Optimization of shot peening parameters by response ...

Surface & Coatings Technology 305 (2016) 99–109

Contents lists available at ScienceDirect

Surface & Coatings Technology journal homepage: www.elsevier.com/locate/surfcoat

Optimization of shot peening parameters by response surface methodology Okan Unal Mechanical Engineering Department, Karabuk University, Karabuk 78050, Turkey

a r t i c l e

i n f o

Article history: Received 2 February 2016 Revised 21 May 2016 Accepted in revised form 2 August 2016 Available online 4 August 2016 Keywords: Response surface methodology Almen intensity Shot peening ANOVA Regression

a b s t r a c t In this study, the shot peening parameters which directly influence the arc height of the Almen strip and its characteristics are optimized within the context of Almen intensity, surface roughness and surface hardness via response surface methodology. Determination of the Almen intensity by trial and error method depending on the experience of the technician (measuring the arc height of Almen strips by changing the parameters repeatedly for each shot peening process) makes the optimization approaches valuable. The optimization is considered to perform by selecting surface roughness and surface hardness as the responses in order to classify the shot peening processes by taking into consideration of wide range of plastic deformation level. The effect of input parameters air pressure, shot diameter and peening duration on the Almen intensity, surface roughness and surface hardness is to be determined by using ANOVA regression analysis. Based on the estimated models, optimum peening conditions are introduced via response optimizer. The model adequacy is verified by the confirmation tests. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Almen intensity plays a crucial role in determining the amount of the plastic deformation of shot peening process [1–3]. Thus, it is located on the top level of the parameters list considered against the failure mechanisms such as fatigue [4–6] and corrosion [7]. Also it has remarkable influence on important mechanical properties such as surface roughness [8], surface hardness and residual stress [9,10]. Therefore, Almen intensity is determined with shot peening of Almen strips until the desired plastic deformation rate is obtained by adjusting the pre-conditions such as air pressure, shot diameter, peening duration and surface coverage before exactly the same application is exposed to the actual machine parts [11]. This situation leads to a large number of Almen strip waste depending on the technician's technical expertise [12]. The studies performed by mathematical [13,14] and modelling approaches [15,16] to determine the Almen intensity have attracted attention and provide useful insights to the practical applications. Even in recent years, the studies have been oriented to residual stress and ultra finenano structure formations [13,17–19]. The reason for this is the Almen intensity is directly considered as the major causal agent of all the mechanical and microstructural changes in the material. In recent years, studies related with shot peening dwell on the plastic deformation effect on the microstructural changes and integration of ultra fine-nano grain formation on the surface only by using the parameters higher than the conventional ones [20]. In other words, only creating excessive plastic defomation via severe shot peening could ensure E-mail address: [email protected].

http://dx.doi.org/10.1016/j.surfcoat.2016.08.004 0257-8972/© 2016 Elsevier B.V. All rights reserved.

the surface nanocrystallization [21]. However, surface severe plastic deformation methods increase the surface hardness up to higher levels and leads to form a brittle layer on the surface. Selection of bigger medias and higher deformation ranges causes surface roughness increase up to 2–3 times with compared to conventional shot peening [22]. Most of literature studies depict the improvement of fatigue strength is observed by means of surface nanocrystallization subjected to severe shot peening [23–26]. In rare cases, although the formation of nanocrystalline layer on the surface, significant improvements are not observed on the fatigue strength [22] and even in some cases the fatigue strength is negatively affected [27]. This is due to lack of the surface finish before and after surface treatment, materials behavior alteration and prolonged treatments. Therefore, two or three stage shot peening processes are applied and the results are analyzed [20,28]. In this study, the factors that influence the Almen intensity are intended to be optimized by taking into consideration of not only the level of desired plastic deformation insurance but also desired surface roughness and surface hardness. The response of Almen intensity, surface roughness and surface hardness have been investigated with the alteration of shot peening process parameters by means of “response surface methodology”. 2. 2. Experimental methods The most widely used strip (A strip) has been selected for the shot peening process. It is manufactured from AISI 1070 spring steel [2]. Fig. 1 shows the A strip specifications and the arc height (deflection) after the shot peening treatment.

100

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

Fig. 1. Comparison of shot peened and as received A strip surface and arc height (deflection).

