Вы находитесь на странице: 1из 10

Brandeice Barrett- 1201539

University of Technology, Jamaica


Program: Masters of Philosophy in Chemical Engineering
Name: Brandeice Barrett
ID #: 1201539
Course: Design of Experiments-Semester 2 Test
Date: March 24, 2018
Brandeice Barrett- 1201539

Question 1
a) I there any indication that the fluids differ at 5% level of significance?
When the ANOVA produces a significant result, this is only indicating that there is at least one
mean which differs from the others and not exactly which group. In order to analyze the
differences between the means and to identify which group is different, the ANOVA is often
followed by specific comparisons. This can be carried out using a pairwise comparison method
called Fisher LSD. The analysis was carried out using One Way ANOVA in the SPSS software,

Table 1 showing Pairwise Comparisons using the Fisher LSD Method


Dependent Variable: Life (h) at 45 KV Load

(I) Fluid Type (J) Fluid Type Mean Difference Std. Error Sig.b 95% Confidence Interval for
(I-J) Differenceb

Lower Bound Upper Bound

2 .700 1.049 .512 -1.488 2.888

1 3 -2.300* 1.049 .040 -4.488 -.112

4 -.167 1.049 .875 -2.354 2.021


1 -.700 1.049 .512 -2.888 1.488
2 3 -3.000* 1.049 .010 -5.188 -.812
4 -.867 1.049 .418 -3.054 1.321
1 2.300* 1.049 .040 .112 4.488
3 2 3.000* 1.049 .010 .812 5.188
4 2.133 1.049 .055 -.054 4.321
1 .167 1.049 .875 -2.021 2.354

4 2 .867 1.049 .418 -1.321 3.054

3 -2.133 1.049 .055 -4.321 .054

Based on estimated marginal means


*. The mean difference is significant at the .05 level.
b. Adjustment for multiple comparisons: Least Significant Difference (equivalent to no adjustments).

Based on the comparison of the means using the LSD method there is indication that fluid 3
differs significantly from fluid 1 and fluid 2. This can be concluded because there is a significant
difference in the mean for fluid life for fluid 3 when compared life for fluids 1 and 2. The p value
for mean difference between fluid 3 and 1 was 0.04 (p=0.040) and that of 3 and 2 p=0.01, both of
which are less than 0.05 at the significance level of 5%.
Brandeice Barrett- 1201539

b) Which fluid would you select, given that the objective is long life?

The choice of fluid would be dependent on the fluid having the longest life. This information
can be obtained from the descriptive statistics graph as well as the mean plot.

Table 2 showing the Descriptive Statistics obtained from SPSS


Life (h) at 45 KV Load

N Mean Std. Deviation Std. Error 95% Confidence Interval for Mean Minimum Maximum

Lower Bound Upper Bound

1 6 18.6500 1.95218 .79697 16.6013 20.6987 16.30 21.60


2 6 17.9500 1.85445 .75708 16.0039 19.8961 15.30 20.30
3 6 20.9500 1.87910 .76714 18.9780 22.9220 18.50 23.60
4 6 18.8167 1.55488 .63478 17.1849 20.4484 16.90 21.10
Total 24 19.0917 2.04470 .41737 18.2283 19.9551 15.30 23.60

Figure 1 showing the mean plot of life versus fluid type

From the table and the graph it can be seen that fluid 3 would be selected as it had the longest
mean life of 20.95 hours at 45KV (please see graph and table above).
Brandeice Barrett- 1201539

c) Does the residual support the basic assumptions of analysis?

The normal probability plot of residuals to verify the assumption that the residuals are normally
distributed. From the plot above it can be seen that the points are normally distributed as they are
all close to the diagonal. Therefore the assumption of normal is distribution is met.
Brandeice Barrett- 1201539

Figure 2 showing the Scatter Plot showing the distribution of variance

The residuals versus fluid plot is used to verify the assumption that the residuals are randomly
distributed and have constant variance. Ideally, the points should fall randomly on both sides of
0, with no recognizable patterns in the points. From the plot above it can be seen that the
residuals are randomly distributed on both sides of 0. Therefore it can be concluded that the
assumption of constant variance or variance equality is also met.

Both residual plots support the basic assumption of ANOVA which include normal distribution
of the points and variance equality.
Brandeice Barrett- 1201539

Question 2
a) Is there any evidence that the stirring rate affects grain size at the 5 % level of
significance?

To test whether or not the stirring rate affects the grain size, an analysis of variance
(ANOVA) test is carried out. The output of ANOVA obtained from the SPSS software is
presented below.

Table 3 showing Tests of Between-Subjects Effects


Dependent Variable: Number of grain size

Source Type III Sum of df Mean Square F Sig.


