliability of the edges, and
r1 , . . . , rm be the reliability of the nodes of the first level, where m is the number of nodes of the
first level; qi1 , . . . , qini be the reliability of edges of second level in the i-th subtree; ri1 , . . . , rini
be the reliability of nodes of the second level in the i-th subtree, where ni is the number of
edges (nodes) in the i-th subtree. It goes evident to see that the number of leaves is equal
to N, where; N = n1 + n2 + . . . + nm such that n1 ≤ n2 ≤ . . . ≤ nm . For the purpose of
generality, consider that an arbitrary path from the root to the end is unfailing (successful), if
all nodes and edges on said path is unfailing (successful). Now we find the reliability of the
tree (φ) through V paths, throughout the derivation such reliability is detonated ℘( φ; V ); where;
V = 1, 2, . . . , N. N ∈ N-{0}.

Theorem 1.

" !!#
m ni
∑ ∑ ∏ 1 − ∏ 1 − Pij


℘( φ; V ) = P0 × 1 − Pi ×
k=0 A,A⊂{1,...,m} i∈ A j =1
| A|=k
     

∑  ∏ ( Pi ×  ∑  ∏ Pij ×  ∏
    


     1 − Pij  

a1 +...+ am =V i∈
/A B,B⊂{1,...,ni } j∈ B ∈B
j/
i ∈ A =⇒ ai =0 | B|= ai 1≤ j ≤ n i

Where:
Pi = ri qi , Pij = rij qij , i = 1, . . . , m i ≤ j ≤ nj

∏ = 1, ∑ = 0
ϕ ϕ

Proof.
We proof the theorem by showing that the probability of failure throughout all paths from the

83
M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

root to the i-th subtree with ni nodes is equal to

nj
!
1 − ∏ 1 − Pij


1 − Pi
j =1

The reliability of the fault tree through exactly a paths (1≤ a ≤ni ) from the root to the i-th subtree
is equal to    

∑  ∏ Pij ×  ∏
  

Pi × 
   1 − Pij  

B,B⊂{1,...,ni } j∈ B ∈B
j/
| B|= ai 1≤ j ≤ n i

Further we remark that the event of having exactly V paths to operate successfully is considered
the sum of all mutually exclusive events of the types; there are exactly a1 paths operate simul-
taneously successfully ending in the first subtree, a2 paths operate simultaneously successfully
ending in the second subtree, and am paths operate simultaneously successfully ending in the
m-th subtree; where a1 + a2 + . . . + am = V. From here it follows
 
m
∑ ∏ Ω(i, ai )
 
℘( φ; V ) = P0 × 
 
a1 +...+ am =V i =1
0≤ a j ≤ n j

where  !
 nj
1 − Pi 1 − ∏ 1 − Pij


ai = 0






 j =1
 
Ω(i, ai )=
 
■
 

Pi ×  ∑ ∏ j∈ B Pij ×  ∏ 1 − Pij  ai ̸ = 0

  


∈B
 

 B,B ⊂{ 1,...,n i } j/
1≤ j ≤ n i

| B|= ai

3.3. Special Case

For the special case that the system behaves in a homogeneous matter as per Definition 3.1.
Namely, this implies that: q1 = q2 = . . . = qm = q0 , and for the second level operating on the same
type and identical conditions: q11 = . . . = q1n = q21 = . . . = q2n2 = . . . = qm1 = . . . = qmnm = q1 .
1
It trivially follows that r11 = . . . = r1n1 = r21 = . . . = r2n2 = . . . = rm1 = . . . = rmnm = r1 , and
n1 = n2 = . . . = nm = n.
By transforming the formula of Theorem 1, by change of assertions; namely: P1 = P2 = Pm =
q0 r = p1 ; P11 = P12 = . . . = Pm1 = . . . = Pmnm = q1 r = p2 .
Let ℘( φ1 ; V ) denote the reliability of such case, from Section 2.3 it follows:
Corollary 1.
m
k×n
    
m n
m−k
∑ k V (k×n)−V

℘ ( φ1 ; V ) = p0 × × p1 × 1 − p1 × 1 − (1 − p2 ) × × p2 × (1 − p2 )
k V
k=⌈ Vn ⌉
l m
Where; (mk) = m!
k!(m−k)!
And V
n is the smallest integer greater than V
n

4. Conclusion
A theoretical study is introduced for a mathematical model that can describe a hierarchical
control system (homogeneous and non-homogeneous) in order to evaluate its reliability. This
investigation of the reliability is highly suitable for study of many Engineering applications
ranging from industrial process control, through production management to Economic and other
systems.

84
 M. S. Fahmy, A. I. Ahmed, M. Khalil RT&A, No 1 (72)
ON RELIABILITY: A MATHEMATICAL FAULT TREE Volume 18, March 2023

References
[1] Barlow, E. (1973). Introduction to Fault Tree Analysis. Research Air Force Systems Command.
[2] Fussell, J. B. (1974). Fault Tree Analysis - Concepts and Techniques. Nato Advanced Study
Institute On Generic Techniques Of System Reliability Assessment, Nordhoff Publishing
Company.
[3] Rausand, M., Hyland, A. (2004). System Reliability Theory: Models, Statistical Methods,
and Applications (2nd ed.). Wiley.
[4] Epstein, Weissman. (2009). Mathematical Models for Systems Reliability. CRC Press.
[5] Mavko, B., V., A., Marko, C. (2009). Application of the fault tree analysis for assessment of
power system reliability (Vol. 94). Reliability Engineering System Safety.
[6] Curcuru, G., Galante, G. M., La Fata, C. M. (2013). An imprecise Fault Tree Analysis for the
estimation of the Rate of Occurrence Of Failure. J Loss Prev Process Ind.
[7] Carpitella, S., Certa, A., Izquierdo Sebastin, J., La Fata, C. M. (2018). k-out-of-n systems:
an exact formula for the stationary availability and multi-objective configuration design
based on mathematical programming and TOPSIS. Journal of Computational and Applied
Mathematics.

85

View publication stats