Chapter 2

Introductory Mathematical Concepts

for Mining Equipment Reliability,
Maintainability, and Safety Analysis

2.1 Introduction

As in other areas of engineering analysis, various mathematical concepts play a piv-

otal role in mining equipment reliability, maintainability, and safety analysis. Al-
though the history of our current number symbols can be traced back to the stone
columns erected by the Scythian emperor Asoka of India in 250 B.C., the applica-
tion of mathematical concepts in engineering in general is relatively new [1].
In particular, probability plays a central role in the analysis of mining equipment
reliability, maintainability, and safety problems; its history may only be traced back
to the 16th-century writings of Girolamo Cardano (15011576) [1, 2]. In these writ-
ings, Cardano considered some interesting questions on probability. In the 17th cen-
tury, the problem of dividing the winnings in a game of chance was solved indepen-
dently and correctly by Blaise Pascal (16231662) and Pierre Fermat (16011665).
In the 18th century, probability concepts were further developed and successfully
applied to areas other than games of chance by Pierre Laplace (17491827) and
Karl Gauss (17771855) [2, 3].
A detailed history of mathematics including probability is available in
Refs. [1, 2]. This chapter presents various introductory mathematical concepts con-
sidered useful for performing mining equipment reliability, maintainability, and
safety analysis [4, 5].

2.2 Range, Arithmetic Mean, Mean Deviation,

and Standard Deviation

Many statistical measures are used to analyze reliability-, maintainability-, and

safety-related data. This section presents a number of such measures considered
useful for application in the area of mining equipment reliability, maintainability,
and safety.

B.S. Dhillon, Mining Equipment Reliability, Maintainability, and Safety 11


c Springer 2008
12 2 Introductory Mathematical Concepts for Mining Equipment Reliability

2.2.1 Range

This is a measure of dispersion or variation. More specifically, the range of a data

set is the difference between the largest and the smallest values in the set.

Example 2.1

A mining facility reported the following monthly equipment failures over a period
of 12 months:

40, 5, 10, 15, 20, 46, 50, 19, 25, 17, 35, and 16 .

Find the range of the above data set values.

By examining the given data values, we conclude that the largest and the small-
est values are 50 and 5, respectively. Thus, the range, R, of the given data set is
expressed by

R = Largest value Smallest value = 50 5 = 45 .

Thus, the range of the given data set is 45.

2.2.2 Arithmetic Mean

The arithmetic mean is defined by

n
mj
j=1
m= , (2.1)
n
where
m is the mean value,
mj is the data value j; for j = 1, 2, 3, . . ., n,
n is the total number of data values.

Example 2.2

A mining equipment manufacturing organization inspected ten identical mining sys-

tems and found 5, 10, 3, 2, 7, 15, 20, 1, 9, and 8 defects in each system. Calculate
the average number of defects per mining system (i.e., arithmetic mean of the data
set).
2.2 Range, Arithmetic Mean, Mean Deviation, and Standard Deviation 13

Inserting the specified data values into Eq. (2.1) we obtain

5 + 10 + 3 + 2 + 7 + 15 + 20 + 1 + 9 + 8
m= =8.
10
Thus, the average number of defects per mining system (i.e., arithmetic mean of the
data set) is 8.

2.2.3 Mean Deviation

This is a widely used measure of dispersion that indicates the degree to which given
data tend to spread about a mean value. The mean deviation is defined by
n  
 m j m
j=1
md = , (2.2)
n
where
n is the total number of data values,
mj is the data value j; for j = 1, 2, 3, . . ., n,
md is the mean deviation,
m
  is the mean value,
 m j m is the absolute value of the deviation of m j from m.

Example 2.3

Calculate the mean deviation of the data values given in Example 2.2. Using the
given data and calculated values of Example 2.2 in Eq. (2.2) yields

|5 8| + |10 8| + |3 8| + |2 8| + |7 8| + |15 8|
md =
10
|20 8|+ |1 8| + |9 8| + |8 8|
+
10
3 + 2 + 5 + 6 + 1 + 7 + 12 + 7 + 1 + 0
=
10
= 4.4 .