Squares

Corrected Model 423.375a 6 70.563 16.907 .000


Intercept 3164.063 1 3164.063 758.111 .000
SR 367.188 3 122.396 29.326 .000
Furnace 56.187 3 18.729 4.488 .035
Error 37.563 9 4.174
Total 3625.000 16
Corrected Total 460.938 15

a. R Squared = .919 (Adjusted R Squared = .864)

Yes there is evidence that the Stirring Rate (SR) affects the grain size at the 5% level of
significance. This can be seen in the resulting p value which is less than 0.05 this is
indicating that the stirring rate does have an effect on the grain size at the 5 % level of
significance. The null hypothesis is rejected as the probability that the grain size occurred
due to chance is less than 5% (the probability is too low).

b) Is there evidence that the furnace affects the grain size at the 5 % level of
significance?

Yes there is evidence that the furnace affects the grain size at the 5% level of
significance. This can be seen in the resulting p value (p = 0.035) which is less than 0.05
this is indicating that the furnace does have an effect on the grain size at the 5 % level of
significance. The null hypothesis is rejected as the probability that the grain size occurred
due to chance is 3.5 % (the probability is too low).
Brandeice Barrett- 1201539

c) What should the process engineers recommend concerning the choice of stirring
rate and furnace for this grain refiner if small grain size is desirable?

The means plot above illustrates the grain sizes produced based on the furnace and stirring rate.
From the plot it can be seen that if smallest grain size is desired the process engineers should
recommend using furnace 2 at a stirring rate of 5 rpm.
Brandeice Barrett- 1201539

Question 3

a) Is there any evidence that the brand of battery affects the life in hours at 5% level of
significance?

To test whether or not the brand of battery affects the life of the battery, an analysis of
variance (ANOVA) test is carried out. The output of ANOVA obtained from the Minitab 18
software is presented below.

Table showing ANOVA results

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


Model 5 23.3275 4.6655 54.36 0.000
Linear 3 23.2458 7.7486 90.28 0.000
Brand of Battery 1 0.8008 0.8008 9.33 0.022
Device 2 22.4450 11.2225 130.75 0.000
2-Way Interactions 2 0.0817 0.0408 0.48 0.643
Brand of Battery*Device 2 0.0817 0.0408 0.48 0.643
Error 6 0.5150 0.0858
Total 11 23.8425

Yes there is evidence that the brand of battery affects the life in hours at the 5% level of
significance. This can be seen in the resulting p value (p=0.022) which is less than 0.05
this is indicating that the brand of battery does have an effect on the life of the battery at
the 5 % level of significance. The null hypothesis is rejected as the probability that the
life occurred due to chance is 2.2 % (the probability is too low).

b) Is there any evidence that the device affects the life in hours at 5% level of
significance?

Yes there is evidence that the device the life in hours at the 5% level of significance. This
can be seen in the resulting p value (p<0.01) which is less than 0.05 this is indicating that
the device does have an effect on the life of the battery at the 5 % level of significance.
The null hypothesis is rejected as the probability that the life occurred due to chance is
less than 5% (the probability is too low).
Brandeice Barrett- 1201539

c) Is there any interaction effect at 5 % level of significance?


To determine if the there are any interaction effect at the 5% significance level, the
ANOVA table is again referred and a interaction plot can also be generated.

Table showing ANOVA results


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

Model 5 23.3275 4.6655 54.36 0.000

Linear 3 23.2458 7.7486 90.28 0.000


Brand of Battery 1 0.8008 0.8008 9.33 0.022

Device 2 22.4450 11.2225 130.75 0.000

2-Way Interactions 2 0.0817 0.0408 0.48 0.643

Brand of Battery*Device 2 0.0817 0.0408 0.48 0.643


Error 6 0.5150 0.0858

Total 11 23.8425

From the table the interaction between Brand of battery and device resulted a p value of
0.643 which is greater than 0.05. This would suggest that the interaction between these
two variables does not create an effect on the life of the battery at the significance level
of 5%.

The interaction plot above also supports the data obtained from the ANOVA table that
indicates that there is no interaction effect at the 5% level of significance. An interaction
effect would have been indicated by intersecting line(s).
Brandeice Barrett- 1201539

d) What brand of battery would you recommend?

Contour Plot of Life (h) vs Device, Brand of Battery


3.0
Life
(h)
< 6
6 – 7
2.5 7 – 8
8 – 9
> 9
Device

2.0

1.5

1.0
1.0 1.2 1.4 1.6 1.8 2.0
Brand of Battery

** Brand A = 1
** Brand B= 2
*** Devices: 1 =Radio 2= Camera 3= DVD Player

The contour plot above shows the life of the battery in colour bands based on the relationship
with the brand of battery and devices. The plot is based on the model generated from the Minitab
18 analysis of the data. The plot reveals that the Brand B of battery has the highest life ( >9
hours)

This is also observed in the main effect graph where brand B is seen to have a greater mean life
than brand A. Therefore I would recommend this brand of battery.

Вам также может понравиться