Al Akhawayn University at Ifrane

Department of Mathematics
School of Science and Engineering
Ifrane 53000, Morocco

Final Examination, Spring Semester 2023

Course Title: Engineering Probability and Statistics
Course Code: EGR 3303-01
Time: 13:00 to 14:50
Day: Wednesday, 15 May, 2024
Total Marks 35

SECTION: 1 ID: NAME:

Instructor: Anwar H. Joarder, [email protected]

Answer all Five (5) questions

Q Total Marks Marks Obtained Comment

1 6
2 6
3 11
4 6
5 6
35

1
Example 10.3.4A A farm-equipment manufacturer wants to compare the average daily downtime for
two sheet-metal stamping machines located in two different factories. Investigation of company records
for 25 randomly selected days on each of the two machines gave the following results:

Construct a 90% confidence interval estimate for the difference of the mean daily downtime.

2
11.3 Muzzle velocities of eight shells tested with a new gunpowder yield a sample mean of
3059 feet per second and a standard deviation of 39.4. The manufacturer claims that the new
gunpowder produces an average velocity exceeding 3000 feet per second. Does the sample
provide enough evidence to contradict the manufacturer's claim? Assume that the velocities
follow a normal distribution.

3
12.2 An Example of SLRA with Stress and Time

A study of stainless steel stress-corrosion cracking by potential measurements reported on the

relationship between applied stress in and time to fracture in hours, for 18-
8 stainless steel under uniaxial tensile stress in a 40% solution at . Ten different
settings of applied stress were used, and the resulting data were

Table 12.2a (Sample Data on Stress and Time to Fracture)

2.5 5 10 15 17.5 20 25 30 35 40

63 58 55 61 62 37 38 45 46 19

The following is a table for important parameters and statistics in the question:

Table 12.2b (Population and Sample Characteristics of Stress and Time to Fracture)

Mean Standard Deviation Intercept Slope Correlation

Population (rho)
(beta one)

Sample

All the mathematical results can be solved by formulae.

Table 12.2c (Descriptive Statistics of Independent Variable: Stress)

Some Minitab Printouts are given below:

Statistics

4
 Sum of
Variable Mean StDev Variance Squares Minimum Maximum

x 20.00 12.53 156.94 5412.50 2.50 40.00

Table 12.2d (Descriptive Statistics of Dependent Variable: Time to Fracture)

Statistics
Sum of
Variable Mean StDev Variance Squares Minimum Maximum

y 48.40 14.19 201.38 25238.00 19.00 63.00

Figure 12.2e (Scatterplot of Correlation Between Stress and Time to Temperature)

5
Figure 12.2f (CI for Pearson Correlation)

Click Stat > Basic Stat > Correlation > Graphs > Correlation & p-value to get

6
Figure 12.2g (Pearson Correlation with P-Value)

Table 12.2h (Two Sided Hypotheses on Correlation)

Correlation Domain

Population

Sample

Note that implies that the correlation between stress and time to fracture is insignificant.

Similarly, implies that the correlation between stress and time to fracture is significant.

Since the p-value is 0, we accept the alternative hypotheses

7
Table 12.2i (Left Sided Hypotheses on Correlation)
Correlation Domain

Population

Sample

Since the p-value is 0/2, we accept the alternative hypotheses

Figure 12.2j (Fitted Line Plot)

Table 12.2 k (Regression Equation)

Regression Equation
y = 66.42
- 0.901 x

Table 12.1l (Explanation of Slope)

8
 Serial

2 5

3 10

It is estimated that a 1-unit increase in x would result in unit change ('increase' if is

positive, and 'decrease' if is negative) in the expected value of the dependent variable. For
our example, it is estimated that if stress increases from 5 to 10, i.e., if it increases by 5, then it
would decrease time to fracture by In other word, the estimated average change in the
time to fracture per unit change in stress is

You may observe in the graph or in the above table that if the stress is 5 (2nd observation), the
estimated average time to fracture ( ) is more than the observed time to fracture ( )
making the error negative. Also observe in the graph or in the above table that if the stress is 10
(3rd observation), the estimated average time to fracture ( ) is more than the observed
time to fracture ( ) making the error negative.

Table 12.2m (Model Summary)

Model Summary
S R-sq R-sq(adj)

9.1243 63.25% 58.66%

1

Table 12.2n (Two Sided Hypotheses on Slope)

Slope Domain

Population

Sample

9
The hypotheses means that the slope, and consequently, the regression is

insignificant. Similarly, the hypotheses means that the slope, and consequently,
the regression is significant.