Shot peening treatment has been performed manually in Peenmatic 2000S device with 200% constant coverage, 90 mm nozzle distance and 900 impact angle. The coverage finally has been applied particularly for the reduction of surface roughness and also leads to the saturation of arc height after the exposure of plastic deformation substantially. So, preconditions for determining the characteristics of shot peening, shot diameter, air pressure and peening duration are aimed to assess in terms of Almen intensity, surface roughness and surface hardness. Shot peening parameters and the responses are shown in Table 1. In the process, S110, S180 and S230 cast steel shot media is used. Commonly used shot media in shot peening process is presented in Fig. 2 (with the permission of Celik Granul Sanayi A.S.). Surface roughness of Almen strips is measured by using Mitutoyo Surface Roughness Tester, surface hardness measurements are performed by using Qness GmbH Q10 microhardness tester. Ra is obtained by measuring the values on three different points and averaging them. The hardness values are determined via 25 gf load through 10 s by measuring on five different points and averaging them. The optimization of the shot peening process parameters by means of arc height, surface roughness and surface hardness are performed via DOE statistical method by using Minitab. The effect of the parameters

are investigated on the outputs via response surface methodology (RSM) and L18 factorial design is applied. 2.1. Response surface methodology Design of experiment (DOE) is a statistical approach and applied for determining the large number of coefficients that affect the experimental investigations and the other input parameters [29,30]. In this study, the effect of the input parameters shot diameter (s), air pressure (p) and peening duration (s) on the Almen intensity, surface roughness and surface hardness are discussed. Arc height : f ðA; B; CÞ

ð1Þ

Surface roughness : f ðA; B; CÞ

ð2Þ

Surface hardness : f ðA; B; CÞ

ð3Þ

Response surface approach could be expressed by response surface “Y″ as second order polynomial regression equation. Y ¼ b0 þ ∑bi xi þ ∑bii x2i þ ∑bij xi x j þ er

Table 1 Shot peening input parameters and obtained arc heights (Almen Intensity) surface roughness and surface hardness. Test Peening No Duration (s)

Shot size (mm)

Air pressure (kPa)

Arc height (mm)

Almen Roughness Hardness intensity Ra (um) (HV)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

0,4572 0,4572 0,5842 0,2794 0,2794 0,2794 0,4572 0,5842 0,2794 0,2794 0,5842 0,2794 0,5842 0,5842 0,2794 0,5842 0,5842 0,2794

520 520 207 482 482 448 414 380 482 620 414 690 380 413 690 448 482 723

0,07 0,09 0,14 0,16 0,18 0,18 0,24 0,25 0,26 0,29 0,29 0,31 0,31 0,31 0,33 0,33 0,34 0,35

7A 9A 14 A 16 A 18 A 18 A 24 A 25 A 26 A 29 A 29 A 31 A 31 A 31 A 33 A 33 A 34 A 35 A

2 3 10 20 30 40 15 25 50 45 15 45 50 25 50 25 20 50

1266 1317 1893 1611 1786 1707 2412 2666 2229 2844 3111 2946 3385 3449 3318 3664 3748 3496

286 289 293 295 302 307 374 376 371 394 399 404 402 407 419 427 454 448

ð4Þ

3. Results and discussion 3.1. Analysis of variance and regression model for Almen intensity The regression coefficients evaluated by using ANOVA has determined the statistical significance of each factor in terms of Almen intensity (arc height) (Table 2). Although the selection of the confidence level of 95%, the whole factors have P values less than 0.05. The analysis reveals they are statistically significant for the optimization. The model for the arc height is compatible with total variance by means of R2 = 93,87%. If the R2 is closer to 1, model predicts the response quite better. R2(adj) coefficient is also close to 1, so achieves the model high significance [31]. Data normality has been investigated by means of normal probability plot. Residual normal probability plot for Almen intensity (arc height) is shown in Fig. 3. According to the normal probability plots, the all residuals has fallen to the straight line. The result confirms that errors are positioned normally and scattering has not been observed for Almen intensity (arc height).

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

101

Fig. 2. Shot media used in shot peening applications.

