Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54 Received for review: 2023-06-15
© 2024 The Authors. CC BY 4.0 Int. Licensee: SV-JME Received revised form: 2023-10-02
DOI:10.5545/sv-jme.2023.692 Original Scientific Paper Accepted for publication: 2023-11-15
*Corr. Author’s Address: Nguyen Tat Thanh University, 300A Nguyen Tat Thanh, Ho Chi Minh, Vietnam, lvan@ntt.edu.vn 42
Multi-performance Optimization of the Rotary Turning Operation 
for Environmental and Quality Indicators
Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
Tat-Khoa Doan
1
, Trung-Thanh Nguyen
1
, An-Le V an
2,*
1 
Le Quy Don Technical University, Faculty of Mechanical Engineering, Vietnam 
2 
Nguyen Tat Thanh University, Faculty of Engineering and Technology, Vietnam
In this investigation, two environmental metrics (the comprehensive energy used (TU) and turning noise (TN)) and a quality metric (surface 
roughness (SR)) of the rotary turning process for the Ti6Al4V were optimized and reduced using the optimal factors (the inclined angle-i, depth 
of cut-d, feed-f, and turning speed-V). The TU model was proposed comprising the embodied energy of the insert and lubricant. The method 
based on the removal effects of criteria (MEREC), an improved quantum-behaved particle swarm optimization algorithm (IQPSO), and TOPSIS 
were applied to select weight values and the best optimal solution. The machining cost (MC) was proposed in terms of process parameters. 
The outcomes presented that the optimal values of the i, d, f, and V were 35 deg., 0.30 mm, 0.40 mm/rev., and 190 m/min, respectively, while 
the TU, SR, TN, and MC were saved by 6.7 %, 22.3 %, 23.5 %, and 8.5 %, respectively. The turning responses were primarily affected by the 
feed rate and turning speed, respectively. The developed turning process could be employed for machining hard-to-cut alloys. The developed 
approach could be applied to deal with optimization problems for other machining operations. 
Keywords: rotary turning, total energy consumption, surface roughness, noise emission, IQPSO
Highlights
• 	 A new rotary turning tool was designed and fabricated. 
• 	 Process parameters, including the spindle speed, depth of penetration, feed rate, and inclination angle were optimized. 
• 	 The total energy consumption, surface roughness, and turning noise were enhanced.
• 	 An improved quantum-behaved particle swarm optimization algorithm was proposed.
0  INTRODUCTION
The machining operation using rotary inserts is an 
effective solution to deal with hard-to-cut materials. 
The cutting temperature, force components, and 
pressure at the nose are reduced with the support of the 
rotational motion of the round piece. Additionally, a 
higher tool life is obtained due to the even distribution 
of the cutting temperature, leading to higher 
productivity and quality indicators, as compared to the 
conventional processes. 
Different milling and turning operations having 
rotary inserts have been developed and optimized 
by many investigators. Karaguzel et al. [1] indicated 
that the rotary turning and milling processes provide 
10- and 40-times longer tool life than conventional 
operations. The optimal cutting speed, feed, depth of 
cut, and inclination angle were selected to decrease the 
surface roughness and improve the material removal 
rate [2]. The ultrasonic vibration-based rotary turning 
was developed to machine decrease the machining 
forces and average roughness of the turned AA 7075 
[3]. The results indicated the tool speed of 98.63 m/
min and the feed of 0.08 mm/min were optimal data. 
A simulation model was developed to predict the tool 
wear in the rotary turning [4]. The authors stated that 
the tool wear was effectively decreased due to the 
disengagement duration. The energy efficiency and 
surface roughness were enhanced by 8.9 % and 24.8 
%, respectively using the optimal process parameters 
[5]. Nguyen emphasized that the energy consumption, 
surface roughness, and material removal rate of the 
turned SKD11 were affected by the speed, feed, depth 
of cut, and inclination angle [6]. Umer et al. indicated 
that an increased speed and/or depth caused a higher 
temperature of the turned 51200 steel [7]. Ahmed et 
al. stated that surface roughness and tool wear of the 
turned AISI 4140 were decreased by 24.6% and 32.6 
%, respectively using optimal process parameters [8].  
Nieslony et al. [9] indicated that a higher speed 
caused a decrease in the surface roughness and a 
stable turning operation, while an increased depth led 
to a higher intensity of the vibration. A rotary milling 
process was developed to machine the titanium alloy, 
in which a low speed was recommended to reduce the 
tool wear rate [10]. Ahmed et al. [11] emphasized that 
low process parameters (speed, feed, and depth of cut) 
and high inclination angle decreased the temperature 
of the rotary turning. Similarly, Chen et al. [12] 
emphasized that the surface roughness produced by the 
rotary process was lower than the conventional one. A 
novel simulation model was developed to forecast the 
temperature of the turned nickel and titanium alloys 
[13]. The authors stated that low process parameters 