Thus, the mean deviation of the given data values is 4.4.

14 2 Introductory Mathematical Concepts for Mining Equipment Reliability

2.2.4 Standard Deviation

This is another measure of dispersion of data in a data set about the mean value. The
standard deviation is defined by [3]
n 1/2
2
(m j m)
j=1
=


, (2.3)
n

where
is the standard deviation.
The following three standard deviation properties are associated with the normal
distribution presented later in the chapter:
99.73% of the all data values are included between m 3 and m + 3 .
95.45% of the all data values are included between m 2 and m + 2 .
68.27% of the all data values are included between m and m + .

Example 2.4

Calculate the standard deviation of the data values given in Example 2.2.
Using the given data and calculated value of Example 2.2 in Eq. (2.3) we get

(5 8)2 + (10 8)2 + (3 8)2 + (2 8)2 + (7 8)2 + (15 8)2
=
10

1/2
(20 8)2 + (1 8)2 + (9 8)2 + (8 8)2
+
10
 
9 + 4 + 25 + 36 + 1 + 49 + 144 + 49 + 1 + 0 1/2
= = 5.64 .
10

Thus, the standard deviation of the data values given in Example 2.2 is 5.64.

2.3 Boolean Algebra Laws and Probability Definition

and Properties

Boolean algebra plays an important role in probability theory and is named after
mathematician George Boole (18131864). Some of its laws are as follows [6, 7]:

C + D = D +C , (2.4)
2.3 Boolean Algebra Laws and Probability Definition and Properties 15

where
C is a set or an event,
D is a set or an event,
+ denotes the union of events or sets
C D = D C , (2.5)
where
dot between C and D or D and C denotes the intersection of events or sets.
Sometimes the intersection of events is written without the dot
(e.g., CD), but it still conveys exactly the same meaning.

DD = D , (2.6)
C +C = C , (2.7)
C (C + D) = C , (2.8)
D + DC = D , (2.9)
C (D + E) = CD + CE , (2.10)

where
E is a set or an event.

C+0 = C , (2.11)
(C + D)(C + E) = C + DE . (2.12)

Probability may be defined as the likelihood of occurrence of a given event. Mathe-

matically, it is expressed as follows [8]:
 
M
P (Y ) = lim , (2.13)
n n

where
P(Y ) is the probability of occurrence of event Y ,
M is the number of times event Y occurs in the n repeated experiments.

Some probability properties are as follows [8]:

The probability of occurrence of event, say X, is

0 P(X) 1 . (2.14)

The probability of occurrence and nonoccurrence of an event, say X, is always

P(X) + P(X) = 1 , (2.15)

16 2 Introductory Mathematical Concepts for Mining Equipment Reliability

where
P(X) is the probability of occurrence of event X,
P(X) is the probability of nonoccurrence of event X.
The probability of an intersection of K independent events is given by

P(X1 X2 X3 . . . XK ) = P(X1 )P(X2 )P(X3 ) . . . P(XK ) , (2.16)

where

P(Xi ) is the probability of occurrence of event Xi , for i = 1, 2, 3, . . ., K.

The probability of the union of K independent events is expressed by

K
P(X1 + X2 + . . . + XK ) = 1 (1 P(Xi )) . (2.17)
i=1

The probability of the union of K mutually exclusive events is

K
P(X1 + X2 + . . . + XK ) = P (Xi ) . (2.18)
i=1

Example 2.5

Assume that in Eqs. (2.17) and (2.18) we have K = 2, P(X1 ) = 0.05, and P(X2 ) =
0.12. Calculate the probability of the union of events X1 and X2 using Eqs. (2.17)
and (2.18) and comment on the resulting probability values.
Inserting the given data into Eq. (2.17) we get

P (X1 + X2) = P (X1 ) + P(X2 ) P(X1 ) P (X2 )

= 0.05 + 0.12 (0.05)(0.12)
= 0.164 .