According to Table 2, air pressure (p), “peening duration × air pressure (t × p)” and “shot diameter × air pressure (s × p)” have the most significant influence on the Almen intensity. Besides, “shot diameter (s)” and “peening duration (t)” factors are also crucial for Almen intensity due to very low P values. However, the interaction of peening duration and shot diameter “peening duration × shot diameter (t × s)” influence is shown inadequate which has the P value of 0.04 with compared to the other factors mentioned above. Response surface regression equation can be obtained for Almen intensity is given below:

Arc height ðmmÞ ¼ 2101–0; 02895  t ðsÞ‐3132  s ðmmÞ‐0; 04429  p ðkPaÞ þ 0; 01; 601  t ðsÞ  s ðmmÞ þ 0; 059  t ðsÞ  p ðkPaÞ þ 0; 00721  s ðmmÞ  p ðkPaÞ ð5Þ

Fig. 4 presents 2D contour plots of Almen intensity versus three input parameters (shot diameter, air pressure and peening duration). Regions with different colors identify the interaction of the input parameters and their impact on the Almen intensity (arc height). When the 2D contour plots are examined, if peening duration and shot diameter increase together with closer rates, Almen intensity reveals considerable increase. However, if one parameter remains the same or rises with low levels and the other parameter increases with very high rates, the interaction graphs of “air pressure- peening duration” and “air pressure-shots diameter “ show Almen intensity is not much affected and also stays at lower values than expected. Specifically, the interaction graphs present another certain results that the air pressure plays an effective and trigger role on the Almen intensity. Miao et al. [2] investigate the shot diameter and shot velocity effect on the Almen intensity. The results prove the shot velocity is much more

Table 2 Regression coefficients and model summary of Almen intensity. Term

Adj SS

DF

Adj MS

F-value

P-value

Model Shot diameter (s) Peening duration (t) Air pressure (p) t×s t×p s×p Error Total Model Summary S = 0,02721

0,124,684 0,012,829 0,010,623 0,017,691 0,004,011 0,019,726 0,019,609 0,008,144 0,132,828

6 1 1 1 1 1 1 11 17

0,020,781 0,012,829 0,010,623 0,017,691 0,004,011 0,019,726 0,019,609 0,00,074

28,07 17,33 14,35 23,89 5,42 26,64 26,48

0.000 0.002 0.003 0.000 0,040 0,000 0,000

R2(adj) = 90,52%

R2(pre) = 73,69%

R2 = 93.87%

102

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

Residual Plots for Arc height (mm) Normal Probability Plot

Versus Fits

99

0,04

Residual

Percent

90 50

0,00 - 0,02

10 1

0,02

- 0,04 - 0,050

- 0,025

0,000

0,025

0,050

0,1

0,2

Residual

0,4

Versus Order 0,04

3

0,02

Residual

Frequency

Histogram 4

2

0,00 - 0,02

1 0

0,3

Fitted Value

- 0,04 - 0,03 - 0,02 - 0,01 0,00

0,01

0,02

0,03

0,04

2

4

6

Residual

8

10

12

14

16

18

Observation Order

Fig. 3. Residual normal probability plots for Almen intensity (arc height).

effective than the shot diameter. If shot velocity is considered for the air blast shot peening, it is directly associated with the air velocity and air pressure.

Fig. 5 shows a 3D surface plot of Almen intensity versus shot diameter and air pressure. In cases where the shot diameter is greater than 0.5 mm, the increase of air pressure is due to the increase of

Contour Plots of Arc height (mm) Shot diamet er (mm)*Peening durat ion (s)

Air pressure (kPa)*Peening durat ion (s)

700 0,54 600 0,48 500 0,42 400 0,36 300 0,30 10

20

30

40

50

10

20

30

Air pressure (kPa)*Shot diamet er (mm)

700

Arc height (mm) < 0,0 0,0 – 0,2 0,2 – 0,4 0,4 – 0,6 0,6 – 0,8 > 0,8

600 500 400 300

0,30

0,36

0,42

0,48

0,54

Fig. 4. 2D contour plots of Almen intensity (arc height) versus input parameters.

