International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
853
AN APPLICATION OF ARTIFICIAL NEURAL NETWORK TO
PREDICT SURFACE ROUGHNESS DURING DRILLING OF AISI 1020
STEEL
Abhinaw Roy & Mr. Kulbhushan Bhagat
Student, Suresh Gyan Vihar University, Jaipur Rajasthan
Assistant Professor, Suresh Gyan Vihar University, Jaipur, Rajasthan
ABSTRACT
Metal machining has been an interesting topic of
research for decades. Early research efforts include
topics, such as, chip formation, heat generation at tool-
chip and tool-work interfaces, tool wear, tool life,
cutting force analysis, machinability, power
requirements, manufacturing cost and productivity etc.
With more emphasis on quality of surface finish with
high production rate, in past few decades, research
efforts seem to have been diverted towards the study of
surface finish/roughness in metal machining.
The objectives of the present study is the development
of the surface roughness prediction model using
Artificial neural network in terms of feed, cutting
speed and depth of hole during the drilling of AISI
1020 steel and to Investigate the effect of drilling
parameters on surface roughness using 3D plots
between the parameters and surface roughness .
In the present research a multilayer feed-forward
neural network has been employed using MATLAB
software. The input layer has four neurons
corresponding to each of the three input variables
(cutting speed, feed and depth of cut) and one neuron
in the output layer corresponding to output variable
(surface roughness). The number of hidden layers and
neurons in hidden layers has been determined on the
basis of trial and error method. The mean square error
has been selected as performance function. The ANN
training simulation was carried out using batch
gradient descent with variable learning rateprocedure
“traingdx”. This procedure helps in improving the
performance of back -propagation algorithm by
allowing the learning rate to change during the training
process based on the complexity of the local error
surface, In this research, a portable surface roughness
tester has been used to measure surface roughness
indicators of finished work pieces.
Keywords: Artificial neural network, surface
roughness, AISI 1020 steel.
I. INTRODUCTION
In the manufacturing industries, various machining
processes (turning, milling, drilling etc.) are adopted
for removing the material from the workpiece to obtain
finished product. Among the various metal removing
www.ijsret.org
Volume 4, Issue 8, August 2015
processes, drilling is the one of most important metal
removing process as compared to other traditional
machining processes. Drilling is use for making the
hole in the workpieces. Hole making is a most
important machining process in manufacturing. During
the drilling, the drill rotates and fed into the work piece
to remove the material in the form of chips that move
along the fluted shank of the drill.Different drilling
tools and hole making methods are used for drilling.
The selection of different tools and methods depends
on the type of workpiece, size of the hole, the quantity
of holes, and the quantity of the holes in given time
periods. Also, drilling is used as an initial process for
many machining operations, such as reaming, tapping
and boring. Drilling process is widely used in the
aerospace, aircraft and automotive industries. Now
days, due to global competitiveness, manufacturing
industries are more concerned about the quality of their
products with high metal removal rate. These focus on
producing high quality products in time at minimum
cost.
The surface roughness is considered to be a measure of
the technological quality of a product. Surface
roughness is the one of the critical performance
parameter that has an appreciable effect on several
mechanical properties of machined parts such as
fatigue behaviour, corrosion resistance, creep life, etc.
It also affects other functional attributes of machined
parts like friction, wear, light reflection, heat
transmission, lubrication, electrical conductivity, etc.
Hence, achieving the desired surface quality is of great
importance for the functional behaviour of the
mechanical parts.
During the drilling of the workpiece, it has long been
recognized that the drilling conditions (drill point
geometry, drill and workpiece materials, drilling
parameters like feed rate and spindle speed) affect the
performance of the operation to a greater extent. These
drilling conditions should be selected to optimize the
economics of drilling operations. So it can be achieved
by empirical modeling of performance as a function of
machining conditions using artificial neural network.
The proposed work will be employed for investigate
the effect of drilling parameters on surface roughness
and development of surface roughness prediction
model using artificial neural network (ANN).
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
854
Volume 4, Issue 8, August 2015

“traingdx”. This procedure helps in improving the

performance of back -propagation algorithm by
allowing the learning rate to change during the training
process based on the complexity of the local error
surface.