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
43 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
(speed, feed, and depth of cut) and high inclination 
angle decreased the temperature. Umer et al. [14] 
revealed that a lower depth of cut was recommended 
to reduce the temperature and forces. The total energy 
consumed and machining time of the turned SKD61 
were decreased by 17.0 % and 17.8 %, respectively, 
using the optimal factors [15]. Additionally, the carbon 
emission of the rotary turning operation was reduced 
by 5.0 % using the PSO [16]. He et al. revealed 
that the cutting temperature of the turned K417 
alloy decreased with a higher inclination angle and 
cutting speed [17]. However, the shortcomings of the 
aforementioned works can be expressed as follows.  
An efficient self-propelled rotary tool having 
high stiffness to machine high-hardness steels has 
not been designed and fabricated to replace the fixed 
turning one. Low rigidity is a primary drawback of the 
proposed tools in previous publications.
The noise emission damages the inner ear and 
causes occupational hearing loss as well as chronic 
stress; hence, minimizing the sound intensity of the 
rotary turning operation is a necessary requirement. 
Moreover, the optimal process variables have 
not been determined to make reductions in energy 
consumed, roughness, and noise emission. 
The next section presents the framework. The 
experimental setting and discussions are then shown. 
Finally, the obtained findings are drawn.  
1  THE CONCEPT OF THE ROTARY TURNING OPERATION
The principle of the rotary turning process to produce 
external surfaces of hardened materials is presented 
in Fig. 1a. The workpiece is rotated around its axis, 
while the motion of the round piece is conducted 
using the friction between the body and specimen. 
The manufactured tool is shown in Fig. 1b, including 
the screws, bolts, the round insert, the base, and the 
holder. The milled grooves on the base are utilized 
to change the inclination angle. The round insert is 
conducted self-rotation using two bearings. The round 
inserts having a rake angle of 11
o
 and a hardness of 92 
HRC are utilized for all tests.
2  OPTIMIZATION APPROACH
The TU consists of the turning energy (TE), embodied 
energy for the insert (EI), and embodied energy for the 
coolant (EC). 
 TU TE EI EC  . (1)
The EC is computed as:
 TE EEEEEE
st sb tn attt c
 , (2)
where E
st
, E
sb
, E
tn
, E
at
, E
t
, E
tc
 are energy consumed 
in the startup, standby, transition, air-turning, turning, 
and tool change stages (Fig. 2).
Fig. 2.  The machining load of the rotary turning process
The start-up state presents the shortest period 
for turning on the lathe. The standby state denotes 
the stable period, which starts with turning on the 
machine tool and stops with the spindle rotation. The 
transition state refers to the short period for increasing 
and decreasing the spindle speed.  The air-turning 
state presents the duration with spindle rotation but no 
material cutting. The turning state refers to the steady 
period for material removal.
a) 
b) 
Fig. 1.  The concept of the rotary turning process; a) the schematic 
principle, and b) the fabricated rotary tool (1. the screws; 2. the 
bolts; 3. the round insert, 4. the base; and 5. the holder)

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
44 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
 
TE Pt Pt aV bV c
Pc Vc tP tP t
t
T
oo sb sb
st ac cs tt c
c
T
  
 




2
12
()
 

, (3)
where P
o
, P
sb
, and P
c
 are the power used in the 
startup, standby, and turning states, respectively. 
t
o
, t
sb
, t
a
, and t
c
 are the startup, standby, air-cutting, 
and turning time, respectively. a, b, and c denote 
the experimental coefficients. t
tc
 and T
T
 are the tool 
change time and tool life, respectively. The T
T
 is 
expressed as:
 T
A
Vfd
T


, (4)
where A, α, β, and γ are the experimental coefficients.
The EI is computed as:
 EI
t
T
SE I
c
T
iv
= , (5)
where SE
i
 and I
v
 are the fabricating energy and volume 
of each insert, respectively.
The EC is computed as:
 EC
t
T
VE
c
L
uL
  , (6)
where T
L
 and E
L
 denote the cycle time and fabricating 
energy of the lubricant, respectively. V
u
 is the lubricant 
volume. ρ and η are the density and concentration of 
the lubricant, respectively.
The SR is computed as:
 SR
R
n
ai
i
n



1
, (7)
where R
ai
 is the average roughness at the i
th
 measured 
point.
The TN is computed as:
 TN
TN
n
i
i
n



1
, (8)
where TN
i
 is the turning noise at the i
th
 measured time.
In this study, the characteristics of the coolant 
system, cutting piece, and specimen are named as 
constants. The factors considered and their levels are 
presented in Table 1. The ranges are determined based 
on the specifications of the lathe. Moreover, these 
values are confirmed with the published works related 
to the rotary turning processes. The optimization issue 
is presented as:
 Find X = [i, V, f, and d].
 Minimizing TU, SR, and TN;
	 Constraints:	20	deg	≤	i 	≤	50	deg;	
0.3	mm	≤	d	≤	0.7	mm;	
	 0.4	mm/rev	≤	f 	≤	0.8	mm/rev;	
	 90	m/min	≤	V 	≤	190	m/min.
Table 1.  Process parameters of the rotary turning
Symbol Parameters 1 2 3
i Inclination angle [deg] 20 35 50
d Turning depth [mm] 0.3 0.5 0.7
f Feed rate [mm/rev] 0.4 0.6 0.8
V Turning speed [m/min] 90 140 190
3.2  Optimization Framework
The optimizing approach is depicted in Fig. 3. 
Fig. 3.  Systematic optimizing procedure
Step 1: Performing experimental tests using the 
Box-Behnken design [18] and [19].
The Box-Behnken design requires three levels for 
each factor, which presents the lowest, middle, and 
highest ranges. The design points are placed on the 
middle points of the edge and the centre of the block. 
The advantages of the Box-Behnken design are the 
low number of tests and ensuring predictive accuracy. 
The number of experiments (NE) in the Box-Behnken 
design is computed as [20]:
 NE nn N
c
  21 () , (9)
where n and N
c
 are the number of variables and the 
number of centre points, respectively. In this work, 29 
experiments are performed for 4 process parameters 
and 5 replications.
Step 2: Developing regression models for energy 
components, SR, and TN [21].

