Sei sulla pagina 1di 8

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013

by M. J. Baker and H. Tan




BASIC CONCEPTS IN PROBABILITY THEORY
1. Introduction
A good understanding of statistics and probability theory is needed for this module. The current chapter
provides most, if not all, the basic theory in statistics and probability that will be required for Course
EG50S1.
It is likely that many of you will have already covered much of this material in undergraduate degree
courses; so much of it should seem familiar. However, those of you who have not studied it before will
need to spend more time on this material. In all cases, your aim in studying should be to gain sufficient
understanding of all the topics to enable you to appreciate the later units on quantitative risk and
reliability analysis. It is suggested that you work slowly through the chapter making sure that you
understand all the concepts and studying the examples. This applies especially to those of you taking
the course by distance learning.
These notes are reasonably comprehensive, but for those of you wishing to refer to textbooks a short
list of suitable texts are given at the end of this chapter.
2. Sample space
The theory of probability is concerned formally with experiments and their outcomes. The collection of
all possible outcomes of an experiment is called its sample space and is here denoted O. This space
consists of a set of points called sample points, each of which is associated with one and only one
distinguishable outcome. For example, the outcome of the throw of a six-sided die has as its sample
space the numbers one to six inclusive, O = {1, 2, 3, 4, 5, 6}, where each point in the sample space O
is characterized by the number on the upward side of the die.

Figure 1, Sample space of the outcome of throwing a six-sided die.
In probability theory, an event is a set of outcomes (a subset of the sample space) to which a probability
is assigned. Events may be visualised more easily by using Venn diagrams. For example suppose the
event A is {face an odd number} and the event B is {face a multiple of 3}, then A = {1,3,5} and B =

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

{3,6}. These events can be represented symbolically by circles as shown in Figure 1. Because the
number 3 is contained in both events A and B, the circles overlap.
In Figure 1, the rectangle represents the sample space, which is all of the possible outcomes of the
experiment. The circle labelled as A represents event A. In other words, all of the points within A
represent possible ways of achieving the outcome of A. Similarly all of the points within B represent
possible ways of achieving the outcome of B.
The above is an example of a discrete sample space, but continuous sample spaces are generally more
common in engineering for example, the sample space of annual maximum wind speed at a particular
location.
3. Union and intersection
The terms union and intersection are defined as follows. If we have two events A and B: The union
( A B ) means all the sample points contained by either A or B or both. The intersection ( A B )
means only the sample points common to both A and B. These are illustrated in Figure 2. In the Venn
diagrams, the shaded areas represent the sets under discussion.

(a) (b)
Figure 2, Union and intersection of events.

Examples:
As shown in Figure 1, event { } 1, 3, 5 A= , { } 3, 6 B = . Then { } 1, 3, 5, 6 A B = and {3} A B = .
4. Complementary and empty sets
It is conventional to denote the sample space by the symbol O. Let A be an event. The event containing
all the sample points in O that are not in A is called the complementary event and is denoted by A. It

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

follows that A A = O and A A | = where the symbol | denotes the impossible event (with a
probability zero of occurring) represented by a null or empty set.
Examples:
For events A and B shown in Figure 1, { } 2, 4, 6 A = , { } 1, 2, 4,5 B = ,
{ } 2, 4 A B = , { } 1.2, 4, 5, 6 A B =
{ } { } 1, 3, 5, 6 2, 4 A B A B = = = ,
{ } { } 3 1, 2, 4, 5, 6 A B A B = = = ,
( )
( )
A B A B | =
5. Axioms of probability
In probability theory, an event is a set of outcomes (a subset of the sample space) to which a probability
is assigned. The probability of the occurrence of an event A is denoted by P(A).
Axiom 1: The probability of an event is a number greater than or equal to zero, but less than or equal to
unity:
0 ( ) 1 P A s s (1)
Axiom 2: The probability of a certain event is unity, that is:
( ) 1 P O = (2)
since O is the event associated with all sample points in the sample space.
In probability theory, events E
1
, E
2
, ..., E
n
are said to be mutually exclusive if the occurrence of any one
of them automatically implies the non-occurrence of the remaining n 1 events. Therefore, two
mutually exclusive events cannot both occur. Formally said, the intersection of each two of them is
empty (the null event | ): A B | = . In consequence, for mutually exclusive events A and B,
( ) 0 P A B = .
Axiom 3: The probability of an event which is the union of two mutually exclusive events A and B is
the sum of the probabilities of these two events:
( ) ( ) ( ) P A B P A P B = + (3)

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

The first two axioms are simply convenient conventions so that all probabilities will be positive
numbers and their sum over any mutually exclusive and collectively exhaustive set of events (that is a
non-overlapping set of events that occupy the whole sample space) will be normalised to unity.
6. Intersecting events
If two events A and B are not mutually exclusive, then their intersection A B will be associated with
a non-zero probability, as shown in Figure 2(b). The following can be easily proved:
( ) ( ) ( ) ( ) P A B P A P B P A B = + (4)
7. Theorem of conditional probability
The conditional probability of the event A, given that the event B has occurred, is denoted ( | ) P A B ,
and is defined as

( )
( | )
( )
P A B
P A B
P B

= (5)
The conditional probability ( | ) P A B can be interpreted as the probability that A will occur given the
knowledge that B has occurred. The condition that B has occurred restricts the outcome to the set of
sample points of A in B, but should not change the relative likelihood of the simple events in B.