40

50

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

103

Surface Plot of Arc height (mm) vs Air pressure (kPa); Shot diameter (mm)

0,3

Arc height (mm) 0,2 800

0,1 600 0,3

400 0,4 0,5

Shot diamet er (mm)

0,6

Air pressure (kPa)

200

Fig. 5. 3D surface plot of Almen intensity versus shot diameter and air pressure.

Almen intensity. Therefore, the bigger shot media ensures the effective plastic deformation. However, smaller shots (0.3 mm) triggers the arc height when the air pressure level is so high. Only higher levels of air pressure compensate the limited capacity of the smaller shots on the plastic deformation. Model predicts Almen intensity is adversely affected when the shot diameter increase at very low levels of air pressure. The phenomena proves the dominant effect of air pressure to the arc height. In Guagliano's work [3], shot velocity increase provides a descending increase on the Almen intensity. And also the raise has been vanished at the shot velocity of 110 m/s. The results strengthen the presented approach of Fig. 5 that the shot peening treatment ensures the saturation of plastic deformation at very high level of the air pressure. 3.2. Analysis of variance and regression model for surface roughness The results for the surface roughness of ANOVA regression are presented in Table 3. R2 is close to 1 and 96,02% of total variations can be identified by the model. After ignoring the insignificant factors 77,77% of the total variations could be explained by the model. R2 (pre) is also compatible with the R2 and R2(adj) [30]. Fig. 6 shows the residual plots for surface roughness. The figure includes normal probability plot, residual fitted values, histogram of the residuals and residuals versus order of data. According to the residual plots and estimated regression coefficients (Table 3), the whole input parameters and their interactions are significant for the model. This is because the P values are quite smaller against α-value of 0,05 [32].

After assessing the coefficients, the regression equation for surface roughness is as follows: Surface roughness ðμmÞ ¼ 18; 14‐0; 2796  t ðsÞ‐25; 97  s ðmmÞ‐0; 03; 807  p ðkPaÞ þ 0; 1694  t ðsÞ  s ðmmÞ þ 0; 000536  t ðsÞ  p ðkPaÞ þ 0; 0621  s ðmmÞ  p ðkPaÞ ð6Þ The contour plots of surface roughness are shown in Fig. 7. Surface roughness (Ra) is raised by the increase of shot diameter and peening duration. Similarly, observed in “air pressure-peening duration” and “air pressure-shot diameter” interaction graphs, the increase of input parameters ensures the increase of surface roughness. In addition, even at low values of the parameters, the surface roughness is kept higher than the expected values. However, despite the high rate of one parameter, if another of the parameter is low, surface roughness remains at very low values and the outcome can be considered as an important result for the model. If the parameters that affect the plastic deformation and grain size alteration on the surface would be investigated and determined whether they satisfy the severe shot peening conditions or not, the study could prepare a groundwork in future of severe plastic deformation with reasonable levels of surface roughness. Bagherifard et al. [8] have been studied the surface roughness evaluation of low alloy steel 39NiCrMo3 (UNIEN10083) by theoretical and experimental approaches and observed the surface roughness increase by raising of shot diameter and shot velocity. Also, increase of surface coverage causes to the increase of roughness up to a certain value, but after that value remains constant regardless of the surface coverage increase.

Table 3 Regression coefficients and model summary of surface roughness. Term

Adj SS

DF

Adj MS

F-value

P-value

Model Shot diameter (s) Peening duration (t) Air pressure (p) t×s t×p s×p Error Total Model Summary S = 0,209055

11,6059 0,8815 0,9911 1,3072 0,4493 1,6523 1,4572 0,4807 12,0866

6 1 1 1 1 1 1 11 17

1,93431 0,88152 0,99112 0,30718 0,44929 1,65233 1,45721 0,04370

44,26 20,17 22,68 29,91 10,28 37,81 33,34

0.000 0.001 0.001 0.000 0,008 0,000 0,000

R2(adj) = 93,85%

R2(pre) = 89,73%

R2 = 96,02%

104

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

Residual Plots for Surface roughness Ra (um) Normal Probability Plot

Versus Fits

99 0,2

Residual

Percent

90 50

0,0

10 - 0,2 1 -0,50

- 0,25

0,00

0,25

0,50

1,5

2,0

2,5

3,0

Residual

Fitted Value

Histogram

Versus Order

3,5

0,2

Residual

Frequency

3

2

1

0,0

- 0,2 0

- 0,2

- 0,1

0,0

0,1

0,2

0,3

2

4

6

Residual

8

10

12

14

16

18

Observation Order

Fig. 6. Residual normal probability plots for surface roughness.