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
45 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
Step 3: The MEREC is utilized to compute the 
weights. 
For the maximizing aim, the normalized response 
(n
ij
) is computed as:
 n
y
y
ij
i
i
=
min
. (10)
For the minimizing aim, the n
ij
 is computed as:
 n
y
y
ij
i
i
=
max
. (11)
The performance of the alternatives S
i
 is 
computed as:
 S
i
n
n
ij j
  












ln ln(), 1
1
 (12)
where n is the number of responses.
The performance of i
th
 alternative is computed as:
 S
ij
n
n
ij kk j
'
 














ln . ln()
,
1
1
 (13)
The removal effect of the j
th
 response (E
j
) is 
computed as:
 ES S
i ji ji
 
'
. (14)
The weight (ω
i
) is computed as:
 
i
j
k
k
E
E


. (15)
Step 4: Generation of the optimality using the 
IQPSO.
In the QPSO, the updated position of each particle 
is expressed as: [22] and [23]:
 
xt pt mt
xt
u
k
ij ij best ij
ij
,, ,
,
() () (( )
() )l n
 








1
1
0

If   . ., 5 (16)
 
xt Pt
mt xt
u
k
ij ij
best ij ij
,,
,,
() ()
() () ln










1
1
0

If   . ., 5 (17)
 pt Pt Gt
ij ij j ,,
() () () () ,   11  (18)
 mt
N
Pt
best ij ij
ij
NM
,,
,
,
() () . 


1
11
 (19)
In this work, the IQPSO combining the QPSO 
and theCauchy-Lorentz distribution is proposed to 
expand the perturbation [24]. The probability density 
function (f(x)) of the Cauchy-Lorentz distribution is 
given as:
 fx x
xx
o
o
(, ,)
()
, 











1
22
 (20)
where x
o
 and γ are the locations of the peak of the 
distribution and scale parameter, respectively.
In the mutation stage, each vector is added by a 
Cauchy-Lorentz random value (D(.)) and expressed: 
 xxD '( .),   (21)
where x ′	 is	 the	 new	 location	 after	 mutatation	 with	
random value to x.
The convergence of the QPSO-CL is enhanced 
with the aid of natural selection, which is expressed 
as:
 FXtF xt Fx t
N
(( )) (( )),..., (( )) , 
1
 (22)
where X(t) and F(X(t)) are the position vector of 
particles and fitness function of swarm, respectively. 
The particles are sorted based on fitness values, which 
is expressed as: 
 
FXtF xt Fx t
Xt xt xt
N
N
(( )) (( ))... (( ))
() () ,... ()
,
,
'' '
'''




1
1
 (23)
The operating steps of the IQPSO are illustrated 
in Fig. 4. Matlab 2019 commercial software entitled is 
used to conduct the IQPSO. 
Fig. 4.  The operating procedure of the IQPSO

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
46 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
Step 4: The best solution is selected by the 
TOPSIS. 
The normalized value of each alternative (g
ij
) is 
computed as: 
 g
e
e
ij
ij
ij
i
m



2
1
, (24)
where e
ij
 presents the value of the alternative j
th
.
The positive ideal solution (P
+
) and the negative 
idea solution (N
–
) are computed as:
 P v
ij
v
j
j
m





 
2
1
, (25)
 N v
ij
v
j
j
m

 




1
2
. (26)
The best point is found with the highest selection 
index (S
i
), which is calculated as:
 S
N
PN
i




. (27)
3  EXPERIMENTAL SETTING
A turning machine entitled EMCOTURN E45 is 
utilized to execute the turning trials. The Ti6Al4V bar 
with an outside diameter of 60 and a length of 400 
mm is utilized as the specimen (Fig. 5). The chemical 
compositions of the Ti6Al4V produced by EDX 
results are presented in Table 2 and Fig. 6. The outside 
diameter, inside diameter, and thickness of the round 
insert are 12 mm, 4.4 mm, and 4.76 mm, respectively. 
A KEW6305 electrical sensor, Mitutoyo SJ-301, and 
EXTECH 407730 sound meter are employed to obtain 
the power components, machined roughness, and 
turning noise.
Fig. 5.  The turned specimens
Table 2.  Chemical compositions of the Ti6Al4V
Elements Al Ti V Fe C O Others
[%] 6.01 83.74 3.26 0.16 0.28 5.08 Allowance
The representative data of the rotary turning 
operation are depicted in Fig. 7. Fig. 7a presents the 
power used at the experimental No. 16, while the 
roughness profile and SEM image are shown in Figs. 
7b and c, respectively. The wear and fracture have not 
been found on the edges of round inserts, as shown in 
Fig. 7d. The noise profile is presented in Fig. 7e.
a)            b) 
Fig. 6.  Investigation of a) the microstructure, and b) chemical compositions of Ti6Al4V; produced by EDX results
Table 3.  Regression models of the energy consumed in the transition state and operational power
No. Regression model R
2
Adjusted R
2
Predicted R
2
1 EC
ts
 = 0.000025V 
2
 – 0.0014V + 0.4682 0.9882 0.9794 0.9654
2 P
op
 = 0.0025V + 0.03682 0.9924 0.9826 0.9758

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
47 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
a)    b) 
c)      d) 
e) 
Fig. 7.  Representative experiments at experimental No. 16; a) power consumed, b) average roughness,  
c) the SEM image, d) the SEM image of the round insert, e) turning noise
4  RESULTS AND DISCUSSIONS
4.1  Development of EC
ts
 and P
op
 Models 
The EC
ts
 and P
op
 models are shown in Table 3. 
4.2  Development of E
t
, SR, and TN models
The obtained data for the E
t
, SR, and TN are presented 
in Table 4. 
The ANOVA results of the E
t
, SR, and TN are 
shown in Tables 5 to 7, respectively. The values of 
the R
2
, the adjusted R
2
, and the predicted R
2
 values 
indicate that the E
t
, SR, and TN correlations are 
significant. 
For the E
t
 model, the contributions of the i, d, 
f, and V are 2.11 %, 6.04 %, 22.79 %, and 27.33 %, 
respectively. The contributions of the if, df, dV, and fV 
are 1.22 %, 2.44 %, 2.89 %, and 4.23 %, respectively. 
The contributions of the i
2
, f
2
, and V
2 
are 6.89 %, 9.33 
%, and 13.2 %, respectively.
For the SR model, the contributions of the i, d, 
f, and V are 6.37 %, 18.18 %, 22.15 %, and 23.36 %, 
respectively. The contributions of the id, iV, dV, and fV 
are 1.06 %, 2.42 %, 3.34 %, and 2.59 %, respectively. 
The contributions of the i
2
 and d
2 
are 15.22 % and 
3.38 % respectively.

