Applied Engineering Statistics CEP

COMPLEX ENGINEERING PROBLEM

SUBJECT CODE

MA-241 AES

SUBMITTED TO

Dr. Muhammad Shabir

SUBMIITED BY

2023-ME-122

Muhammad Arslan Ali

DATE

31st May, 2025

1|P a ge
 Applied Engineering Statistics CEP

PART 1

Introduction to the Normal Distribution

The normal distribution, commonly referred to as the Gaussian distribution, constitutes one of
the most fundamental and extensively employed probability distributions in the domains of
statistics and data analytics. As a continuous probability distribution, it encapsulates the behavior
of a random variable that conforms to a symmetric, bell-shaped curve centered around its mean.
A defining characteristic of the normal distribution is that its mean, median, and mode coincide,
thereby ensuring perfect symmetry. The standard deviation governs the degree of dispersion or
spread of the data around the mean, offering a quantitative measure of variability. Due to its
mathematical elegance and empirical applicability, the normal distribution underpins a multitude
of inferential statistical methodologies.

Applications of the Normal Distribution Across Diverse Contexts

The normal distribution exhibits remarkable versatility, rendering it an indispensable tool across
a wide array of scientific, industrial, and social disciplines. Its utility stems from its ability to
approximate numerous naturally occurring phenomena. Common applications include:

1. Physical Measurements: Quantitative traits such as human height, body mass, and
systolic blood pressure frequently exhibit distributions that approximate the normal
curve, particularly within large populations.

2. Financial Modeling: In the field of finance, normal distributions are often employed—
albeit sometimes as idealized assumptions—to model asset returns, interest rate
fluctuations, and other stochastic financial processes.

3. Statistical Quality Control: In manufacturing and process engineering, the normal

distribution is pivotal for monitoring process stability, identifying variation, and detecting
anomalies in product quality.

4. Biomedical and Life Sciences: Variables such as reaction times, biochemical marker
levels, and drug absorption rates are often normally distributed, allowing for effective
statistical treatment in medical research.

5. Psychometrics: Intelligence quotient (IQ) scores and other standardized psychological

assessments are typically designed to follow a normal distribution, facilitating
comparative analysis across populations.

6. Behavioral and Social Sciences: Constructs such as personality traits, social attitudes,
and preference scales often yield normally distributed results when measured across
sufficiently large and diverse samples.

2|P a ge
 Applied Engineering Statistics CEP

7. Environmental Science: Climatic variables like ambient temperature, rainfall intensity,

and atmospheric particulate concentration are frequently modeled using the normal
distribution due to their tendency to cluster symmetrically around a mean.

The Empirical Rule: Data Dispersion Within Standard Deviations

An elegant feature of the normal distribution is encapsulated in the Empirical Rule, also known
as the 68-95-99.7 Rule, which delineates the proportion of data contained within specific
intervals of standard deviations from the mean:

 Approximately 68% of the observations reside within one standard deviation of the
mean.
 Around 95% of the data falls within two standard deviations.
 Nearly 99.7% of all values lie within three standard deviations.

These proportions are instrumental for probabilistic inference and are particularly useful in
assessing outliers and evaluating the significance of deviations from the mean. Though the
percentages are approximate, they offer robust practical guidance for interpreting normally
distributed datasets.

Utilizing Normal Distribution Tables (Z-Tables)

Normal distribution tables, colloquially known as z-tables, serve as essential statistical tools
for evaluating the cumulative probabilities associated with the standard normal distribution—
a normalized form of the normal distribution with a mean of 0 and a standard deviation of 1.

To utilize a z-table effectively, one must proceed through the following analytical steps:

1. Compute the Z-Score:

The z-score represents the standardized value of a given
observation and is calculated as:

𝑥−μ
𝑧=
σ
where x denotes the value of interest, μ is the mean, and σ is the standard deviation. This
transformation allows disparate normal distributions to be mapped onto the standard
normal curve.

3|P a ge
 Applied Engineering Statistics CEP

2. Consult the Z-Table:

Once the z-score is obtained, the corresponding cumulative
probability (i.e., the area under the curve to the left of the z-score) can be extracted from
the table.