3D surface plot of surface roughness versus the input parameters air pressure and peening duration shown in Fig. 8. The surface plot reveals at lower level of peening duration, the increase of air pressure leads to

the fluctuation (decrease and increase together) on the surface roughness. However, the increase of peening duration provides the surface roughness increase at certain air pressure. It could be mentioned that

Contour Plots of Surface roughness Ra (um) Shot diameter (mm)*Peening duration (s)

Air pressure (kPa)*Peening duration (s)

700 0,54 600 0,48 500 0,42 400 0,36 300 0,30 10

20

30

40

50

10

20

30

Air pressure (kPa)*Shot diameter (mm)

700

Surface roughness Ra (um) < 0,0 0,0 – 1,5 1,5 – 3,0 3,0 – 4,5 4,5 – 6,0 6,0 – 7,5 > 75

600 500 400 300

0,30

0,36

0,42

0,48

0,54

Fig. 7. 2D contour plots of surface roughness versus input parameters.

40

50

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

105

Surface Plot of Surface roughness Ra (um vs Air pressure (kPa); Peening duration (s)

4

Surface roughness Ra (um) 3

2

1

45

0 800

30

600 15

400

Air pressure (kPa)

200 0

Peening duration (s)

0

Fig. 8. 3D surface plot of surface roughness versus shot diameter and air pressure.

at relatively long period of peening duration, the increase of air pressure induces negative effects on the surface roughness. Mylonas and Labeas [33] has demonstrated the impact of shot velocity on the surface roughness in terms of the numerical approach study on residual stress, surface roughness and cold work prediction. When the experimental results compared to numericals it is observed that the surface roughness shows increase abruptly by the increase of shot velocity. 3.3. Analysis of variance and regression model for surface hardness Table 4. presents the estimated regression coefficients and ANOVA for the surface hardness of different peened Almen strips. The input parameters shot diameter, peening duration and air pressure probability values (P value) are 0,001, 0,003 and 0,000, respectively. The values are all below 0,05 and therefore the terms take a role as significant influence on the response. However, “peening duration × shot diameter” interaction with P value of 0,096 could be assessed as insignificant term for the response. R2 value can be seen as satisfactory with 95,50%. The empirical equation of the surface hardness could be generated as shown below: Surface hardness ðHVÞ ¼ 1387–16; 41  t ðsÞ–1789  s ðmmÞ–2; 620  p ðkPaÞ þ 7; 07  t ðsÞ  s ðmmÞ þ 0; 03; 504  t ðsÞ  p ðkPaÞ þ 4; 527 sðmmÞ  p ðkPaÞ ð7Þ

The residual plots of the surface hardness of normal probability plot, residual versus fitted values, histogram of the residuals and residual

versus order of data has been shown in Fig. 9. It can be assumed that the residuals of the model for surface hardness are normally distributed. Fig. 10 shows the 2D contour plots of surface hardness versus the input parameters. It is observed from the contours that the increase of shot diameter and peening duration plays an important role on surface hardening. When the air pressure-peening duration contour is considered, even lower values of the input parameters, the model predicts the surface hardness rises up to much higher values (400–500 HV) with compared to interior core structure of the strip. This is interpreted to significant influence of air pressure and peening duration on surface hardening. However, the alteration in hardness is not observed at very high pressures if sufficient peening duration is not applied. In addition, peening duration has no influence alone (at very low air pressures) on the surface hardness. The effect of peening duration and shot diameter on surface hardness is shown in 3D surface plot (Fig. 11). Model predicts the surface hardness does not increase when smaller shot media is selected even the raise of peening duration. This is because smaller shot media can not provide sufficient plastic deformation for ensuring the surface hardening. In the previous studies, the selection of smaller media and lower air pressure (shot velocity) prevents the desired impact level of plastic deformation. At the same time, softer materials give better responses to the plastic deformation with compared to harder materials [34]. However, shot diameter increase generally raises the surface hardness dramatically. Model predicts the bigger shot media enhances surface hardness to maximum values despite lower peening durations. Also the grain refinement is one of the important output of the shot peening application which is exposed to materials as a severe plastic deformation [8]. The experimental investigations are concentrated on

