MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp.

(55-61)
ISSN No. 2230 - 7680 MIT Publications

55

Surface Roughness Modeling with

Machining Parameters
(Speed, Feed & Depth of Cut) in CNC Milling
Nitin Agarwal

Associate Professor
Department of Mechanical Engineering
Moradabad Institute of Technology, Moradabad

ABSTRACT
Surface roughness is one of the most significant requirements in metal machining operations. In order to attain enhanced
surface quality, the appropriate setting of machining parameters is important before the cutting operation take place. The
objective of this research is to analyze the effect of machining parameters on the surface quality of aluminum alloy in CNC
milling operation with HSS tool. A multiple regression model is developed with spindle speed, feed rate and depth of cut
as the independent variables and surface roughness parameter Ra as the dependant variable. The prediction ability of the
model has been tested and analyzed using t-test and it has been observed that there is no significant different between the
mean of Ra values of theoretical and experimental data at 5% level of significance.
Keywords: Surface Roughness, Machining Parameters, CNC Milling.

INTRODUCTION
Now-a-days, due to the growing demand of superior quality
components for its functional aspect, surface roughness of a
machined part plays a significant role in the modern
manufacturing process. A good quality machined surface
appreciably improves fatigue strength, corrosion resistance,
and creep life. Surface roughness also influences some
functional characteristics of parts, such as, contact causing
surface friction, wearing, light reflection, heat transmission,
ability of distributing and holding a lubricant, load bearing
capacity, coating etc. Consequently, the desired finish surface
is generally specified and the appropriate cutting parameters
are preferred to attain the required quality [1]. In manufacturing
industries, manufacturers focused on the quality and
productivity of the product. To increase the productivity of
the product, computer numerically machine tools have been
implemented during the past decades. Surface roughness is
one of the most important parameters to determine the quality
of product. The mechanism behind the formation of surface
roughness is very dynamic, complicated, and process
dependent. Several factors will influence the final surface
roughness in a CNC milling operations such as controllable
factors (spindle speed, feed rate and depth of cut) and
uncontrollable factors (tool geometry and material properties
of both tool and workiepce). Most of the machine operator
using trial and error method to set-up milling machine cutting

conditions [2]. This method is not effective and efficient and

the achievement of a desirable value is a repetitive and
empirical process that can be very time consuming. In order
to solve the problem, a surface prediction technique which is
termed the multiple regression prediction models to optimize
the cutting conditions is developed. This method can find the
best conditions required for the machining independent
variables such as speed, feed and depth of cut that would result
in the best machining response.

LITERATURE REVIEW
Surface finish of milled components has massive influence
on the quality of the finished product. Surface finish in milling
had been found to be influenced in varying amounts by a
number of factors such as feed rate, work hardness, built-up
edge, coolent used, cutting speed, depth of cut, cutting time,
cutting edge. According to these parameters, a comprehensive
literature survey is carried out as follows. Srikanth and Kamala
[3] developed a Real Coded Genetic Algorithm (RCGA) to
locate optimum cutting parameters and explained its
advantages over the existing approach of binary coded genetic
algorithm (BCGA). Franic and Joze [4] used Binary Coded
Genetic Algorithm (BCGA) for the optimization of machining
parameters. This genetic algorithm optimizes the machining
conditions having an influence on production cost, time and
quality of the final product. David et al. [5] addressed a