Table 12.2o (Two Sided Student Test on Slope)

Coefficients

Term Coef SE Coef T-Value P-Value VIF

Constant 66.42 5.65 11.76 0.000

x -0.901 0.243 -3.71 0.006 1.00

Table 12.2p (Left Sided Hypotheses on Slope)

Slope Domain

Population

Sample

Table 12.2q ( Two Sided Fisher Test on Slope)

Analysis of Variance
Source DF SS MS F P

Regression 1 1146.38 1146.3 13.77 0.006

8

Error 8 666.02 83.25

10
 Total 9 1812.40

Table 12.2s (CI & PI)

Table 12.2u (CI & PI Band)

Figure 12.2v (Normal Probability Plot)

11
Figure 12.2w (Residual-Fit Graph)

Report of the SLR Analysis:

The sample correlation coefficient between stress and time to fracture is negative

(See Figure 11.2e Scatter Plot). The true linear correlation coefficient between stress

and time to fracture is estimated to be (see Figure 11.2f). The 95% confidence

interval of linear correlation is (see Figure 11.2f) which does not include

0, hence we accept The p-value of 0.006 of the test of correlation (Figure 11.2g)

further corroborates the fact. This indicates that there is significant true linear correlation

between and time to fracture

The average time to fracture denoted by depends on the stress through the linear

equation It is estimated symbolically by or which for

our sample is (See Table 11.2k or 2nd column of Table 11.2o). That is, if the

12
stress level increases by 1 unit, then the average time to fracture decreases (because of

minus sign in the slope) by 0.901 unit. When the stress is 20, the time to fracture is

given by 37. Thus the error in estimating by so that the error is

It is also denoted by Show this error in the Figure 11.2j.

The 10 errors will add to zero.

Since the coefficient of determination is 0.6325 (see Table 11.2d) explaining that 63.25% of the

total variation in the (mean-centered squared) time to fracture is explained by that of the

stress

The sample Student statistic with 8 degrees of freedom (see the intersection

of 3nd row and 2th column of Table 11.2o) is given by …………..( see the intersection of

3nd row and 4th column of Table 11.2o) with probability value ( See
rd th
the intersection of 3 row and 5 column of Table 11.2o. Note that the probability value can be

calculated by Hence we accept i.e., we conclude the

significance of the rate of change of average time to fracture with change in stress

The hypotheses against can be tested by the probability value

( See the intersection of 3rd row and 5th column of

Table 11.2p). Hence we accept which indicates that the higher the stress level, the
lower the time to fracture.

13
The sample variance ratio statistic with 1 and 8 (See the intersection of row 2 and row 3,

column 2 of Table 11.2f) degrees of freedom is ( see the intersection of row 2 and

column 4 of Table 11.2r) with probability value (See the

intersection of row 2 and column 5 of Table 11.2r). Hence we accept i.e., we

conclude the significance of the rate of change of average time to fracture with

change in stress

When the stress level is 60, the average time to fracture is can be estimated by a
point 12.3646 and is called a point estimate, a fitted value, or simply fit. The 95% Confidence
Interval for the above average time to fracture is and the related
Prediction Interval is See Table 11.2s. We observe that the confidence
interval for is narrower and the related prediction interval is wider as always
expected.

The normality probability plot (Figure 11.2v) does not exhibit linearity well implying that the
errors are not normally distributed. The plot “Residuals Versus Fit” (Figure 11.2w) is also not
that scattered implying that the error variances are not constant.

13.1 Multiple Linear Regression Analysis of Pull Strength Depending

on Wirelength and height of the die
EXA 13.1A A sample on three variables collected in an observational study in a semiconductor
manufacturing plant. In this plant , the finished semi-conductor is wire-bonded to a frame. The
variables reported are pull strength ( a measure of the amount of force required to break the
bond), the wire length and the height of the die .

14
Serial

1 9.95 2 50

2 24.45 8 110

3 31.75 11 120

4 35.00 10 550

5 25.02 8 295

6 16.86 4 200

7 14.38 2 375

8 9.60 2 52

9 24.35 9 100

10 27.50 8 300

11 17.08 4 412

12 37.00 11 400

13 41.95 12 500

14 11.66 2 360

15 21.65 4 205

16 17.89 4 400

17 69.00 20 600

18 10.30 1 585

19 34.93 10 540

20 46.59 15 250

21 44.88 15 290

22 54.12 16 510

23 56.63 17 590

15
 24 22.13 6 100

25 21.15 5 400

Montgomery, D.C. and Runger, George C. (2014). Applied Statistics and Probability for
Engineers. Wiley, 6E pp 11-13, 472-494