Table 4 Regression coefficients and model summary of surface hardness. Term

Adj SS

DF

Adj MS

F-value

P-value

Model Shot diameter (s) Peening duration (t) Air pressure (p) t×s t×p s×p Error Total Model Summary S = 15,3476

54,972,6 4185,1 3414,2 6188,4 783,0 7051,1 7740,1 2591,0 57,563,6

6 1 1 1 1 1 1 11 17

9162,1 4185,1 3414,2 6188,4 783,0 7051,1 7740,1 235,5

38,90 17,77 14,49 26,27 3,32 29,94 32,86

0,000 0,001 0,003 0,000 0,096 0,000 0,000

R2(adj) = 93,04%

R2(pre) = 85,94%

R2 = 95,50%

106

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

Residual Plots for Surface hardness (HV) Versus Fits 20

90

10

Residual

Percent

Normal Probability Plot 99

50

0 - 10

10 - 20 1

- 30

- 15

0

15

30

300

350

Residual

Histogram

450

Versus Order 20

Residual

4

Frequency

400

Fitted Value

3 2

10 0 - 10

1 - 20 0

- 20

- 10

0

10

20

2

4

6

Residual

8

10

12

14

16

18

Observation Order

Fig. 9. Residual normal probability plots for surface hardness.

determining the distribution of grain size on and just below the surface by using high resolution transmission electron microscopy (HRTEM) and electron back scatter diffraction (EBSD) studies which are extensive

microstructural characterization equipments. However, numerical approaches for the grain refinement after shot peening process is limited [18,35,36].

Contour Plots of Surface hardness (HV) Shot diameter (mm)*Peening duration (s)

Air pressure (kPa)*Peening duration (s)

700 0,54 600 0,48 500 0,42 400 0,36 300 0,30 10

20

30

40

50

10

20

30

Air pressure (kPa)*Shot diameter (mm)

Surface hardness (HV) < 200 200 – 300 300 – 400 400 – 500 500 – 600 600 – 700 > 700

700 600 500 400 300

0,30

0,36

0,42

0,48

0,54

Fig. 10. 2D contour plots of surface hardness versus input parameters.

40

50

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

107

Surface Plot of Surface hardness (HV) vs Shot diameter (mm); Peening duration (s)

450

Surface hardness (HV)

400

350

300

0,6 0,5

0 15

0,4 30

Peening duration (s)

0,3

45

Shot diameter (mm)

Fig. 11. 3D surface plot of surface hardness versus shot diameter and peening duration.

Grain refinement or surface nanocrystallization is performed by accumulating strain via shot peening. Bagherifard et al. [35] constitute an approach based on the strain called equivalent plastic strain (PEEQ) in order to determine the conditions of nanocrystallization occurrence. They reveal the effectiveness of PEEQ profiles on the optimization of surface grain refinement. Hassani-Gangaraj et al. [18] reveal a hybrid model of finite element simulation and also dislocation density evolution model. The model gives compatible results with compared to experimentally evaluated cell size and capable of determining the grain refinement level with only one impingement. The model shows the media velocity is more effective than the shot size and hardness on the grain refinement. The optimization in this study also shows the air pressure leads to acceleration of the shot media provides higher surface hardness even at lower parameters. The hardenability can be observed as the capacity for plastic deformation and also nanograin formation. So surface nanocrystallization analysis with response surface methodology can be assessed as the objective for Almen strips for further investigations. 3.4. Validation of the model The prediction of the model has been confirmed with the experimental results for the aim of validating the model approach. Table 5 shows the predicted and experimented results of the arc height, surface roughness and surface hardness. From the observations in Table 5, it can be inferred the calculated errors are between 4.76% and 17.68%. The errors can be assessed as small and the validity of the model can be defined as satisfactory.