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

methodology to calculate surface roughness in a high speed

end-milling process and used Artificial Neural Networks
(ANN) and statistical tools to develop different surface
roughness predictors. Oktem et al. [6] used response surface
methodology to generate a mathematical model for surface
roughness in terms of cutting parameters: Feed, cutting speed,
axial depth of cut, radial depth of cut and machining tolerance.
An optimum surface roughness predictive model using binary
coded genetic algorithm (BCGA) is suggested by Suresh et
al. [7]. This model gives minimum and maximum values of
surface roughness and their respective optimal machining
conditions. Yang and Tarng [8] used Taguchi method for
design optimization on surface quality. An orthogonal array,
the signal-to-noise (S/N) ratio and the analysis of variance
(ANOVA) were used to explore the cutting characteristics.
Uros and Franci [9] suggested a neural network-based
technique to optimize the cutting parameters taking into
consideration the technological, economic and organizational
limitations. Al-Ahmari [10] proposed empirical models for
tool life, surface roughness and cutting force for turning
operations. Two significant techniques used were response
surface methodology and Neural Networks. Huang and Joseph
[11] suggested multiple regression model to calculate inprocess surface roughness in turning operation via
accelerometer. Hossain et al. [12] developed an artificial
neural network algorithm for calculating the surface finish in
end milling of Inconel 718 alloy. Avisekh et al. [13] performed
a feasibility study of on-line supervising of surface roughness
in turning operations using a developed opto-electrical
transducer. Regression and neural network (NN) models were
developed to predict surface roughness and compared to actual
and on-line measurements. Groover and Mikell [14] illustrated
the effect of three factors, namely, the feed, nose radius, and
cutting-edge angles, on surface roughness. Azouzi et al. [15]
suggested an on-line prediction of surface finish and
dimensional accuracy in turning using neural network based
sensor fusion. Feng and Hu [16] performed a comparative
study of the ideal and actual surface roughness in finish
turning and also developed the fractional factorial
experimentation approach for analysing the impact of turning
parameters on the surface quality of turned surfaces and used
ANOVA technique to examine the impact of turning factors
and factor interactions on surface roughness. Muammer et
al. [17] proposed regression analysis and neural networkbased method to predict the surface roughness and compare
it for various cutting conditions in turning. Bajic et al. [18]
modeled machined surface roughness and optimization of
cutting parameters in face milling and scrutinized the
influence of cutting parameters on surface roughness in face
milling. Sakir et al. [19] worked on the calculation of surface
roughness using artificial neural network in lathe machine
operations and examined the effect of tool geometry on
surface roughness on AISI 1040 steel in dry cutting condition
at a depth of cut of 0.5 mm. Optimization of machining
parameters not only enhance the utility for machining
economics, but also the product quality to a great extent.

56

Most of the techniques discussed above are based on trial

and error approach to set-up machining conditions in order
to achieve desired surface quality.
Therefore, the purpose of this study is to develop a multiple
regression model for the prediction of surface roughness
parameter Ra and to evaluate the prediction ability of the model.

PROBLEM DEFINITION
The goal of present analysis is to develop a technique to predict
the surface roughness of a part to be machined and to avoid
trial and error approaches to set-up machining conditions
in order to achieve the desired surface roughness. The objective
is to predict surface roughness parameter (Ra) under multiple
cutting conditions determined by spindle speed, feed rate and
depth of cut. Surface roughness would be measured directly
by surface roughness measuring instruments. Experimental
results are expected to show that parameters of spindle speed,
feed rate and depth of cut that could calculate surface roughness
(Ra) under different combinations of cutting parameters.

METHODOLOGY
Experiments have been carried out in order to examine the
impact of one or more factors of the process parameters
(spindle speed, feed rate and depth of cut) on the surface finish
of the machined surface in vertical milling operation. When
an experiment involves two or more factors, the factors can
influence the response individually or interactally. Figure 1
defines the steps involved in the process.

Figure 1: Steps involved in the process

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

Figure 2: Surface deviations

SELECTING THE VARIABLES

Figure 3: Machining process on the

workpiece in CNC machine

In this study the dependant variable is the surface roughness

(Ra) and the independent variables are the spindle speed, feed
rate and depth of cut. Because these variables are controllable
machining parameters, they can be used to calculate the surface
roughness in vertical milling which will then improve the
product quality. The variables are defined as follows:
Spindle speed: The rate at which the machine spindle
rotates. Spindle speed is typically measured in rpm.
Feed: It is the relative velocity at which the cutter is
advanced along the workpiece. Feed rate is measured in mm/
min.
Depth of cut: It is the thickness of the metal that is removed
in one cut. It is the perpendicular distance measured between
the machined surface and non-machined surface of the
workpiece.
Ra (surface roughness parameter): This parameter is also
known the arithmetic mean roughness value. It is defined as
the average value of the ordinates (y1, y2, y3,.., yn) from
the mean line. Ra is universally recognized and the most used
as international parameter of roughness. It is the arithmetic
mean of the departure of the roughness profile from the mean
line (Figure 2).
N
1
Ra =
Yi
N i =0

Figure 4: Dimensions of workpiece

Where Ra = the arithmetic average deviation from the mean

line.

Figure 5: Actual workpiece

y = the ordinate of the profile curve.