V. EXPERIMENTAL WORK
The process parameters that were chosen for
experimentation are given as under:
1. Spindle speed (RPM)
2. Feed (mm/rev.)
3. Depth of hole (mm)

MACHINE USED:
Fig 1: Drilling process The drilling operations have been carried out on a
CNC MILL MT250 Machining Center
II. OBJECTIVES
 Development of the surface roughness
prediction model using Artificial neural
network in terms of feed, cutting speed and
depth of hole during the drilling of AISI 1020
steel.
 Investigating the effect of drilling parameters
on surface roughness using 3D plots between
the parameters and surface roughness.
 To investigate the most significant parameters
that affects the surface roughness.

III. ARTIFICIAL NEURAL NETWORK

Artificial neural network (ANN) is a most attractive
branch of artificial intelligence. It is based on the basic
model of the human brain with the capability of
generalization and learning. An ANN has the potential
to handle problems such as modeling, estimating,
prediction, optimization, diagnosis, and adaptive
control in complex nonlinear systems. It can acquire,
store, and utilize knowledge gained from experience.
The capabilities of ANNs in capturing the
mathematical mapping between input variables and
output features are of primary significance for
modeling machining processes.
Fig 2: CNC MILL MT250 Machining Center used for
drilling
CUTTING INSERTS:
Coated carbide tool performs better than uncoated
carbide tools. On this basis, commercially accessible
K- series, tungsten carbide high speed core drills with
Tialn coating having 10 mm diameter with two flute
has been used for experimentation.

IV. PROPOSED NEURAL NETWORK

MODELS
In the present research a multilayer feed-forward
neural network has been employed using MATLAB
software. The input layer has four neurons
corresponding to each of the three input variables Fig 3: Tungsten carbide high speed core drills
(cutting speed, feed and depth of cut) and one neuron
in the output layer corresponding to output variable COOLANT:
(surface roughness). The number of hidden layers and Coolant has been used in all the experiments.
neurons in hidden layers has been determined on the SUPERCUT cutting oil by SHELL COMPANY has
basis of trial and error method. The mean square error been used in the ratio of 20:1 i.e 20 litres of water and
has been selected as performance function. The ANN 1 litre of cutting oil in it.
training simulation was carried out using batch S/R MEASUREMENT
gradient descent with variable learning rateprocedure

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
855
Volume 4, Issue 8, August 2015

Surface roughness is defined as the finer irregularities 5 600 0.01 20 1.021 0.25258
of the surface texture that usually result from the 6 300 0.02 10 1.207 0.37232
inherent action of the machining process or material 7 600 0.03 10 1.133 0.32468
condition. There are many parameters used related to 8 600 0.02 10 1.011 0.24614
surface roughness in literatures. In this research, a 9 300 0.01 15 1.301 0.43283
portable surface roughness tester (Model No TR 210 10 900 0.01 15 1.311 0.43927
manufactured by Beijing TIME High Technology Ltd. 11 600 0.03 20 1.527 0.57833
12 300 0.01 20 1.485 0.55129
Beijing City, China) has been used to measure surface
13 600 0.02 20 1.229 0.38648
roughness indicators of finished work pieces.
14 900 0.02 20 1.782 0.74249
15 900 0.03 10 1.608 0.63047
16 900 0.03 20 2.182 1.00000
17 900 0.03 15 1.811 0.76116
18 600 0.03 15 1.342 0.45923
19 300 0.02 15 1.474 0.54421
Testing data set
20 900 0.02 10 1.259 0.40579
Fig 4: surface roughness tester 21 300 0.03 20 1.972 0.86481
22 300 0.01 10 1.182 0.35622
FORMATION OF DESIGN MATRIX: 23 600 0.02 15 1.199 0.36717
S.n Spindle feed Depth Surface Validation data set
speed of hole roughnes 24 300 0.03 10 1.54 0.58670
1 600 0.01 15 0.982 25 900 0.01 20 1.438 0.52103
2 300 0.02 20 1.653 26 600 0.01 10 0.784 0.10000
3 900 0.01 10 1.118 27 300 0.03 15 1.774 0.73734
4 900 0.02 15 1.498
5 600 0.01 20 1.021 Neural network design and training:
6 300 0.02 10 1.207 The simulation for ANN training was conducted using
7 600 0.03 10 1.133 batch gradient descent with variable learning
8 600 0.02 10 1.011 rateprocedure “traingdx”. The trial and error method
9 300 0.01 15 1.301 was used for values of training parameters for
10 900 0.01 15 1.311 “traingdx” and is shown in table.
11 600 0.03 20 1.527 Training parameters Values
12 300 0.01 20 1.485
Network configuration 4-11-1
13 600 0.02 20 1.229
14 900 0.02 20 1.782 Number of hidden layer 1
15 900 0.03 10 1.608 Maximum number of 10000
16 900 0.03 20 2.182 epochs to train
17 900 0.03 15 1.811
18 600 0.03 15 1.342 Performance goal 0.00001
19 300 0.02 15 1.474 Learning rate 0.03
20 900 0.02 10 1.259
21 300 0.03 20 1.972 Ratio to increase learning 1.001
22 300 0.01 10 1.182 rate
23 600 0.02 15 1.199 Maximum validation 4
24 300 0.03 10 1.54 failures
25 900 0.01 20 1.438 Momentum constant 0.43
26 600 0.01 10 0.784 Minimum performance 1.00E-04
27 300 0.03 15 1.774 gradient
Number of patterns used 19
for training
DEVELOPMENT OF ANN MODEL FOR Number of patterns used 4
SURFACE ROUGHNESS: for testing
SN Spindle feed Depth S/R NOR Number of patterns used 4
speed Of S/R for validation
hole
Training data sets TRAINING OF MODEL:
1 600 0.01 15 0.982 0.22747
2 300 0.02 20 1.653 0.65944
3 900 0.01 10 1.118 0.31502
4 900 0.02 15 1.498 0.55966