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
48 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
For the TN model, the contributions of the i, d, 
f, and V are 17.37 %, 15.91 %, 16.82 %, and 18.09 
%, respectively. The contributions of the if, dV, and 
fV are 1.64 %, 1.44 %, and 1.04 %, respectively. The 
contributions of the i
2
, d
2
, f
2
, and V
2 
are 21.23 %, 1.82 
%, 1.35 %, and 2.22 % respectively.
The deviations between the actual and predictive 
values of the E
t
, SR, and TN change from –0.99 % to 
1.26 %, from –0.97 to 0.80, and –1.26 % to –0.47 %, 
respectively (Table 8). Therefore, the E
t
, SR, and TN 
models are significant. 
The probability plots of three responses are 
presented in Fig. 8. It can be stated that observed 
data are distributed on straight lines, indicating the 
goodness of the fit of the proposed models.
Table 4.  Experimental data for developing the E
t 
, SR, and TN 
models
No. i d f V E
t
 SR TN
Experimental data for developing models
1 50 0.5 0.6 190 8.75 2.17 98.1
2 20 0.3 0.6 140 9.69 2.04 78.2
3 35 0.3 0.8 140 7.57 2.24 78.2
4 20 0.5 0.8 140 8.67 2.78 91.4
5 50 0.5 0.4 140 14.90 2.16 79.4
6 20 0.7 0.6 140 10.78 2.75 92.1
7 50 0.5 0.6 90 15.73 2.99 81.4
8 35 0.7 0.4 140 14.24 2.19 73.3
9 20 0.5 0.4 140 13.98 1.97 77.4
10 35 0.5 0.6 140 9.60 2.18 76.3
11 50 0.3 0.6 140 9.99 2.31 80.1
12 35 0.5 0.4 90 19.34 2.14 59.8
13 20 0.5 0.6 190 8.36 1.81 96.5
14 35 0.5 0.6 140 9.62 2.16 76.8
15 35 0.3 0.4 140 12.08 1.41 58.8
16 35 0.5 0.4 190 11.47 1.37 76.7
17 35 0.5 0.8 190 6.52 2.11 94.9
18 35 0.7 0.6 190 8.52 2.22 93.4
19 50 0.5 0.8 140 8.99 3.02 96.7
20 50 0.7 0.6 140 11.29 2.94 94.2
21 35 0.5 0.8 90 12.32 3.07 75.9
22 35 0.3 0.6 190 7.55 1.49 77.3
23 20 0.5 0.6 90 15.09 2.81 79.2
24 35 0.7 0.6 90 15.56 2.94 74.1
25 35 0.7 0.8 140 8.53 2.99 91.4
26 35 0.3 0.6 90 13.18 2.46 60.9
Experimental data for testing developed models
27 25 0.5 0.4 100 17.72 2.15 65.7
28 30 0.4 0.5 120 14.57 2.17 63.3
29 40 0.6 0.7 140 8.95 2.63 85.7
30 25 0.7 0.5 130 13.08 2.51 80.7
31 40 0.5 0.7 150 8.11 2.36 84.3
32 45 0.4 0.6 160 8.73 2.06 81.9
Table 5.  ANOVA results for the E
t
 model
So. SS MS F-value p-value Con. [%]
Mo. 249.1 17.8 35.6 < 0.0001
i 37.4 37.4 74.8 0.003 2.11
d 107.1 107.1 214.1 < 0.0001 6.04
f 404.0 404.0 807.9 < 0.0001 22.79
V 484.4 484.4 968.9 < 0.0001 27.33
if 21.6 21.6 43.2 0.007 1.22
dV 51.2 51.2 102.5 0.003 2.89
fV 75.0 75.0 150.0 0.002 4.23
i
2
122.1 122.1 244.3 < 0.0001 6.89
f  
2
165.4 165.4 330.8 < 0.0001 9.33
V 
2
234.0 234.0 467.9 < 0.0001 13.2
Re. 5.5 0.5 
To. 254.6  
R
2
 = 0.9784; Adj. R
2
 = 0.9692; Pred. R
2
 = 0.9578
Table 6.  ANOVA results for the SR model
So. SS MS F-value p-value Con. [%]
Mo. 6.33 0.45 38.90 < 0.0001 6.37
i 0.42 0.42 154.86 < 0.0001 18.18
d 1.19 1.19 441.98 < 0.0001 22.15
f 1.45 1.45 538.49 < 0.0001 23.36
V 1.53 1.53 567.91 < 0.0001 1.06
id 0.07 0.07 25.77 0.010 6.37
iV 0.16 0.16 58.83 0.007 2.42
dV 0.22 0.22 81.20 0.009 3.34
fV 0.17 0.17 62.97 0.007 2.59
i
2
1.00 1.00 370.02 < 0.0001 15.22
d 
2
0.22 0.22 82.17 0.009 3.38
Re. 0.13 0.01 
To. 6.45  
R
2
 = 0.9802; Adj. R
2
 = 0.9784; pred. R
2
 = 0.9662
Table 7.  ANOVA results for the TN model
So. SS MS F-value p-value Con. [%]
Mo. 3168.7 226.3 44.4 < 0.0001
i 2729.5 2729.5 535.1 0.0010 17.37
d 2500.1 2500.1 490.1 < 0.0001 15.91
f 2643.0 2643.0 518.1 < 0.0001 16.82
V 2842.6 2842.6 557.3 < 0.0001 18.09
if 257.7 257.7 50.5 0.0068 1.64
dV 226.3 226.3 44.4 0.0075 1.44
fV 163.4 163.4 32.0 0.0078 1.04
i
2
3336.0 3336.0 654.0 < 0.0001 21.23
d 
2
286.0 286.0 56.1 0.0066 1.82
f  
2
212.1 212.1 41.6 0.0074 1.35
V 
2
348.8 348.8 68.4 0.0062 2.22
Re. 56.1 5.1
To. 3224.8
R
2
 = 0.9826; Adj. R
2
 = 0.9794; pred. R
2
 = 0.9685