EXECUTION OF EXPERIMENTS
The test specimens are prepared on a CNC vertical milling
machine (Figure 3). A groove 10 mm X 1.5 mm is cut on
aluminium workpiece using HSS milling cutter of diameter
10 mm. The dimensions of the workpiece is shown in Figure 4
and actual workpiece is shown in Figure 5. Surface roughness
measurement is done off line with the usage SJ201 surface
roughness tester (Figure 6). A computer numeric control
(CNC) program has been written to perform the grooving
operation. The parameters defined in the CNC machine are:
Spindle speed (N), feed rate (F), depth of cut (D).Different
levels of cutting parameters are shown in Figure 7. All
specimens in this experiment are machined under dry cutting
conditions.

Figure 6: Arrangement of surface roughness

tester and workpiece

57

58

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

3. Surface roughness values measured within the

measuring area are sufficient to represent the roughness
of entire work piece.
4. Vibration is negligible.
5. Cutting edge of the tool is constant.
The measured values of surface roughness parameter Ra at
different combinations of machining parameters are shown
in Table 1.
Table 1: Measured surface roughness of the machines surface
at different machining parameters
S.No. Speed Feed Depth
Ra Value
(rpm) (mm/ of cut
min) (mm) Trial 1 Trial 1 Trial 1 Average

Figure 7: Different levels of cutting parameters

Also, after every specimen, the cutting tool was cleaned to
avoid chip formation or a built-up edge (BUE) which might
affect the surface roughness of the specimens. In addition, the
following assumptions are made:
1. The cutting tools used are identical in property.
2. The hardness of each work piece is same throughout the
length of the work piece.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

800
800
800
800
800
800
800
800
800
800
800
800
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1000
1200
1200
1200
1200
1200
1200
1200
1200
1200
1200
1200
1200

200
200
200
300
300
300
400
400
400
500
500
500
200
200
200
300
300
300
400
400
400
500
500
500
200
200
200
300
300
300
400
400
400
500
500
500

0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75
0.25
0.50
0.75

2.566
2.566
2.341
2.543
2.518
2.134
2.278
2.546
2.543
2.554
2.314
2.342
2.234
2.456
2.564
1.980
2.001
2.011
1.897
1.892
1.786
1.556
1.446
1.456
1.432
1.400
1.423
1.323
1.399
1.299
1.119
1.101
1.109
1.098
1.101
1.102

2.783
2.657
2.512
2.842
2.823
2.443
2.876
2.765
2.432
2.654
2.345
2.134
1.999
2.546
2.543
1.988
1.987
2.009
1.881
1.990
1.699
1.456
1.463
1.471
1.428
1.401
1.456
1.387
1.301
1.281
1.187
1.108
1.111
1.001
1.002
1.001

2.534
2.453
2.123
2.543
2.123
2.223
2.567
2.654
2.761
2.654
2.000
2.145
2.176
2.559
2.543
1.999
1.945
2.018
1.883
1.838
1.700
1.442
1.487
1.499
1.415
1.401
1.500
1.332
1.300
1.272
1.139
1.109
1.123
1.006
1.001
1.002

2.627
2.558
2.325
2.642
2.488
2.266
2.573
2.655
2.578
2.620
2.345
2.207
2.136
2.520
2.550
1.989
1.977
2.012
1.887
1.906
1.728
1.484
1.465
1.475
1.425
1.400
1.459
1.347
1.333
1.284
1.148
1.106
1.114
1.035
1.034
1.035

59

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

BUILDING THE MULTIPLE REGRESSION

MODEL
The general equation of multiple regression model is as
follows:
Ra = a + bN + cF + dD + eNF + fND + gFD
Where a,b,c,d,e,f and g are constants, N is spindle speed in
rpm, F is the feed rate in mm/min and D is the depth of cut in
mm.

ANALYZING AND VALIDATING THE

MODEL
The model has been tested using t-test at 5% level of
significance. This test is used to test the data of two samples
of size n1 and n2, mean x1 and x2 and standard deviation S1
and S2 from normal populations with the same variances. We
have to test the hypothesis that the population means are the
same. That is
Null hypothesis, H0: 1 = 2
i.e. the mean of both the population are same
Alternative hypothesis, H 1: 1 2
To carry out the test, we calculate the statistic.
x1  x2

1 1

n1 n2

S
x1

S2

1
n1

n1

i 1

xi , x2

1
n2

n2

j 1

n1
1

( xi 
n1  n2  2 i 1

xj
n

( x j  x2 ) 2
x1 )2 

j 1

If the calculated value of |t| be < t0.05, the difference between

the sample mean is said to insignificant at 5% level of
significance.