Using the specified data values in Eq. (2.18) we get

P (X1 + X2 ) = P (X1 ) + P(X2 )

= 0.05 + 0.12
= 0.17 .

This means that the probability of the union of mutually exclusive events X1 and X2
is higher than the probability of the union of independent events X1 and X2 .
2.4 Useful Mathematical Definitions 17

2.4 Useful Mathematical Definitions

This section presents a number of mathematical definitions considered useful to

perform reliability, maintainability, and safety studies in the mining industry [3, 8].

2.4.1 Cumulative Distribution Function

For continuous random variables, the cumulative distribution function is defined by

t
F(t) = f (x) dx , (2.19)

where
F(t) is the cumulative distribution function,
f (x) is the probability density function of continuous random variable x,
t is time.

For t = , Eq. (2.19) yields

F () = f (x) dx = 1 . (2.20)

This simply means that the total area under the probability density curve is equal to
unity.
Usually, in reliability work Eq. (2.19) is simply expressed as

t
F (t) = f (x) dx . (2.21)
0

2.4.2 Probability Density Function

For continuous random variables, the probability density function is defined by

dF(t)
f (t) = , (2.22)
dt
where
f (t) is the probability density function (in reliability work, it is often called
failure density function).
18 2 Introductory Mathematical Concepts for Mining Equipment Reliability

2.4.3 Reliability Function

The reliability function is defined by

t
R(t) = 1 F(t) = 1 f (x) dx , (2.23)
0

where

R(t) is the reliability function or simply reliability at time t.

2.4.4 Expected Value

For continuous random variables, this is defined by

E(t) = m = t f (t) dt , (2.24)

where
E(t) is the expected value of the continuous random variable t.
m is the mean of the continuous random variable t. In reliability work,
it is referred as mean time to failure.

2.4.5 Variance

The variance of a random variable t is defined by

2 (t) = E(t 2 ) [E(t)]2 (2.25)

or

(t) =
2
t 2 f (t) dt m2 , (2.26)
0

where
2 (t) is the variance of random variable t.
2.4 Useful Mathematical Definitions 19

2.4.6 Laplace Transform

This is defined by

f (s) = f (t) est dt , (2.27)
0

where
s is the Laplace transform variable,
t is the time variable,
f (s) is the Laplace transform of the function f (t).
Laplace transforms of some commonly occurring functions in mining equipment
reliability, maintainability, and safety studies are presented in Table 2.1 [9, 10].

2.4.7 Laplace Transform: Final Value Theorem

If the following limits exist, then the final-value theorem may be stated as

lim f (t) = lim [s f (s)] . (2.28)

t s0

Table 2.1 Laplace transforms of some commonly occurring functions

in mining equipment reliability, maintainability, and safety studies

f (t) f (s)
1
e t
(s + )
1
t e t
(s + )2
d f (s)
t f (t)
ds
c
c (a constant)
s
d f (t)
s f (s) f (0)
dt
m!
t m , for m = 1, 2, 3, . . .
sm+1
t
f (s)
f (t) dt
s
0
20 2 Introductory Mathematical Concepts for Mining Equipment Reliability

2.5 Probability Distributions

Over the years, a large number of probability distributions have been developed
to perform various types of statistical analysis [11]. This section presents some of
these probability distributions considered useful for application in the area of mining
equipment reliability, maintainability, and safety.

2.5.1 Binomial Distribution

This is a discrete random variable distribution and it was developed by Jakob

Bernoulli (16541705) [1]. Thus, it is also called a Bernoulli distribution. The distri-
bution is used in situations where one is concerned with the probabilities of outcome
such as the total number of occurrences (e.g., failures) in a sequence of given num-
ber of trials. More specifically, each of these trials has two possible outcomes (e.g.,
success or failure), but the probability of each trial remains constant or unchanged.
The binomial probability density function, f (x), is defined by

m!
f (x) = px qmx , for x = 0, 1, 2, 3, . . . , m , (2.29)
x!(m x)!

where
p is the single trial probability of occurrence (e.g., success),
q is the single trial probability of nonoccurrence (e.g., failure),
x is the number of nonoccurrences (e.g., failures) in m trials.