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
49 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
a) 
b)      
c) 
Fig. 8.  The probability plots of three responses;  
a) for E
t
 model, b) for SR model, c) for TN model
a) 
b) 
c) 
Fig. 9.  The residuals versus the observations for three responses; 
a) for E
t
 model, b) for SR model, c) for TN model
The residuals versus the observations of three 
responses are presented in Fig. 9. The errors of the 
responses are systematically distributed, presenting 
constant errors for each model. 
4.3  Parametric Impacts
The E
t
 is first reduced by 2.8 % with a higher i (Fig. 
10a). However, the E
t
 is increased by 7.9 % with 
a further i. An increased i causes a reduction in the 
cutting volume, leading to a decrease in the resistance; 
hence, the E
t
 reduces. A higher i increases the cutting 
volume due to the perpendicular tool, resulting in 
higher friction; hence, the E
t
 increases.
The E
t
 is increased by 14.9 % with an increment 
in the d (Fig. 10a). A higher d increases the thickness 
of the chip; hence, the E
t
 increases. 
The E
t
 is decreased by 33.6 % with an increment 
in the f (Fig. 10b). A higher f reduces the turning time; 
hence, the E
t
 increases.
The E
t
 is decreased by 38.9 % with an increment 
in the V (Fig. 10b). When the V increases, the 
turning time reduces; hence, the energy consumption 
decreases.

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
50 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
The SR is first decreased by 11.2 % with an 
increment in the i (Fig. 11a). However, the SR is 
increased by 21.9 % with a further i. An increased 
i decreases the turning volume, resulting in a low 
resistance; hence, a low SR is produced. A higher i 
causes an increased turning volume, leading to a hard 
turning; hence, a rough surface is generated.
The SR is increased by 30.9 % with a higher d 
(Fig. 11a). A higher d causes an increase in the turning 
volume, leading to higher resistance; hence, a higher 
SR is produced.
The SR is increased by 47.6 % with an increment 
in the f (Fig. 11b). A higher f causes an increase in the 
turning volume, leading to a higher friction; hence, the 
SR increases. Moreover, A higher f causes an increase 
in the turning marks, resulting in a higher roughness.
The SR is decreased by 47.6 % with an increment 
in the V (Fig. 11b). The cutting temperature increases 
with an increment in the V, resulting in softer 
specimen; hence, the SR reduces.  
The TN is decreased by 8.72 % with an increment 
in the i (Fig. 12a). However, the TN is increased by 
47.5 % with further i. A higher i decreases the material 
removal volume, resulting in low friction between the 
turning insert and workpiece; hence, the TN decreases. 
In contrast, a further i increases the material removal 
volume, leading to greater resistance; hence, the TN 
increases.
Table 8.  Confirmations of the precision of the developed models
No.
E
t
 [kJ] SR [µm] TN [dB]
Exp. Pred. Err. Exp. Pred. Err. Exp. Pred. Err.
27 17.72 17.86 -0.79 2.15 2.16 -0.47 65.7 66.1 -0.61
28 14.57 14.62 -0.34 2.17 2.18 -0.46 63.3 64.1 -1.26
29 8.95 9.04 -1.01 2.63 2.62 0.38 85.7 86.1 -0.47
30 13.08 13.16 -0.61 2.51 2.49 0.80 80.7 81.2 -0.62
31 8.11 8.19 -0.99 2.36 2.38 -0.85 84.3 84.9 -0.71
32 8.73 8.62 1.26 2.06 2.08 -0.97 81.9 82.6 -0.85
a)                 b) 
Fig. 10.  Interactions of process parameters on the E
t
; a) E
t
 vs. i and d, b) E
t
 vs. V and f
a)               b) 
Fig. 11.  Interactions of process parameters on the SR; a) SR vs. i and d, b) SR vs. V and f

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
51 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
The TN is increased by 27.6 % with an increment 
in the d (Fig. 12a). A higher D increases the material 
removal to be cut, leading to higher friction; hence, 
the TN increases. Moreover, a higher d causes greater 
resistance, resulting in higher turning noise. 
The TN is increased by 28.2 % with an increment 
in the f (Fig. 12b). An increased f causes higher 
material removal to be cut, leading to higher friction; 
hence, the TN increases. Additionally, a higher f 
increases the machining power of the drive system; 
hence, a higher TN is produced.
The TN is increased by 29.3 % with an increment 
in the V (Fig. 12b). A higher V increases the 
engagement frequency of the spindle system; hence, 
the TN increases. Additionally, an increased V causes 
higher material removal to be cut, leading to higher 
friction; hence, the TN increases.
The E
t
, SR, and TN are expressed as:
E
t
 = 0.43126 – 0.21181i+14.09801d – 49.89973f 
 – 0.25773V + 0.017943id – 0.04979if – 0.000085iV 
 – 7.46871df – 0.03535dV + 0.05171 + 0.00374i
2
 