Write a Report on the Pull strength depending on wire length and die height

Solution:

Since corresponding to regression is (See Row 2

and Column 6 of ANOVA Table) the regression model is significant at 5% level of significance.
That is, both the predictors together ( wire length and die height ) have significant effect
on the mean response ( pull strength). Since corresponding to the predictor wire
length is (Row 3 and Column 5 of Coefficients Table) it has a significant
effect on the mean pull length at 5% level of significance. Also the corresponding to the
predictor die height (Row 4 and Column 5 of ANOVA Table) is it has a
significant effect on the mean pull length at 5% level of significance.

16
14.1 Anova (Tensile Strength n=4)

A manufacturer of paper used for making grocery bags is interested in improving the
tensile strength of the product. Product engineering thinks that tensile strength is a
function of the hardwood concentration in the pulp and that the range of hardwood
concentrations (HC) of practical interest is between 5 and 20%. A team of engineers
responsible for the study decides to investigate four levels of hardwood concentration:
5%, 10%, 15%, and 20%. They decide to make up six test specimens at each
concentration level, using a pilot plant. All 24 specimens arc tested on a laboratory
tensile tester, in random order. The data from this experiment are shown in Table 14-1
below:

17
Table 14-1

HC Observations

1 2 3 4 5 6 Total Average

5 7 8 15 11 9 10 60 10

7 12 17 13 18 19 15 94 15.67

15 14 18 19 17 16 18 102 17

20 19 25 22 23 18 20 127 21.17

383 15.96

a. Test the hypothesis that the mean tensile strengths of the bags due to different hardwood
concentration are the same

Solution: Different from

One-way ANOVA: C2 versus C1

Method

Null hypothesis All means are equal

Alternative hypothesis At least one mean is different
Significance level α = 0.05

Equal variances were assumed for the analysis.

Factor Information

Factor Levels Values

C1 4 5, 10, 15, 20

Analysis of Variance

Source DF Adj SS Adj MS F-Value P-Value

C1 3 382.8 127.597 19.61 0.000
Error 20 130.2 6.508
Total 23 513.0

Solution by Probability Value Method: Since we reject

and accept Different from

18
b. Test that the effects due to due to different hardwood concentration are negligible

Solution: The same as (a).

f. Testing against

Solution: The Gosset test statistic with 20 degrees of freedom is given by

By the sample data and the Analysis of Variance Table, we have

The Gosset test statistic with 20 degrees of freedom has the sampling distribution

Distribution Plot
T, df=20

0.4

0.3
Density

0.2

0.1

0.025 0.025
0.0
-2.086 0 2.086
X

19
Decision by Percentile Method: Since the sample produces which is too large
compared to the 97.5th percentile or too small compared the 2.5th
percentile, we accept

Decision by Probability Value Method: Since the sample produces probability value
we accept

Fisher Pairwise Comparisons

Fisher Individual Tests for Differences of Means

Difference Difference SE of Adjusted

of Levels of Means Difference 95% CI T-Value P-Value
10 - 5 5.67 1.47 ( 2.59, 8.74) 3.85 0.001
15 - 5 7.00 1.47 ( 3.93, 10.07) 4.75 0.000
20 - 5 11.17 1.47 ( 8.09, 14.24) 7.58 0.000
15 - 10 1.33 1.47 (-1.74, 4.41) 0.91 0.376
20 - 10 5.50 1.47 ( 2.43, 8.57) 3.73 0.001
20 - 15 4.17 1.47 ( 1.09, 7.24) 2.83 0.010

CI Method: By “Fisher Pairwise Comparisons” we have ( 8.09, 14.24), i.e.,

Since the interval does not include 0, we reject accept null hypothesis
and accept

f. Testing against

Solution: The Gosset test statistic with 20 degrees of freedom is given by

By the sample data and the Analysis of Variance Table, we have

The Gosset test statistic with 20 degrees of freedom has the sampling distribution

20
 Distribution Plot
T, df=20

0.4

0.3
Density

0.2

0.1

0.025 0.025
0.0
-2.086 0 2.086
X

Decision by Percentile Method: Since the sample produces which is too large
compared to the 97.5th percentile or too small compared the 2.5th
percentile, we accept

Decision by Probability Value Method: Since the sample produces probability value
we accept

CI Method: By “Fisher Pairwise Comparisons” we have ( 8.09, 14.24), i.e.,

Since the interval does not include 0, we reject accept null hypothesis
and accept

21