The cumulative distribution function is given by [8, 11]

x
m!
F(x) = pi qmi , (2.30)
i=0 i!(m i)!

where

F(x) is the probability of x or less nonoccurrences in m trials.

2.5.2 Exponential Distribution

This is a continuous random variable distribution and is widely used in reliability,

maintainability, and safety work. Two principal reasons for its widespread use are
as follows:
Easy to handle in performing various types of reliability, maintainability, and
safety analyses.
2.5 Probability Distributions 21

Constant failure rates of many engineering items during their useful life periods,
particularly electronic ones [12].
The distribution probability density function is expressed by

f (t) = e t , for > 0, t 0 , (2.31)

where
f (t) is the probability density function,
t is time,
is the distribution parameter. In reliability work, it is known
as the constant failure rate.

Substituting Eq. (2.31) into Eq. (2.21) we get the following expression for the cu-
mulative distribution function:
t
F(t) = e x dx = 1 e t . (2.32)
0

2.5.3 Rayleigh Distribution

This continuous random variable distribution is named after its originator, John
Rayleigh (18421919) [1]. The distribution probability density function is defined
by
2
f (t) = 2 t e( ) , for > 0, t 0 ,
t 2
(2.33)

where
is the distribution parameter.
Using Eq. (2.33) in Eq. (2.21) yields the following cumulative distribution function:

t
2 ( x )2
x e dx = 1 e( ) .
t 2
F(t) = (2.34)
2
0

2.5.4 Weibull Distribution

This continuous random variable distribution is named after W. Weibull, a Swedish

mechanical engineering professor, and it can be used to represent many differ-
ent physical phenomena [13]. The distribution probability density function is ex-
22 2 Introductory Mathematical Concepts for Mining Equipment Reliability

pressed by

1 ( t )
f (t) = t e , for t 0, > 0, > 0 , (2.35)

where

and are the distribution shape and scale parameters, respectively.

By substituting Eq. (2.35) into Eq. (2.21), we get the following equation for the
cumulative distribution function:
t
1 ( x ) t
F(t) =
x e dx = 1 e( ) (2.36)

0

It is to be noted that both exponential and Rayleigh distributions are the special
cases of Weibull distribution for = 1 and = 2, respectively.

2.5.5 Normal Distribution

This is one of the most widely used continuous random variable distributions and is
also known as the Gaussian distribution after Carl Friedrich Gauss (1777 1855).
However, the distribution was actually discovered by De Moivre in 1733 [11].
The probability density function of the distribution is defined by

1 (t )2
f (t) = exp , for < t < + , (2.37)
2 2 2

where
is the distribution mean,
is the distribution standard deviation.
Using Eq. (2.37) in Eq. (2.21) yields the following equation for the cumulative dis-
tribution function:
t

1 (x )2
F(t) = exp dx . (2.38)
2 2 2

2.5.6 Lognormal Distribution

This is another continuous random variable distribution and is often used to repre-
sent failed equipment repair times. The distribution probability density function is
2.6 Solving Differential Equations Using Laplace Transforms 23

expressed by

1 (lnt m)2
f (t) = exp , for t 0 , (2.39)
t 2 2 2

where
m and are the distribution parameters.
Using Eq. (2.39) in Eq. (2.21) yields the following cumulative distribution function:
t

1 1 (lnx m)2
F(t) = exp dx . (2.40)
2 x 2 2

2.6 Solving Differential Equations Using Laplace Transforms

Sometimes mining equipment reliability, maintainability, and safety studies may re-
quire finding solutions to a system of linear first-order differential equations. Under
such circumstances, the application of Laplace transforms has proven to be a very
effective approach. The following example demonstrates the application of Laplace
transforms to finding solutions to a set of linear first-order differential equations
describing a mining system:

Example 2.6

Assume that an engineering system used in mines can be, at any time t, in either of
the three distinct states: working normally, failed in open mode, or failed in short
mode. The following three linear first-order differential equations describe the min-
ing system:
dPw (t)
+ (om + sm ) Pw (t) = 0 , (2.41)
dt
dPom (t)
omP (t) = 0 , (2.42)
dt
dPsm(t)
smP (t) = 0 , (2.43)
dt
where
Pj (t) is the probability that the mining system is in state j at time t,
j = w (working normally), j = om (failed in open mode),
and j = sm (failed in short mode),
sm is the mining system constant short mode failure rate,
om is the mining system constant open mode failure rate.
At time t = 0, Pw (0) = 1, Pom (0) = 0, and Psm (0) = 0.
24 2 Introductory Mathematical Concepts for Mining Equipment Reliability

Find solutions to differential Eqs. (2.41)(2.43) using Laplace transforms.

Using Table 2.1, Eqs. (2.41)(2.43) and the given initial conditions we get

sPw (s) 1 + (om + sm ) Pw (s) = 0 , (2.44)

sPom (s) omPw (s) = 0 , (2.45)
sPsm (s) smPw (s) = 0 . (2.46)

Solving Eqs. (2.44)(2.46) we obtain

1
Pw (s) = , (2.47)
s + om + sm
om
Pom (s) = , (2.48)
s (s + om + sm )
sm
Psm (s) = . (2.49)
s (s + om + sm )

Taking the inverse Laplace transforms of Eqs. (2.47)(2.49) we get

Pw (t) = e(om +sm )t , (2.50)

om  
Pom (t) = 1 e(om +sm )t , (2.51)
om + sm

and  
sm
Psm (t) = 1 e(om +sm )t . (2.52)
om + sm
Thus, Eqs. (2.50)(2.52) are the solutions to differential Eqs. (2.41)(2.43).

2.7 Problems

1. A mining equipment manufacturing company inspected eight identical mining

systems and found 10, 11, 2, 20, 6, 9, 4, and 5 defects in each system. Calculate
the average number of defects per mining system (i.e., arithmetic mean of the
data set).
2. Calculate the mean deviation of the data values given in the above problem (i.e.,
Problem 1).
3. Discuss the history of probability.
4. Define standard deviation.
5. Prove Eq. (2.12).
6. Discuss five important properties of probability.
7. Mathematically, define probability.
8. Determine expected value of Eq. (2.31).
References 25

9. Write down probability density functions for the following statistical distribu-
tions:
Normal distribution
Weibull distribution
10. Obtain the Laplace transform for the following function:

f (t) = t e t , (2.53)

where
t is the time variable,
is a constant.

References

1. Eves, H.: An Introduction to the History of Mathematics. Holt, Rinehart, and Winston, New
York (1976)
2. Owen, D.B. (ed.): On the History of Statistics and Probability. Dekker, New York (1976)
3. Shooman, M.L.: Probabilistic Reliability: An Engineering Approach. McGraw-Hill, New
York (1968)
4. Spiegel, M.R.: Probability and Statistics. McGraw-Hill, New York (1975)
5. Spiegel, M.R.: Statistics. McGraw-Hill, New York (1961)
6. Lipschutz, S.: Set Theory. McGraw-Hill, New York (1964)
7. Fault Tree Handbook, Report No. NUREG-0492, US Nuclear Regulatory Commission,
Washington, DC (1981)
8. Mann, N.R., Schafer, R.E., Singpurwalla, N.D.: Methods for Statistical Analysis of Reliability
and Life Data. Wiley, New York (1974)
9. Spiegel, M.R.: Laplace Transforms. McGraw-Hill, New York (1965)
10. Oberhettinger, F., Badic, L.: Tables of Laplace Transforms. Springer, New York (1973)
11. Patel, J.K., Kapadia, C.H., Owen, D.B.: Handbook of Statistical Distributions. Dekker, New
York (1976)
12. Davis, D.J.: An analysis of some failure data. J. Am. Stat. Assoc. 113150 (June 1952)
13. Weibull, W.: A statistical distribution function of wide applicability. J. Appl. Mech. 18, 293
297 (1951)
http://www.springer.com/978-1-84800-287-6