 – 1.60143d 
2
 + 28.51704f  
2
 + 0.00064V 
2                          
(28)
SR = 3.29271 – 0.08815i – 0.4125d + 2.77125f
  – 0.0126V – 0.00666id + 0.00416if + 0.00006iV 
 – 0.1875df + 0.00625dV – 0.00475fV + 0.00126i
2
 
 + 1.58333d 
2
 – 0.07292f  
2
 + 0.0000053V 
2                  
(29)
TN = 56.38004 – 2.88317i + 57.05417d +49.7375f 
 – 0.00301V + 0.01666id + 0.275if – 0.0002iV 
 – 8.125df + 0.0725dV + 0.0525fV + 0.04743i 
 – 22.91667d 
2
 – 16.97917f  
2
+0.00044V 
2
             
(30)
4.4  Optimizing Outcomes Produced by the IQPSO 
Table 9 shows the coefficients for turning objectives. 
The values of the TU, SR, and TN are presented in 
Table 10. The weight values of the TE, SR, and TN are 
0.43, 0.37, and 0.20, respectively.
The Pareto fronts generated by IQPSO are 
exhibited in Fig. 13. As a result, turning objectives 
have contradictory trends. The reduction in the SR 
leads to a higher TU (Fig. 13a). Similarly, a decreased 
TU leads to a higher TN (Fig. 13b).  
The TOPSIS is utilized to select the best point 
among feasible solutions. The optimum values of the 
i, d, f, and V are 35 deg, 0.30 mm, 0.40 mm/rev., and 
190 m/min, respectively. The reductions in the TE, SR, 
and TN are 6.7 %, 22.3 %, and 23.5 %, respectively in 
comparison with the initial values (Table 11).
4.5  Comparisons with the Optimization Results Produced 
by the MOPSO
The optimum findings generated by the MOPSO 
of the i, d, f, and V are 29 deg, 0.30 mm, 0.40 mm/
rev, and 172 m/min, respectively. The reductions in 
the TU, SR, and TN are 6.0 %, 20.9 %, and 23.0 %, 
respectively, as compared to the initial values. The 
number of feasible solutions generated by the IQPSO 
and MOPSO are 426 and 286, respectively. It can be 
stated that the IQPSO provides better optimization 
results than the MOPSO. 
4.6  Evaluation of the Total Turning Cost
The comprehensive model for the MC is expressed as: 
MC kT Uk
t
T
kt tttt
kt
t
T
k
ec
c
T
labor os tatc c
labor ch
c
T
fp
 

()
(   

 
kttttt V
T
kk ttttt
T
fd os tatc cu
L
md mr os tatc c
m
)( )
() ()
( ()
()
,
31

 kt tttt
T
no st at cc
W
a)         b) 
Fig. 12.  Interactions of process parameters on the TN; a) TN vs. A and D, b) TN vs. V and f

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
52 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
where k
e
, k
c
 k
labor 
are the costs of energy, tool, and 
labour, respectively. V
u
 is the lubricant volume. k
fp
 
and k
fd
 present the cost for the lubricant preparation 
and disposal, respectively. k
md
, k
mr
, and T
m
 are the cost 
of the degradation and remanufacturing for the lathe, 
respectively. T
m
 is the service life of the machine. k
n
 
Table 9.  Experimental coefficients for the rotary turning process
p
o
 [kW] t
o
 [s] P
st
 [kW] t
st
 [s] t
a
 [s] t
tc
 [s] A α β
0.48 4 0.72 6 8 8 16.2×105 2.65 0.27
γ U
m
 [kJ/m
3
] T
L
 [month] V
in
 [cm
3
] V
ad
 [cm
3
] H [%] ρ [g/cm
3
] E
L
 [J/g) U
m
 [kJ/m
3
]
0.37 9.16×103 1 8.5 4.5 5 0.92 422984 9.16x103
a)      b) 
Fig. 13.  Pareto fronts generated by IQPSO; a) TE and SR, b) TN and SR
Table 10.  The values of total energy consumption, average roughness, and turning noise
No. i [deg] D [mm] f [mm/rev] V [m/min] TU [kJ] SR [µm] TN [dB]
1 20 0.3 0.6 140 25.31 2.04 73.6
2 50 0.3 0.6 140 25.72 2.32 90.9
3 20 0.7 0.6 140 26.69 2.76 89.5
4 50 0.7 0.6 140 27.31 2.96 107.2
5 35 0.5 0.4 90 33.65 2.16 63.9
6 35 0.5 0.4 190 28.52 1.38 81.3
7 35 0.5 0.8 90 27.05 3.08 79.8
8 35 0.5 0.8 190 23.98 2.11 992
9 35 0.3 0.6 90 28.05 2.41 64.3
10 35 0.3 0.6 190 24.65 1.41 81.1
11 35 0.7 0.6 90 30.24 2.96 78.9
12 35 0.7 0.6 190 25.43 2.21 98.5
13 20 0.5 0.4 140 29.84 1.93 74.1
14 20 0.5 0.8 140 24.57 2.73 89.4
15 50 0.5 0.4 140 30.66 2.14 89.9
16 50 0.5 0.8 140 24.79 3.00 108.5
17 35 0.3 0.4 140 28.31 1.47 62.1
18 35 0.3 0.8 140 23.33 2.31 79.7
19 35 0.7 0.4 140 30.38 2.16 78.8
20 35 0.7 0.8 140 24.21 2.98 95.3
21 20 0.5 0.6 90 29.73 2.83 74.3
22 20 0.5 0.6 190 25.75 1.87 92.8
23 50 0.5 0.6 90 30.37 2.98 92.1
24 50 0.5 0.6 190 26.14 2.20 109.9
25 35 0.5 0.6 140 25.48 2.17 80.5
26 35 0.3 0.4 190 27.77 1.03 71.1