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
856
Volume 4, Issue 8, August 2015

VALIDATION OF MODEL:

Fig 5: Correlation between the experimental (Target)

and predicted (Output) surface roughness in training.

Fig 6: Comparison of the experimental and predicted

values of surface roughness in training. Fig 9: Comparison of the experimental and predicted
values of surface roughness in validation
TESTING OF MODEL:
Performance of the network:

Fig 7: Correlation between the target and output

surface roughness in testing.

Fig 10: Correlation between the experimental (Target)

and predicted (Output) surface roughness for entire
data set

VALUES OF PREDICTED SURFACE

ROUGHNESS:
S.n Spindle feed Depth EXP. PRD.
speed of hole S/R S/R
1 600 0.01 15 0.982 0.926
2 300 0.02 20 1.653 1.699
3 900 0.01 10 1.118 1.093
Figure 8: Comparison of the experimental and 4 900 0.02 15 1.498 1.503
predicted values of surface roughness in testing. 5 600 0.01 20 1.021 1.044

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
857
Volume 4, Issue 8, August 2015

6 300 0.02 10 1.207 1.237

7 600 0.03 10 1.133 1.077
8 600 0.02 10 1.011 1.021
9 300 0.01 15 1.301 1.361
10 900 0.01 15 1.311 1.295
11 600 0.03 20 1.527 1.574
12 300 0.01 20 1.485 1.467
13 600 0.02 20 1.229 1.272
14 900 0.02 20 1.782 1.742
15 900 0.03 10 1.608 1.613
16 900 0.03 20 2.182 2.156
17 900 0.03 15 1.811 1.841
18 600 0.03 15 1.342 1.334 Fig 13: Plot between depth of hole and
19 300 0.02 15 1.474 1.403 surfaceroughness
20 900 0.02 10 1.259 1.271
21 300 0.03 20 1.972 1.886
22 300 0.01 10 1.182 1.116
23 600 0.02 15 1.199 1.207
24 300 0.03 10 1.54 1.568
25 900 0.01 20 1.438 1.480
26 600 0.01 10 0.784 0.839
27 300 0.03 15 1.774 1.794

EFFECT OF DRILLING
PARAMETERS ON S/R:

Fig 14: 3D plot for surface roughness between spindle

speed and feed

Fig 11: Plot between spindle speeds on surface

roughness Fig 15: 3D plot for surface roughness between spindle
speed and depth of hole.

