Prediction of Mechanical Properties in Rotational Moulding …
7
Table 1 Oven residence time and corresponding mechanical properties [24]
S. No.
Oven residence time
(min)
Tensile strength
(MPa)
Impact strength (J)
Flexure strength
(MPa)
1
32
16
0.52
18.2
2
34
16.2
0.69
18
3
36
17.2
0.92
17.6
4
38
17.4
0.97
17
5
40
17.4
1
17.2
6
42
16
0.8
16.7
7
44
15.3
0.6
16
2.1 Training Data
The above data was taken from the experimental work done by Ramkumar [24].
Then from the selected model, hypothesis polynomial of the curve is obtained.
And further, from the hypothesis, polynomial values of mechanical properties were
obtained for different oven residence timings. Degree of hypothesis polynomial
changes according to the inputs and their corresponding outputs (Table 1).
3 Result and Discussion
From the two machine learning models, following prediction curves were obtained:
Figure 1 depicts the prediction curve obtained from linear regression model for
various mechanical properties for different oven residence timings. It can be observed
from the prediction curves that the input points deviate from the prediction curve
in a great manner which is not suited for the required machine learning prediction
model.
Figure 2 depicts the prediction curve obtained from polynomial regression model
for various mechanical properties for different oven residence timings. It can be
observed from the prediction curves that the deviation of input points from the
prediction curve is relatively smaller that the linear regression model (Fig. 3).
From the above observation, it can be inferred that polynomial regression model
gives more precise prediction of mechanical properties of roto moulded product.
This finding confirms that polynomial regression model is best suited for the
purpose of predicting the mechanical properties of the roto moulded product when
the oven residence time is varied.
Further from the polynomial regression prediction curve that was obtained from
polynomial regression machine learning prediction model, different polynomial
functions of the best suited degree were obtained that shows the variation of
mechanical properties with the change in oven residence timings.
7
Table 1 Oven residence time and corresponding mechanical properties [24]
S. No.
Oven residence time
(min)
Tensile strength
(MPa)
Impact strength (J)
Flexure strength
(MPa)
1
32
16
0.52
18.2
2
34
16.2
0.69
18
3
36
17.2
0.92
17.6
4
38
17.4
0.97
17
5
40
17.4
1
17.2
6
42
16
0.8
16.7
7
44
15.3
0.6
16
2.1 Training Data
The above data was taken from the experimental work done by Ramkumar [24].
Then from the selected model, hypothesis polynomial of the curve is obtained.
And further, from the hypothesis, polynomial values of mechanical properties were
obtained for different oven residence timings. Degree of hypothesis polynomial
changes according to the inputs and their corresponding outputs (Table 1).
3 Result and Discussion
From the two machine learning models, following prediction curves were obtained:
Figure 1 depicts the prediction curve obtained from linear regression model for
various mechanical properties for different oven residence timings. It can be observed
from the prediction curves that the input points deviate from the prediction curve
in a great manner which is not suited for the required machine learning prediction
model.
Figure 2 depicts the prediction curve obtained from polynomial regression model
for various mechanical properties for different oven residence timings. It can be
observed from the prediction curves that the deviation of input points from the
prediction curve is relatively smaller that the linear regression model (Fig. 3).
From the above observation, it can be inferred that polynomial regression model
gives more precise prediction of mechanical properties of roto moulded product.
This finding confirms that polynomial regression model is best suited for the
purpose of predicting the mechanical properties of the roto moulded product when
the oven residence time is varied.
Further from the polynomial regression prediction curve that was obtained from
polynomial regression machine learning prediction model, different polynomial
functions of the best suited degree were obtained that shows the variation of
mechanical properties with the change in oven residence timings.