Figure 3, Conditional probability.
In effect, the probability measure of those points in A that are also in B, ( ) P A B , is re-normalised by
the factor 1/ ( ) P B

to give ( ) / ( ) P A B P B according to axiom 2, in this case B can be thought of as a
reduced sample space.
Following from Eq. (5) by rearrangement:
( ) ( | ) ( ) ( | ) ( ) P A B P A B P B P B A P A = = (6)

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

8. Statistical independence
If two physical events are not related in any way we would not alter our estimate of the probability of
one even if we knew the other had occurred, e.g.
Event A The butt weld has slag inclusions.
Event B The welder was born on a Wednesday.
This notion leads to the definition of statistical independence.
Two events A and B are said to be independent if and only if
( | ) ( ) P A B P A = (7)
Hence for independent events A and B, from Eqs (5) and (7)

( )
( | ) ( )
( )
P A B
P A B P A
P B

= = (8)
Thus for independent events
( ) ( ) ( ) P A B P A P B = (9)
Equation (9) is known as the multiplication rule and applies for any number of mutually independent
events.
Example:
Consider the statically determinate structural system with 7 elements shown in Figure 4. Let the event
that element i fails be denoted by F
i
and let the probability of failure of element i be ( )
i
P F . Further,
assume that failures of the individual members are statistically independent, that is
( ) ( ) ( )
i j i j
P F F P F P F = for any pair of (i, j). For this statically determinate structure, the failure of
any member will result in system failure. Hence,
( )
( ) ( ) ( )
| || | | |
1 2 7
1 2 7
1 2 7
(Fail of structure)
1 (structure does not fail)
1
1 because of statistical independence
1 1 ( ) 1 ( ) 1 ( )
P
P
P F F F
P F P F P F
P F P F P F
=
=
=
=
(10)
Suppose the probability of failure for elements 1 to 7 are
1 3 5 7
( ) ( ) ( ) ( ) 0.02 P F P F P F P F = = = = ,
2 6
( ) ( ) 0.01 P F P F = = and
4
( ) 0.03 P F = , Then

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

( ) ( ) ( )
4 2
(Fail of structure) 1 1 0.02 1 0.01 1 0.03 0.1231 P = =

Figure 4, A statically determinant structural system.
9. Probability of system failure
Consider the piping system shown in Figure 5 in which valves A, B, C and D are initially closed.
Calculate the probability that flow will not occur between X and Y when a signal is sent to all valves to
open, given that valve failures are statistically independent and the probabilities of failure of individual
valves and their actuators are

3
( ) ( ) ( ) ( ) 10
A B C D
P F P F P F P F

= = = =
where F
A
is the event that valve A will not open, etc.


Figure 5, Valve layout

Solution:
System failure may be expressed in Boolean logic as

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan


{ }
( )
( ) ( )
system failure
A BCD
A BC D
A B C D
F F
F F F
F F F F
=
=
=

where
BCD
F mean failure of the sub-system containing valves B, C and D; and
BC
F mean failure of the
sub-system containing valves B and C. The probability of system failure is therefore
( ) ( ) ( )
, f sys A B C D
P P F F F F =

( )
( )
( )
( ) ( )
( ) ( ) ( ) ( )
( )
1
1
1
1 1 1
BC B C
B C
B C
B C
B C
P F P F F
P F F
P F F
P F P F
P F P F
=
=
=
=
=


( ) ( )
( ) ( )
BCD BC D
BC D
P F P F F
P F P F
=
=

( ) ( ) ( ) ( )
( ) 1 1 1
A BCD A BCD
P F F P F P F =
Numerically,

( )( )
4 4 3
( ) 1 1 10 1 10 1.999 10
BC
P F

= =


6
( ) 1.999 10
BCD
P F

=

3
( ) 10
A BCD
P F F

~

NOTES:
(1) This example deals only with the probabilistic part of the problem and not the engineering part,
which is highly simplified.
(2) All valve failure events are assumed to be statistically independent, which may not be realistic in
practice.

Fundamental Safety Engineering and Risk Management Concepts, 2012/2013
by M. J. Baker and H. Tan

(3) The value of the probability of system is very close to
3
( ) 10
A
P F

~

because of the way in which
the valves are configured, being dominated by the event F
A
.

Potrebbero piacerti anche