Fig 12: Plot between feed on surface roughness

Fig: 3D plot for surface roughness between feed and

depth of hole

www.ijsret.org
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
VI. CONCLUSION & FUTURE SCOPE
The objective of the present work is to optimize
drilling parameters for minimum surface roughness.
An attempt has also been made to investigate the
effects of the drilling parameters on surface roughness.
The important conclusions drawn from the present
work are summarized as follows:
1. The predicted values of surface roughness
using proposed ANN model was found to be in close
agreement with the experimental data.
2. The results so obtained clearly illustrates that
the as the spindle speed increase the surface roughness
decreases up to a certain value of spindle speed
afterward roughness increases with increase in spindle
speed
3. The surface roughness increases with increase
in feed.
4. The surface roughness increases with increases
in depth of hole.
5. The minimum surface roughness is achieved at
low level of feed, low level of depth of hole and
middle level of spindle speed.
REFERENCESES
[1] Ahamed, A.R., Asokan, P., Aravindan, S. and
Prakash, M.K. (2010). Drilling of hybrid Al-5%SiCp-
5%B4Cp metal matrix composites, International
Journal of Advanced Manufacturing Technology,
49(9-12), 871-877.
[2] Bahce, E. and Ozel, C. (2013). experimental
investigation of the effect of machining parameters on
the surface roughness and the formation of built up
edge (BUE) in the drilling of Al 5005, Tribology in
Engineering, InTech, DOI: 10.5772/56027.
[3] Bai, B. H., Kumar, G.V. and Babu, K.A.(2014).
Optimization of surface roughness and burr height in
drilling of AL-6061 under MQL condition using DFA,
Mechanica Confab, 3(4), 10-17.
[4] Cicek, A. (2007), Step based geometric and
topological similarity assessment of mechanical parts,
Mathematical and Computational Applications, 12(3),
141-150.
[5] Cicek, A., Kivak, T. and Samtas, G. (2012).
Application of Taguchi method for surface roughness
and roundness error in drilling of AISI 316 Stainless
Steel, Journal of Mechanical Engineering, 58(3), 122-
129.
[6] Davim J.P. and Reis, P. (2003). Study delamination
in drilling carbon fiber reinforced plastics (CFRP)
using design experiments. University of Aveiro
Campus Santiago, Portugal. 59, 481-487.
[7] Gaitonde V.N. Karnik S.R., Acyutha B.T. and
Siddeswarappa B. (2007). Methodology of Taguchi
optimization for multi-objective drilling problem to
minimize burr size. International J. Advance
Manufacturing Technology, 34, 1-8.
www.ijsret.org
858
Volume 4, Issue 8, August 2015
[8] Gul T. and Muratoglu M. (2004). The drilling of
Al/SiCp metal-matrix composites. Part II: wokpiece
surface intergrity. Firat University, Turkey. 64, 1413-
1418.
[9] Hsu, I. and Tsao C.C. (2008). Optimization of
process parameter in drilling composite materials by
Grey-Taguchi method. Tahua Insitute of Technology,
Taiwan.
[10] Kaneriya, N.G. and Sharma, G.K.
(2014).Experimental evaluation and optimization of
dry drilling parameters of AISI304 austenitic stainless
steel using different twist drills,5th International &
26th All India Manufacturing Technology, Design and
Research Conference (AIMTDR 2014) December
12th–14th, 2014, IIT Guwahati, Assam, India, 1-7.
[11] Karnik S.R., Gaitonde V.N. and Davim J.P.
(2008). Acomparative study of the ANN and RSM
modelling approaches for predicting burr size in
drilling. International J. Advance Manufacturing
Technology, 38, 868-883.
[12] Khan, I., Luthra, D., Verma, M. and Singh,
S.(2012). Selection of optimum drilling parameter in
drilling of commercial acrylic sheet to achieve
minimum hole expansion by using Taguchi approach,
International Journal of Engineering Science and
Technology 4(4), 1256-1261.
[13] Khashaba, U.A., Seif, M.A. and Elhamid, M.A.
(2007). Drilling analysis of chopped composites,
Composites Part A, Applied Science and
Manufacturing, 38(1), 61-70.
[14] Kivak, T., Habali, K.and Seker, U. (2012). The
effect of cutting paramaters on the hole quality and
tool wear during the drilling of Inconel 718, Gazi
University Journal of Science, 25(2),533-540.
[15] Kumar, D., Singh, L.P. and Singh, G.(2012).
Operational modeling for optimizing surface
roughness in mild steel drilling using Taguchi
technique, International Journal of Research in
Management. 2(3) 66-77.
[16] Kurt M., Kaynak Y., Bakır B., KokluU., Atakok
G. and Kutlu L. (2009). Experimental investigation
and Taguchi optimization for the effect of cutting
parameters on the drilling of Al 2024-T4 alloy with
diamond like carbon (DLC) coated Drills, Sixth
International Advanced Technologies Symposium
(IATS’09), Karabuk, Turkey.
[17] Kurt, M., Bagci, E. and Kaynak, Y. (2009).
Application of Taguchi methods in the optimization of
cutting parameters for surface finish and hole diameter
accuracy in dry drilling processes. International J.
Advance Manufacturing Technology, 40, 458-469
[18] Lin, S.C. and Chen, I.K.(1999). Drilling of carbon
fiber-reinforced composite material at high speed,
Wear, 194, 156–162.
[19] Madiwal,S. (2006). Analysis of surface finish in
milling of composite using neural Networks, Wichita
State University.2006.
 International Journal of Scientific Research Engineering & Technology (IJSRET), ISSN 2278 – 0882
859
Volume 4, Issue 8, August 2015