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
53 Multi-performance Optimization of the Rotary Turning Operation for Environmental and Quality Indicators
and T
w
 are the noise tax and working hours per month, 
respectively.
The empirical coefficients of the MC are shown 
in Table 12. It can be stated that, the MC is saved by 
8.5 % at the selected point (Table 13).
Table 13.  Comparative values of the total cost
Method
Optimization parameters Response
i  
[deg]
d  
[mm]
f  
[mm/rev]
V   
[m/min]
MC 
[USD]
Initial values 50 0.30 0.40 140 4.91
Optimal 
results
35 0.30 0.40 190 4.48
Reduction [%] 8.5
4.7  The Contribution Analysis
The proposed cutting tool could be used in the 
practical rotary turning process for other hard-to-cut 
alloys. The new rotary turning tool could be developed 
based on the current device.
The empirical correlations of the performance 
measures could be effectively employed to forecast 
the total energy, turned roughness, and noise emission.  
The optimizing outcomes could be used in the 
practical operation to improve the technological data. 
The proposed turning process could be applied 
to produce external surfaces for other difficult-to-cut 
alloys.
The develop optimization approach could 
be applied to deal with other issues of different 
machining operations. 
The turning expense model could be used to 
compute total cost.
5  CONCLUSIONS
In the current work, the TU, SR, and TN of the rotary 
turning process were optimized, while optimal inputs 
were the i, d, f, and V. The MEREC and IQPSO were 
utilized to select optimal outcomes. The findings are 
expressed as below:
1.  To save the TU, the low data of the i and D were 
used, while the highest data of the f and V were 
utilized. To decrease the SR, the low d and f were 
utilized, while the high i and V were employed. 
For reducing the TN, the lowest process 
parameters could be applied.
2.  The TU and SR models were primarily affected 
by the f and V, followed by the d and I, 
respectively. For the TN model, the V had the 
highest contribution, followed by the f, i, and d, 
respectively. 
3.  The optimal i, d, f, and V were 35 deg, 0.30 mm, 
0.40 mm/rev, and 190 m/min, respectively. The 
TU, SR, and TN were saved 6.7 %, 22.3 %, and 
23.5 %, respectively.
4.  The IQPSO provided better optimization 
outcomes for the rotary turning process, as 
compared to the MOPSO.  
5.  The MC was decreased by 8.5 % at the selected 
point.
6.  The influences of rotary turning factors on the 
production rate and carbon emission will be 
explored in future works.
6  REFERENCES
[1] Karaguzel, U., Olgun, U., Uysal, E., Budak, E., Bakkal, M. 
(2015). Increasing tool life in machining of difficult-to-
cut materials using nonconventional turning processes.  
International Journal of Advanced Manufacturing Technology, 
vol. 77, p. 1993-2004, DOI:10.1007/s00170-014-6588-7.
[2] Rao, T.B., Krishna, A.G., Katta, R.K., Krishna, K.R. (2015). 
Modeling and multi-response optimization of machining 
performance while turning hardened steel with self-propelled 
rotary tool. Advances in Manufacturing, vol. 3, p. 84-95, 
DOI:10.1007/s40436-014-0092-z.
[3] Teimouri, R., Amini, S., Mohagheghian, N. (2017). Experimental 
study and empirical analysis on effect of ultrasonic vibration 
Table 11.  The optimality produced by the IQPSO and MOPSO
Method i [deg] d [mm] f [mm/rev] V [m/min] TU [kJ] SR [µm] TN [dB] S
i
Initial values 50 0.30 0.40 140 28.89 1.48 89.5
Optimal values by IQPSO 35 0.30 0.40 190 26.95 1.15 68.5 0.8624
Reductions by IQPSO [%] 6.7 22.3 23.5
Optimal values by MOPSO 29 0.30 0.40 172 27.08 1.17 68.9
Improvement by MOPSO [%] 6.0 20.9 23.0
Table 12.  Experimental coefficients for the turning cost model
k
e 
[USD/kWh] k
c 
[VND/piece] k
labour 
[USD/h] V
u
 [l] k
fp
 [USD/l] k
fd 
[USD/l] T
L 
[month] k
md
 [USD] k
mr 
[USD] T
m
 [year] k
n 
[USD]
0.15 16.62 8.40 20 0.14 0.45 1 41244.75 1649.79 14 2.68