3. Interpret the Probability:

1. For probabilities less than a given value, the z-table provides a direct answer.
2. For probabilities greater than a value, subtract the cumulative probability from 1.
3. For probabilities between two values, compute the z-scores for both and subtract
the smaller cumulative probability from the larger.

Z-tables are indispensable for solving a broad spectrum of statistical problems, including but not
limited to the construction of confidence intervals, the evaluation of hypothesis tests, and the
determination of critical thresholds. Their relevance persists even in the era of computational
software, particularly in foundational statistical education and manual problem-solving
scenarios.

Example:

Suppose we want to find the probability of getting a score of 80 or higher on an exam with a
mean of 70 and a standard deviation of 5. First, we need to convert the raw score of 80 to a z-
score:

80−70
z= =2
5
Then, we will look up the probability associated with a z-score of 2 in the normal distribution
table, which is 0.9772. This means that the probability of getting a score of 80 or higher is
0.9772, or 97.72%.

Approximating the Normal Distribution Using Other Distributions

The normal distribution is renowned for its mathematical tractability and empirical relevance
across a vast array of disciplines. Its symmetrical, bell-shaped form, defined mean and variance,
and the foundational Central Limit Theorem (CLT) make it a cornerstone in statistical theory
and practice. Nonetheless, in certain analytical contexts—particularly when direct application of
the normal model is infeasible—other probability distributions can be employed to approximate
it.A quintessential example is the Student’s t-distribution, which is particularly suited for
approximating the normal distribution in situations involving small sample sizes (typically n <
30) and unknown population standard deviation. Although the t-distribution resembles the
normal distribution in shape, it possesses heavier tails, thereby accounting for the additional
uncertainty inherent in small-sample estimation. As the sample size increases, the degrees of

4|P a ge
 Applied Engineering Statistics CEP

freedom grow, and the t-distribution converges asymptotically to the standard normal
distribution.

Conditions Under Which the Normal Distribution Serves as

an Approximation for Other Distributions

The appropriateness of using the normal distribution as an approximation for other distributions
is contingent upon several statistical conditions and structural properties of the underlying data:

Large Sample Size (n ≥ 30):

According to the Central Limit Theorem, the sampling
distribution of the sample mean (or proportion) tends toward normality as the sample size
increases, irrespective of the shape of the parent population. This principle justifies the use of the
normal approximation in many inferential procedures, including confidence interval estimation
and hypothesis testing, when dealing with sufficiently large samples.

Symmetry and Unimodality:

The normal distribution is inherently symmetric and unimodal,
peaking at the mean. If the target distribution exhibits a similar structure—i.e., is centered,
symmetric, and possesses a single mode—the normal approximation can often be applied with
minimal loss of accuracy.

Known Mean and Variance:

The presence of a known and finite mean (μ) and variance (σ²)
in the target distribution increases the feasibility of approximating it with a normal distribution.
For instance, the binomial distribution can be approximated by the normal distribution when
the number of trials is large and the probability of success p is neither very small nor very large
(i.e., both np and n(1 − p) are greater than 5 or 10).

Continuity of the Distribution:

Since the normal distribution is continuous, it serves as a more
accurate approximation when applied to other continuous distributions. In the case of discrete
distributions (such as the binomial or Poisson), the continuity correction—an adjustment by
±0.5—is often employed to enhance approximation accuracy.

Well-Behaved Data:
The suitability of a normal approximation deteriorates in the presence of
heavy skewness, multiple modes, or extreme outliers. Distributions with long tails,
pronounced asymmetry, or erratic behavior deviate significantly from the assumptions
underpinning normality and may necessitate alternative modeling strategies.

5|P a ge
 Applied Engineering Statistics CEP

Part B:
Problem Statement:

A machine produces a number of bolts which are measured are in millimeters

(mm). Bolts are measured accurately and any which are smaller than 3.5 mm or
bigger than 4.4 mm are rejected.