[20] Mohan, N.S., Kulkarni, S.M. and Ramachandra,

A. (2007). Delamination analysis in drilling process of
GFRP composite materials. J. Mater. Process.
Technol., 186, 265-271
[21] Navanth, A. and Sharma,K. (2013). A study of
TAGUCHI method based optimization of drilling
parameter in dry drilling of AL 2014 alloy at low
speeds, International Journal of Engineering Sciences
& Emerging Technologies, 6(1) 65-75.
[22] Neseli, S.(2014). Optimization of process
parameters with minimum thrust force and torque in
drilling operation using Taguchi method, Advances in
Mechanical Engineering, 2014, 1-10.
[23] Noorul Haq A., Marimuthu P. and Jeyapaul J.
(2008). Multi response optimizatio of machining
parameters of drilling Al/SiC metal matrix composite
using grey relational analysis in the Taguchi method.
International J. Advance Manufacturing Technology.
37, 250-255.
[24] Nouari M., List G., Girot F. and Gehin D. (2005).
Effect of machining parameters and coating on wear
mechanisms in dry drilling of aluminum alloys.
International Journal of Machine Tools &
Manufacture. 45, 1436–1442.
[25] Shunmugesh, K., Panneerselvam, K. and Thomas,
J. (2014). Optimising drilling parameters of GFRP by
using grey relational analysis, International Journal of
Research in Engineering and Technology,3(6) 302-
305.
[26] Siddiquee, A.N., Khan, Z.A., Goel, P. Kumar,
M., Agarwal, G. and Khan, N.Z.(2014). Optimization
of deep drilling process parameters of AISI 321 steel
using Taguchi method, Procedia Materials Science, 6,
1217 – 1225.
[27] Sundeep, M., Sudhahar, M., Kannan, T. T. M.,
Kumar,V. and Parthipan,N. (2014).Optimization of
drilling parameters on austenitic stainless steel (AISI
316) using TAGUCHI'S methodology, International
Journal of Mechanical Engineering and Robotics
Research, 3(4),388-394.
[28] Tamilselvan and Raguraj (2014). Optimization of
process parameters of drilling in Ti-Tib composites
using Taguchi technique, International Journal on
Mechanical Engineering and Robotics, 2(4), 1-5.
[29]Tosun, N. (2006). Determination of optimum
parameters for multi-performance characteristics in
drilling by using grey relational analysis. International
J. Advance Manufacturing Technology. 28, 450-455.
[30] Tsao C.C. (2006). Taguchi analysis of drilling
quality associated with core drill in drilling of
composite material. International J. Advance
Manufacturing Technology,32, 877-884.
[31] Tyagi, Y., Chaturvedi, V. and Vimal, J. (2012).
Parametric optimization of drilling machining process
using Taguchi design and ANOVA approach, Journal
of Emerging Technology and Advanced Engineering,
2 (7), 339-347.

www.ijsret.org