RESULT AND DISCUSSION

A multiple regression model has been developed which
includes the effect of spindle speed, cutting feed rate and depth
of cut, and any two variable interactions. The regression model
is as follows:
Ra= 3.1788 (0.00079N) + (0.009992F)
(0.77699D) (0.0000123NF) +
(0.00091ND) (0.000221FD)

The calculated values of Ra obtained from this model is

presented in Table 2. The model has been tested by t-test and
it is observed that the calculated value of |t| is less than t0.05,
hence there is no significant different between the mean of Ra
values of theoretical and experimental data at 5% level of
significant.

Table 2: Calculated surface roughness at

different machining parameters
S.No.

Speed
(rpm)

Feed
(mm/min)

Depth
of cut (mm)

Ra Value
(mm)

1
2

800
800

200
200

0.25
0.5

2.5539025
2.530605

800

200

0.75

2.5073075

800

300

0.25

2.5635775

800

300

0.5

2.534755

800

300

0.75

2.5059325

800

400

0.25

2.5732525

800

400

0.5

2.538905

800

400

0.75

2.5045575

10

800

500

0.25

2.5829275

11

800

500

0.5

2.543055

12

800

500

0.75

2.5031825

13

1000

200

0.25

1.9494025

14

1000

200

0.5

1.971605

15

1000

200

0.75

1.9938075

16

1000

300

0.25

1.7130775

17

1000

300

0.5

1.729755

18

1000

300

0.75

1.7464325

19

1000

400

0.25

1.4767525

20

1000

400

0.5

1.487905

21

1000

400

0.75

1.4990575

22

1000

500

0.25

1.2404275

23

1000

500

0.5

1.246055

24

1000

500

0.75

1.2516825

25

1200

200

0.25

1.3449025

26

1200

200

0.5

1.412605

27

1200

200

0.75

1.4803075

28

1200

300

0.25

0.8625775

29

1200

300

0.5

0.924755

30

1200

300

0.75

0.9869325

31

1200

400

0.25

0.3802525

32

1200

400

0.5

0.436905

33

1200

400

0.75

0.4935575

34

1200

500

0.25

0.1020725

35

1200

500

0.5

0.050945

36

1200

500

0.75

0.0011825

The numerical estimates of the effects indicate that the effect

of feed is in positive direction. The positive direction means
that the surface finish declined with increasing the cutting feed.
This is due to the increase in distance between the successive
cuts made by the tool during the cutting action, as the cutting
feed increases.

60

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

The spindle speed has effect in negative direction. This

means that increasing the spindle speed improves the surface
finish. It is generally well known that an increase in cutting
speed improves machinability. This may be due to the
continuous reduction in the build up edge formation as the
cutting speed increases.
The depth of cut also has negative value, which
indicates that increasing the depth of cut improves the surface
finish.
The interaction between the cutting feed and depth of cut
significantly affects the surface roughness. The interaction also
suggests that to get a certain surface finish and maximum metal
removal it is preferable to use a high cutting feed associated
with depth of cut.
It has been observed that increasing the spindle speed
improves the surface finish as the cutting feed decreases. This
supports the earlier discussion about the effect of increasing
cutting speed on the surface roughness of the machined
workpieces.
The interaction between the depth of cut and spindle speed
reveals that increasing the spindle speed and increasing the
depth of cut deteriorates the surface finish.

CONCLUSION
In the present analysis, 36 specimens of Aluminum alloy have
been machined in a CNC milling machine with HSS tool and
then a SJ 201 surface roughness tester has been used to
determine Ra values of all the specimens. A multiple regression
model has been developed and evaluated by means of
significant different between the predicted Ra values and the
actual Ra values with the help of t-test. The significant
conclusions drawn from the present research are recapitulated
as follows:
1. The surface roughness could be efficiently calculated by
using spindle speed, feed rate, and depth of cut as the input
variables.
2. Considering the individual parameters, depth of cut has
been established as most influencing parameter, follow by
feed rate and spindle speed.
3. The prediction ability of the model has been tested and
analyzed using t-test and it has been observed that there
is no significant different between the mean of Ra values
of theoretical and experimental data at 5% level of
significance.
4. As the spindle speed increases for lower feed rates, the
surface roughness decreases. For higher feed rates, the
surface roughness changes considerably.
5. As the depth of cut influences the surface roughness
considerably for a given feed rate, the increase in feed
rate causes the surface roughness to increase. For
lower depth of cut, the feed rate increases with surface
roughness.

