Log Normal Distribution
Uploaded by
ScribdTranslationsLog Normal Distribution
Uploaded by
ScribdTranslationsDISTRIBUTION
LOGARITHMIC
NORMAL
WHAT IS DISTRIBUTION
NORMAL LOGARITHMIC?
It is a probability distribution of a random variable whose logarithm is normally
distributed. That is, if X is a random variable with a normal distribution, then exp(x) has a
log-normal distribution.
The log-normal distribution tends to the probability density function.
f(x; u; o ) = ^ - (ln
(x
) -u ) 2 /2.0 2
xa V 2T0
Where:
X>0
u: Medium
O: Standard deviation
Lognormal Distribution
f(t)= ex p 1 In (t)-pt. j
Distribution Function 2 \G
Lognormal
where p and o are functions of In s
Lognormal Probability Density Function
1-64
RELATIONSHIP WITH THE MEAN AND THE
STANDARD DEVIATION
The log-normal distribution, the geometric mean, and the geometric standard deviation are related. In this
case, the geometric mean is equal to exp(u) and the geometric standard deviation is equal to exp(o)
Confidence interval limit log Geometric
30 1 Finite lower u — 30 /3
Fgeo/ °geo
,, /,2
2nd lower limit u — 20 Hgeoj °geo
the lower limit u-0 Hgeo/
Ogeo
the upper limit u+0 HgeoOgeo
2
2nd upper limit u — 2o Hgeo°geo
3rd upper limit u + 30
HgeoOgeo
Where the geometric mean ^ geo = exp(
u) and the geometric standard deviation a geo = exp(o)
MEAN AND VARIANCE OF X
Half:
Variance:
The equations are derived from the moment generating function.
PROPERTIES
• It assigns values to the rates and probabilities of failure, which in this way can only be positive.
• Since it depends on two parameters, as will be seen, it fits well to a large number of empirical
distributions.
• It is ideal for parameters that are themselves the product of numerous random quantities (multiple
effects that influence the reliability of a component).
• The mathematical expectation or mean in the log-normal distribution is greater than its median. In this
way, it gives more importance to large values of failure rates than a normal distribution with the same
5% and 50% percentiles, therefore tending to be pessimistic.
CHARACTERISTICS
• The log-normal distribution is obtained when the logarithms of a variable are described by a normal
distribution. It is the case in which variations in the reliability of the same class of technical components are
represented by considering the random failure rate 2 instead of a constant variable.
• It is the natural distribution to use when the deviations from the model value are formed by factors,
proportions or percentages rather than absolute values as is the case with the normal distribution.
• The log-normal distribution has two parameters: m (arithmetic mean of the log of the data or failure rate) i
(standard deviation of the log of the data or failure rate.)
APPLICATIONS
• Adjusts to certain types of failures (fatigue of metal components).
• Life of electrical insulation.
• Continuous processes (technical processes) and repair data.
• The log-normal distribution is one in which the logarithm of the variable is normally distributed. Therefore
we can obtain the probability density function of the normal distribution through the transformation.
• The log-normal distribution is important in representing proportional effects phenomena, such as those in
which a change in the variable at any point in a process is a random proportion of the previous value of
the variable. Some failures in the maintenance program fall into this category.
CONCLUSIONS
• The lognormal distribution is a probability used to express the behavior of observations with positive
skewness, where the majority of values occur in the vicinity of a minimum value.
• This distribution is characteristic in data sets where there is a greater frequency of small values, which is
why the mean moves to the right and this makes the best position statistician the mode and not the
arithmetic mean.
EXERCISE 1
The log-normal distribution is the best choice to describe the distribution of maximum well depth data of cast
iron pipes in the soil. Consider u=0.353 and o =0.754 for the maximum well depth (mm) of buried pipes. Find:
a) The mean and variance of the well depth.
b) The probability that the maximum depth is between 1 and 2mm.
Literal to
o2 n - (0.754) 2
E(X ) = e"t 2 =2 e°.353*+ 2 = e0.63726 = 1. 8913
V(X ) = e21+02 (e02—1) = ^ (0.353) (0.754)
. (e(0.754)2 -1
)= 2.7387
Literal b
P(1<X<2)= P[In(1) < In(X) < In(2)]
P(1<X <2) = P(0 < In(X) < 0.693)
0 - 0.353 0.693—0.353)
P(1<X< 0.754 <Z< 0.754 )
_ 0 0.693 - 0.353) ( 0—0.353)
P(1<X < PZ
■> = < 0.754 ) PZ < 0.754 )
P(1<X < 2) = 0(0.451) - O(-0.468)
.03 .114 .05 Z .IM .05 .06
AND
0.0 0.5120 0.5 1 60 0.5199 —0.4 0.3300 0.326+ 0.3228
0.1 0.5517 0.5557 0.5596 —0.3 0.3669 0.3632 0.3594
0.2 0.5910 0.5 948 0.5987 -0.2 0.4052 0.4013 0.3974
0.3 0.6293 0.6331 0.6368 —0.1 0 4+43 0.4404 0,4364
0.4 0.6664 0.6700 0.6736 _0.0 0.4840 0.4801 0.4761
P(1<X<2)= 0.6736 - 0.3228 = 0.3508
EXERCISE 2
It is known that historically the concentration of pollutants produced by chemical plants exhibits a behavior that
resembles a lognormal distribution.
This is important when considering issues related to compliance with government regulations. Suppose that the
concentration of a certain pollutant, in parts per million, has a lognormal distribution with the parameters μ = 3.2 and
σ = 1. What is the probability that the concentration exceeds 8 parts per million?
Solution:
p(X > 8) = 1 - P(X < 8)
Since ln( X ) has a normal distribution with mean μ = 3.2 and standard deviation σ = 1.
We use the symbol Φ to denote the cumulative distribution function of the standard normal distribution.
E .00 .01 .02
—1.4 0.0808 0.0793 0.0778 In (8) - 3.2
-13 0.0968 0.0951 0.0934 d(-1.12) = ^. 1314
0.1151 0 1 111 0.1112
—13
—1,1 CL 1357 0.1335 0.1314
_1.0 0.1 587 0,1562 0.1539
PROPOSED EXERCISE
The life, in thousands of miles, of a certain type of electronic control for locomotives has an approximately
lognormal distribution with μ = 5.149 and σ = 0.737.
Calculate the fifth percentile of the life of an electronic control like this.