Strojniški vestnik - Journal of Mechanical Engineering 70(2024)1-2, 42-54
54 Doan, T.-K. – Nguyen, T.-T. – Van, A.-L.
during rotary turning of aluminum 7075 aerospace alloy. 
Journal of Manufacturing Processes, vol. 26, p. 1-12, 
DOI:10.1016/j.jmapro.2016.11.011.
[4] Mohammad, L., Saeid, A., Mohsen, A. (2018). 3D FEM 
simulation of tool wear in ultrasonic assisted rotary 
turning. Ultrasonics, vol. 88, p. 106-114, DOI:10.1016/j.
ultras.2018.03.013.
[5] Nguyen, T.T., Duong, Q.D., Mia, M. (2020). Sustainability-
based optimization of the rotary turning of the hardened steel. 
Metals, vol. 10, art. ID 939, DOI:10.3390/met10070939.
[6] Nguyen, T.T. (2020). An energy-efficient optimization of 
the hard turning using rotary tool. Neural Computing & 
Applications, vol. 33, pp. 2621-2644, DOI:10.1007/s00521-
020-05149-2.
[7] Umer, U., Kishawy, H., Abidi, M.H., Mian, S.H., Moiduddin, 
K. (2020). Evaluation of self-propelled rotary tool in the 
machining of hardened steel using finite element models. 
Materials, vol. 13, no. 22, art. ID 5092, DOI:10.3390/
ma13225092.
[8] Ahmed, W., Hegab, H., Mohany, A., Kishawy, H. (2021). 
Analysis and optimization of machining hardened steel AISI 
4140 with self-propelled rotary tools. Materials, vol. 14, no. 
20, art. ID 6106, DOI:10.3390/ma14206106.
[9] Nieslony, P., Krolczyk, G., Chudy, R., Wojciechowski, S., 
Maruda, R., Biłous, P., Lipowczyk, M., Stachowiak, L. (2020). 
Study on physical and technological effects of precise turning 
with self-propelled rotary tool. Precision Engineering, vol. 66, 
pp. 62-75, DOI:10.1016/j.precisioneng.2020.06.003.
[10] Yujiang, L., Tao, C. (2021). Research on cutting performance 
in high-speed milling of TC11 titanium alloy using self-
propelled rotary milling cutters. International Journal of 
Advanced Manufacturing Technology, vol. 116, p. 2125-2135, 
DOI:10.1007/s00170-021-07592-4.
[11] Ahmed, W., Hegab, H., Kishawy, H., Mohany, A. (2021). 
Estimation of temperature in machining with self-propelled 
rotary tools using finite element method. Journal of 
Manufacturing Processes, vol. 61, p. 100-110, DOI:10.1016/j.
jmapro.2020.10.080.
[12] Chen, T., Wang, Y., Gao, W., Li, R. (2020). Comparative study 
on the cutting performance of self-propelled rotary cutters and 
indexable cutters in milling TC11 titanium alloy. International 
Journal of Advanced Manufacturing Technology, vol. 111, p. 
2749-2758, DOI:10.1007/s00170-020-06273-y.
[13] Ahmed, W., Hegab, H., Mohany, A., Kishawy, H. (2021). On 
machining hardened steel AISI 4140 with self-propelled rotary 
tools: experimental investigation and analysis. International 
Journal of Advanced Manufacturing Technology, vol. 113, p. 
3163-3176, DOI:10.1007/s00170-021-06827-8.
[14] Umer, U., Mian, S.H., Mohammed, M.K., Abidi, M.H., 
Moiduddin, K., Kishawy, H. (2022). Self-propelled rotary tools 
in hard turning: analysis and optimization via finite element 
models. Materials, vol. 15, no. 24, art. ID 8781, DOI:10.3390/
ma15248781.
[15] Nguyen, T.T. (2021). Analytical approach-based optimization 
of the actively driven rotary turning for environmental and 
economic metrics considering energy footprint of materials. 
Neural Computing & Applications, vol. 33, p. 11937-11950, 
DOI:10.1007/s00521-021-05891-1.
[16] Nguyen, T.T., Duong, Q.D., Mia, M. (2021). Multi-response 
optimization of the actively driven rotary turning for energy 
efficiency, carbon emissions, and machining quality. 
Proceedings of the Institution of Mechanical Engineers, Part 
B: Journal of Engineering Manufacture, vol. 235, no. 13, p. 
2155-2173, DOI:10.1177/09544054211001014.
[17] He, W., Xue, Z., Guo, R., Chen, N,. (2023). On machining K417 
superalloy with sliding self-propelled rotary tools: simulation 
and experimental investigation. International Journal of 
Advanced Manufacturing Technology, vol. 126, p. 1669-1680, 
DOI:10.1007/s00170-023-11218-2.
[18] Sampath, B., Myilsamy, S. (2021). Experimental Investigation 
of a cryogenically cooled oxygenmist near-dry wire-cut 
electrical discharge machining process. Strojniški vestnik - 
Journal of Mechanical Engineering, vol. 67, no. 6, p. 322-330, 
DOI:10.5545/sv-jme.2021.7161.
[19] Ning, P., Zhao, J., Ji, S., Li, J., & Dai, H. (2021). Ultra-precision 
single-point diamond turning of a complex sinusoidal mesh 
surface using machining accuracy active control. Strojniški 
vestnik - Journal of Mechanical Engineering, vol. 67, no. 7-8, 
p. 343-351, DOI:10.5545/sv-jme.2021.7172.
[20] Tzotzis, A., García-Hernández, C., Huertas-Talón, J., Kyratsis, 
P. (2020). 3D FE modelling of machining forces during AISI 
4140 hard turning. Strojniški vestnik - Journal of Mechanical 
Engineering, vol. 66, no. 7-8, p. 467-478, DOI:10.5545/sv-
jme.2020.6784.
[21] Satheesh, C., Sevvel, P., Senthil Kumar, R. (2020). 
Experimental Identification of optimized process parameters 
for FSW of AZ91C Mg Alloy using quadratic regression models. 
Strojniški vestnik - Journal of Mechanical Engineering, vo. 66, 
no. 12, p. 736-751, DOI:10.5545/sv-jme.2020.6929.
[22] Ma, M., Xiong, W., Lian, Y., Han, D., Zhao, C., Zhang, J. (2019). 
Modeling and optimization for laser cladding via multi-
objective quantum-behaved particle swarm optimization 
algorithm. Surface and Coatings Technology, vol. 381, art. ID 
125129, DOI:10.1016/j.surfcoat.2019.125129.
[23] Tang, M., Zhu, W., Sun, S., Xin, Y. (2023). Mathematical 
modeling of resource allocation for cognitive radio sensor 
health monitoring system using coevolutionary quantum-
behaved particle swarm optimization. Expert Systems with 
Applications, vol. 228, art. ID 120388, DOI:10.1016/j.
eswa.2023.120388.
[24] Guo, F., Zhu, J., Huang, L., Li, H., Deng, J., Zhang, X., Wang, 
K., Liu, H., Hou, X. (2023). A modified BRDF model based 
on Cauchy-Lorentz distribution theory for metal and coating 
materials. Photonics, vol. 10, no. 7, art. ID 773, DOI:10.3390/
photonics10070773.