Activity 1: Generation of the Data

Generate 100 values using a random number generator:

4.1 3.8 3.9 4.3 4.2 4.0 3.7 3.8 4.4 4.2 4.3 4.0 3.9 3.6 3.7 4.1 3.6 4.0 4.3 4.1

3.9 4.3 4.0 3.7 3.8 3.9 4.2 4.3 4.4 3.6 4.1 4.2 4.0 3.9 4.4 3.7 3.8 4.0 3.7 3.9

4.3 4.2 3.9 4.4 3.5 4.2 4.3 4.1 3.6 4.1 4.3 4.0 3.8 4.3 4.1 3.9 4.0 4.4 4.2 4.0

4.3 4.2 3.9 4.1 3.8 3.7 3.9 3.5 4.4 4.3 3.6 4.1 4.3 4.4 3.8 4.0 4.2 4.3 4.1 3.6

3.8 4.2 3.9 3.7 4.3 4.4 3.6 4.1 4.2 4.0 3.9 4.3 4.1 4.0 3.7 3.9 4.2 3.8 4.4 3.5

Activity 2: Looking at Your Data

Frequency distribution table:

Diameter (mm) Frequency

3.5 3
3.6 7
3.7 8
3.8 8
3.9 13
4 12
4.1 12
4.2 11
4.3 15
4.4 9

6|P a ge
 Applied Engineering Statistics CEP

HISTOGRAM
16

14

12

10

0
3.5 3.6 3.7 3.8 3.9 4 4.1 4.2 4.3 4.4

As we can see, the histogram has a bell-shaped curve, which is typical of a normal distribution.
However, the distribution is not perfectly symmetrical since we only have a sample size of 100
bolts. To obtain a more symmetrical distribution, we would need to measure thousands of bolts.

Activity 3: Computation of Mean, Standard Deviation, Median & Mode

3.1 Mean:

To compute the mean and standard deviation of the data, we can use the following formulas:

𝑠𝑢𝑚 𝑜𝑓 𝑎𝑙𝑙 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠

𝑚𝑒𝑎𝑛 =
𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑜𝑏𝑠𝑒𝑟𝑣𝑎𝑡𝑖𝑜𝑛𝑠

3.5 × 3 + 3.6 × 7 + 3.7 × 8 + 3.8 × 8 + 3.9 × 13 + 4 × 12 + 4.1 × 12 + 4.2 × 11 + 4.3 × 15 + 4.4 × 9

𝑚𝑒𝑎𝑛 =
100

Mean = 3.939 mm

7|P a ge
 Applied Engineering Statistics CEP

3.2 Standard deviation:

𝑆 = 0.254

3.3 Median:

100 observations → median is the average of 50th and 51st values.

Diameter (mm) Comulative Frequency

3.5 3
3.6 10
3.7 18
3.8 26
3.9 39
4.0 51 (BOTH 50 & 51 FALL HERE)
4.1 63
…. ….

Both 50th and 51st values = 4.0 mm,

So,

Median = 4.0 mm

3.4 Mode:

Most frequent value is called Mode

 f = 15 at 4.3 mm
 Mode is 4.3 mm

8|P a ge
 Applied Engineering Statistics CEP

Summary of Statistics
Mean 3.939 mm
Standard Deviation 0.254 mm
Median 4.0 mm
Mode 4.3 mm

Activity 4 (Solution of the Problem):

To solve the problem, we need to find the proportion of bolts that have a diameter smaller than
3.5 mm or bigger than 4.4 mm. We'll use the z-scores and the standard normal distribution.

Z-score for 3.5 mm:

3.5−3.939
𝑧= ≈ −1.73
0.254
Z-score for 4.4 mm:

4.4−3.939
𝑧= ≈ 1.82
0.254
Now, using the standard normal distribution, we find the probabilities:

 Probability of a bolt being smaller than 3.5 mm P(X<3.5)

≈ 0.04175

 Probability of a bolt being bigger than 4.4 mm P(X>4.4)

≈ 0.03457

The total proportion of rejected bolts is the sum of these probabilities:

0.04175+0.03457 = 0.07632

Number of bolts rejected

= 5000×0.07632 ≈ 382
Therefore, out of a batch of 5000 bolts, approximately 382 bolts would be rejected due to their
diameter being smaller than 3.5 mm or bigger than 4.4 mm.

9|P a ge