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

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

Surface Roughness Modeling with Machining

Parameters

(Speed, Feed & Depth of Cut) in CNC

Milling
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
aluminium 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 Raas 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 errormethod 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, builtup 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 GeneticAlgorithm (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

55

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 speedend-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, thesignal-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 in- process 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-linesupervising
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 network- based 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.

Most of the techniques discussed above are based on trial and error approach to setup machining conditions in order to achieve desired surface quality.

56

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 (R a) 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

57

SELECTING THE VARIABLES

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 andnon-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).
Ra1NYiNi0
Where Ra = the arithmetic average deviation from the mean line.
y = the ordinate of the profile curve.
Figure 3: Machining process on the workpiece in CNC machine

Figure 4: Dimensions of workpiece

Figure 5: Actual workpiece

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

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

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.

58

2.The hardness of each work piece is same throughout the length of the work piece.
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 R a 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

(rpm) (mm/ of cut

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

Ra Value

min)

(mm)

Trial 1 Trial 1

Trial 1 Average

200

0.25

2.566 2.783

2.534

2.627

200
200

0.50
0.75

2.566 2.657
2.341 2.512

2.453
2.123

2.558
2.325

300

0.25

2.543 2.842

2.543

2.642

300

0.50

2.518 2.823

2.123

2.488

300

0.75

2.134 2.443

2.223

2.266

400

0.25

2.278 2.876

2.567

2.573

400

0.50

2.546 2.765

2.654

2.655

400

0.75

2.543 2.432

2.761

2.578

500
500

0.25
0.50

2.554 2.654
2.314 2.345

2.654
2.000

2.620
2.345

500

0.75

2.342 2.134

2.145

2.207

200

0.25

2.234 1.999

2.176

2.136

200

0.50

2.456 2.546

2.559

2.520

200

0.75

2.564 2.543

2.543

2.550

300

0.25

1.980 1.988

1.999

1.989

300

0.50

2.001 1.987

1.945

1.977

300
400

0.75
0.25

2.011 2.009
1.897 1.881

2.018
1.883

2.012
1.887

400

0.50

1.892 1.990

1.838

1.906

400

0.75

1.786 1.699

1.700

1.728

500

0.25

1.556 1.456

1.442

1.484

500

0.50

1.446 1.463

1.487

1.465

500
200

0.75
0.25

1.456 1.471
1.432 1.428

1.499
1.415

1.475
1.425

200

0.50

1.400 1.401

1.401

1.400

200

0.75

1.423 1.456

1.500

1.459

300

0.25

1.323 1.387

1.332

1.347

300

0.50

1.399 1.301

1.300

1.333

300

0.75

1.299 1.281

1.272

1.284

400

0.25

1.119 1.187

1.139

1.148

400

0.50

1.101 1.108

1.109

1.106

400
500

0.75
0.25

1.109 1.111
1.098 1.001

1.123
1.006

1.114
1.035

500

0.50

1.101 1.002

1.001

1.034

500

0.75

1.102 1.001

1.002

1.035

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

59

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.

ANALYZINGAND 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, H1: 1 2
To carry out the test, we calculate the statistic.
t

x1x2
S 1
n1
1

x1

n1

1
n2

j1

1
nn
1

xj
n2

n1i1

S2

1n2

,x2

xi

2
2

n1

n2

(xix1)2

(xj x2)

j1

If the calculated value of |t| be < t 0.05, 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 ttest and
it is observed that the calculated value of |t| is less than t 0.05, 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 Feed
Depth
(rpm) (mm/min) of cut (mm)

1
2
3
4
5

800

200

800

200

800

200

800

300

800

300

0.25
0.5
0.75
0.25
0.5

Ra Value
(mm)

2.5539025
2.530605
2.5073075
2.5635775
2.534755

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

300

800

400

800

400

800

400

800

500

800

500

800

500

1000

200

1000

200

1000

200

1000

300

1000

300

1000

300

1000

400

1000

400

1000

400

1000

500

1000

500

1000

500

1200

200

1200

200

1200

200

1200

300

1200

300

1200

300

1200

400

1200

400

1200

400

1200

500

1200

500

1200

500

0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75
0.25
0.5
0.75

2.5059325
2.5732525
2.538905
2.5045575
2.5829275
2.543055
2.5031825
1.9494025
1.971605
1.9938075
1.7130775
1.729755
1.7464325
1.4767525
1.487905
1.4990575
1.2404275
1.246055
1.2516825
1.3449025
1.412605
1.4803075
0.8625775
0.924755
0.9869325
0.3802525
0.436905
0.4935575
0.1020725
0.050945
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.

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.

60

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 R a values of all the
specimens.Amultiple regression model has been developed and evaluated by means of significant different
between the predicted Ravalues 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 R avalues 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: 1321- 1324.
[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.

[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.

61