Microsoft Word 14 Material Removal

 
 
 
 
 
 
 
 
 
Table 2: Design Matrix with response parameter MRR 
Experiment No 
Factors 
MRR(g) 
A 
B 
C 
1 
- 
- 
- 
0.0145 
2 
+ 
- 
- 
0.0159 
3 
- 
+ 
- 
0.0132 
4 
+ 
+ 
- 
0.0248 
5 
- 
- 
+ 
0.0222 
6 
+ 
- 
+ 
0.0003 
7 
- 
+ 
+ 
0.0004 
8 
+ 
+ 
+ 
0.0529 
9 
0 
0 
0 
0.0206 
10 
0 
0 
0 
0.0482 
11 
0 
0 
0 
0.0141 
12 
0 
0 
0 
0.0003 
 
 
3. EXPERT SYSTEMS 
 
A. Artificial Neural Network

Recent research activities in artificial neural networks (ANNs) have shown that ANNs 

have powerful pattern recognition and classification. ANNs are suited for problems 
which require knowledge that is difficult to specify but for which there are enough data 
or observations. They learn from the training data and capture information about 
relationships among the data even if underlying relationships are unknown or hard to 
Sample 
Machining 
Parameters 
Unit 
-1 
0 
1 
A 
Spindle 
speed 
Rpm 
1500 
2250 
3000 
B 
Feed rate 
mm/min 
30 
35 
40 
C 
Drill size 
Mm 
1 
2 
3 

---

International Journal of Mechanical Engineering and Technology (IJMET), ISSN 0976 – 
6340(Print), ISSN 0976 – 6359(Online) Volume 3, Issue 2, May-August (2012), © IAEME 
131 
describe. ANNs are universal approximations [3] it has been shown that a network can 
approximate any continuous function to any desired accuracy by many researchers [4], 
[5]. Neural network which uses back propagation algorithms for modeling has been 
developed using machining process parameters, spindle speed, feed rate and drill size as 
input parameters and material removal rate as output parameter. Making connections 
from the input layer to the output layer improves the learning efficiency. Out of the 
experimental data generated, the training data is used to train the model and this data is 
not used for testing and validation. 
The BP algorithm was developed by Paul Werbos in 1974 and rediscovered by 
Rumelhart and Parket. Since its rediscovery, the back propagation algorithm has been 
widely used as a learning algorithm in the feed forward multilayer neural network. The 
back –propagation algorithm is applied to feed forward ANN’s with one or more number 
of hidden layers as shown in Fig.1.Based on this algorithm, the network learns a 
distributed associative mapping between the input and output layers. What makes this 
algorithm different than the others is the process by which weights are calculated during 
the learning phase of the network. The input layer is used to feed the data in the network. 
The inputs are subsequently modified based on the interconnection weights between 
layers. The net input to each neuron from the preceding layer will be as: 
net
j=
∑
=
+
N
i
j
i
i
b
jx
w
1
Fig.1 Feed Forward Multilayer Perceptron 
Where, net
j
is the net input, N is the number of neurons of the inputs to the j
th
neuron in 
the hidden layer. W
ij
is the interconnected weight from the i
th
neuron in the forward layer 
to the j
th
neuron in the hidden layer, x
i
is the input from the i
th
neuron in the hidden layer 
and b
j 
is the bias value corresponding to the neuron. The output signal is obtained by 
applying activations to the net input. In this particular analysis, multilayer back 
propagation neural network model is developed using MATLAB neural network toolbox 
for predicting the material removal rate in drilling C/SiC material. There is one input 
vector with three elements. The value of the first element of the input vector ranges 
between 1500 to 3000 rpm, the values of the second element of the input vector ranges 
between 30-40 mm/min and the value of the third element of the input vector ranges 
between 1-3 mm. There are twenty neurons in the hidden layer and one neuron in the 
output layer. The number of neurons in the hidden layer was chosen based on the MSE 
value which is least between 20 to 25 nodes as shown in Fig.2 

---