3.5. Optimization In this study, three responses (Almen intensity, surface roughness and surface hardness) are available and they are directly influenced from shot peening process parameters. It is desirable to obtain higher values of the two responses Almen intensity and surface hardness in order to provide better mechanical and microstructural facilities on the surface. However, the surface roughness between lower reasonable values is desired to prevent surface originated failures such as fatigue, fretting fatigue and corrosion. Desirability function approach is predicted with multiple response applications influenced by more than one input parameters [29]. Considering the outlined shot peening process, the surface hardness and arc height (Almen intensity) are to be maximized and the surface roughness is to be minimized with a target of 0,5 μm. The selection of the target as 0,5 μm is due to acceptation as the maximum value for grinding, honing and lapping. The value could assess as reasonable for the machine critical load carrying parts [37]. The optimization is performed by using Minitab Response Optimizer and the optimum parameters are found to be peening duration of 28 s, a shot diameter of 0,2794 (S110) and an air pressure of 723 kPa (Fig. 12). According to these optimized values, the estimated surface hardness is 397,84 HV with an individual desirability of 0,66576. The arc height is estimated as 0,30 mm (30 A Almen intensity) with an individual desirability of 0,84224. The overall desirability is evaluated as 0,7782 and can be assessed as satisfactory for the response optimization. The model designed for determining the conditions of Almen intensity before exposing to the real machine parts could also be adapted for the real treatment. For instance, constrains could be constructed on the Almen intensity for various types of materials. Aluminum alloys could

Table 5 Confirmation of the tests and validity of the model. Test no

Shot diameter (mm)

Air pressure (kPa)

Peening duration (s)

Pre. arc HEIGHT (mm)

Exp. arc height (mm)

Error (%)

1 2 3

0,4572 0,2794 0,5842

447 585 275

35 50 10

0,2835054 0,2777822 0,1433426

0,27 0,28 0,16

4.76% 7.92% 10.41%

Test no 1 2 3

Shot diameter (mm) 0,4572 0,2794 0,5842

Air pressure (kPa) 447 585 275

Peening duration (s) 35 50 10

Pre. Ra (μm) 3,27785364 2,8570649 2,1621415

Exp. Ra (μm) 2,88 2577 2015

Error (%) 12.13% 9.8% 6.8%

Test no 1 2 3

Shot diameter (mm) 0,4572 0,2794 0,5842

Air pressure (kPa) 447 585 275

Peening duration (s) 35 50 10

Pre. hardness (HV) 515,086 342,359 405,071

Exp. hardness (HV) 424 379 342

Error (%) 17.68% 9.67% 15.57%

108

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

Fig. 12. Optimization of Almen intensity, surface roughness and surface hardness.

reach the desired mechanical and microstructural properties by the application of less Almen intensity with compared to alloy and tool steels. Also parts with geometrical constrains such as small filled radii and small holes could restrict the shot diameter selection range. 4. Conclusion The optimization for the arc height (deflection), roughness and hardness of the Almen strips shot peened with different variations of the input process parameters is essential in order to provide less time and less Almen strip waste. In this study, the input process parameters air pressure, shot diameter and peening duration have been optimized by means of Almen intensity, surface roughness and surface hardness via response surface methodology. Empirical formulations of three responses have been created by using the significant input parameters and their interactions. Some results arise from the study are summarized as follows: 1) The model predicts the “air pressure”, “peening duration × air pressure” and “shot diameter × air pressure” as the most significant parameters for the Almen intensity (arc height). Increase of peening duration and shot diameter leads to the increase of arc height. According to

the 2D contours, air pressure has a decisive and trigger role on the Almen intensity. 2) Surface roughness dramatic increase is observed by raising the peening duration and shot diameter together. However, only one parameter with high amount raise leads to minimal changes of roughness. This model approach should be focused and has to be captured on attention. 3) According to the model, the all input parameters influence the surface hardness with different proportions. However, the most significant factor is observed as air pressure on the hardness. Air pressure increases the hardness significantly even if very short peening durations have been applied. Peening duration-shot diameter interaction is observed as the insignificant term for surface hardness. 4) According to the confirmation tests, the errors are between 4.76% and 17.68%. The values prove that the validity of the model is satisfactory for this shot peening application. 5) As a result of the optimization performed by Response optimizer, for the surface roughness Ra of 0,5 μm, maximum hardness and maximum Almen intensity, peening duration of 28 s, shot diameter of 0,2794 (S110) and the air pressure of 723 kPa are determined as the optimum parameters.