REFERENCES
[1] Mike S.L., Joseph C., Chen, Li M. (1998). Surface Roughness
Prediction for CNC End Milling, Materials and Processes
Quality Control Manufacturing. J. Ind. Technol., 15(1):
(1999).
[2] Zhang J.Z., Joseph C.Chen, E.Daniel Kirby, Surface
Roughness Optimization in an End-milling Operation using
the Taguchi Design Method, 27 January 2006
[3] Kamala, V., Srikanth T. (2008). A Real Coded Genetic
Algorithm for Optimization of Cutting Parameters in Turning.
Int. J. Comput. Sci. Netw. Secur., 8(6):189-193.
[4] Franci, C., Joze, B., (2003). Optimization of Cutting Process
by GA Approach. Robotics and Computer Integrated
Manufacturing, 19:113- 121.
[5] David V., Rubn M., Menndez C., Rodrguez J., Alique
R. (2006). Neural Networks and Statistical Based Models for
Surface Roughness Prediction. International Association of
Science and Technology for Development, Proceedings of the
25th International Conference on Modeling, Indentification
and Control, pp. 326-331.
[6] Oktem H., Erzurumlu T., Kurtaran H. (2005). Application of
Response Surface Methodology in the Optimization of Cutting
Conditions for Surface Roughness. J. Mat. Process. Technol.,
170: 11-16.
[7] Suresh P.V.S., Venkateswara, R.P., Deshmukh, S.G. (2002).
A Genetic Algorithmic Approach for Optimization of
Surface Roughness Prediction Model. Int. J. Mach. Tools
Manuf., 42: 675-680.
[8] Yang, W.H., Tarng, Y.S., (1998). Design Optimization of
Cutting Parameters for Turning Operations Based on the
Taguchi Method. J. Mater. Process. Technol., 84: 122-129.
[9] Uros, Z., Franci, C. (2003). Optimization of Cutting
Conditions During Cutting by using Neural Networks. Robot.
Comput. Integr. Manuf., 19:189-199.
[10] Al-Ahmari A.M.A. (2007). Predictive Machinability Models
for a Selected Hard Material in Turning Operations. J. Mat.
Process. Technol., 190: 305-311.
[11] Huang, L., Joseph, C., Chen (2001). A Multiple Regression
Model to Predict In-process Surface Roughness in Turning
Operation Via Accelerometer. J. Ind. Technol., 17(2): 1-8.
[12] Hossain, M.I., Amin, A.K.M., Patwari, A.U. (2008).
Development of an Artificial Neural Network Algorithm for
Predicting the Surface Roughness in End Milling of Inconel
718 alloy. International Conference on Computer and
Communication Engineering, ICCCE (2008), 13-15: 13211324.
[13] Avisekh, B., Evgueni, V., Bordatchev, S., Kumar, C. (2009).
On-line monitoring of Surface Roughness in Turning
Operations with Optoelectrical Transducer. Int. J. Manuf. Res.,
4(1): 57-73.
[14] Groover, Mikell (1996). Fundamentals of Modern
Manufacturing. Prentice Hall, Upper Saddle River, NJ (now
Published by John Wiley & Sons, New York).
[15] Azouzi, R., Guillot, M. (1997). On-line Prediction of Surface
Finish and Dimensional Deviation in Turning using Neural
Network based Sensor Fusion. Int. J. Mach. Tool Manuf.,
37(9): 1201-1217.

MIT International Journal of Mechanical Engineering, Vol. 2, No. 1, Jan. 2012, pp. (55-61)
ISSN No. 2230 - 7680 MIT Publications

[16] Feng, C. X., Hu ZJ (2001). A Comparative Study of the Ideal

and Actual Surface Roughness in Finish Turning.
[17] Muammer, N., Hasan, G., Iahsan, T. (2007). Comparison of
Regression and Artificial Neural Network Models for Surface
Roughness Prediction with the Cutting Parameters in CNC
Turning. Modelling and Simulation in Engineering, Hindawi
Publishing Corp. New York, NY, United States, 3:2.

61

[18] Bajic, D., Lela, B., Zivkovic, D. (2008). Modeling of

Machined Surface Roughness and Optimization of Cutting
Parameters in Face Milling, 0543-5846.
[19] Sakir, T., Sleyman, N., Ismail, S., Sleyman, Y. (2008).
Prediction of Surface Roughness using Artificial Neural
Network in Lathe. International Conference on Computer
Systems and Technologies - CompSysTech08.