O. Unal / Surface & Coatings Technology 305 (2016) 99–109

References [1] O. Unal, R. Varol, Almen intensity effect on microstructure and mechanical properties of low carbon steel subjected to severe shot peening, Appl. Surf. Sci. 290 (2014) 40–47. [2] H.Y. Miao, S. Larose, C. Perron, M. Lévesque, An analytical approach to relate shot peening parameters to Almen intensity, Surf. Coat. Technol. 205 (2010) 2055–2066. [3] M. Guagliano, Relating Almen intensity to residual stresses induced by shot peening: a numerical approach, J. Mater. Process. Technol. 110 (2001) 277–286. [4] D. Li, H.N. Chen, H. Xu, The effect of nanostructured surface layer on the fatigue behaviors of a carbon steel, Appl. Surf. Sci. 255 (2009) 3811–3816. [5] Y. Lv, L. Lei, L. Sun, Effect of shot peening on the fatigue resistance of laser surface melted 20CrMnTi steel gear, Mater. Sci. Eng. A 629 (2015) 8–15. [6] Y. Lv, L. Lei, L. Sun, Influence of different combined severe shot peening and laser surface melting treatments on the fatigue performance of 20CrMnTi steel gear, Mater. Sci. Eng. A, 657 (2016) 77–85. [7] V. Azar, B. Hashemi, M. Rezaee Yazdi, The effect of shot peening on fatigue and corrosion behavior of 316L stainless steel in Ringer's solution, Surf. Coat. Technol. 204 (2010) 3546–3551. [8] S. Bagherifard, R. Ghelichi, M. Guagliano, Numerical and experimental analysis of surface roughness generated by shot peening, Appl. Surf. Sci. 258 (2012) 6831–6840. [9] L. Xie, C. Jiang, W. Lu, The influence of shot peening on the surface properties of (TiB + TiC)/Ti–6Al–4V, Appl. Surf. Sci. 280 (2013) 981–988. [10] L. Xie, C. Jiang, W. Lu, K. Zhan, Y. Chen, Investigation on the residual stress and microstructure of (TiB + TiC)/Ti–6Al–4V composite after shot peening, Mater. Sci. Eng. A 528 (2011) 3423–3427. [11] O. Unal, R. Varol, A. Erdogan, M.S. Gok, Wear behaviour of low carbon steel after severe shot peening, Mater. Res. Innov. 17 (2013) 519–523. [12] O. Unal, A. Cahit Karaoglanli, R. Varol, A. Kobayashi, Microstructure evolution and mechanical behavior of severe shot peened commercially pure titanium, Vacuum 110 (2014) 202–206. [13] L. Xie, J. Zhang, C. Xiong, L. Wu, C. Jiang, W. Lu, Investigation on experiments and numerical modelling of the residual stress distribution in deformed surface layer of Ti– 6Al–4V after shot peening, Mater. Des. 41 (2012) 314–318. [14] J. Liu, Numerical analysis for effects of shot peening on fatigue crack growth, Int. J. Fatigue 50 (2013) 101–108. [15] S.A. Meguid, G. Shagal, J.C. Stranart, Finite element modelling of shot-peening residual stresses, J. Mater. Process. Technol. 92–93 (1999) 401–404. [16] T. Chaise, J. Li, D. Nélias, R. Kubler, S. Taheri, G. Douchet, V. Robin, P. Gilles, Modelling of multiple impacts for the prediction of distortions and residual stresses induced by ultrasonic shot peening (USP), J. Mater. Process. Technol. 212 (2012) 2080–2090. [17] S.M.H. Gangaraj, M. Guagliano, G.H. Farrahi, An approach to relate shot peening finite element simulation to the actual coverage, Surf. Coat. Technol. 243 (2014) 39–45. [18] S.M. Hassani-Gangaraj, K.S. Cho, H.J.L. Voigt, M. Guagliano, C.A. Schuh, Experimental assessment and simulation of surface nanocrystallization by severe shot peening, Acta Mater. 97 (2015) 105–115. [19] S.M. Hassani-Gangaraj, M. Guagliano, G.H. Farrahi, Finite element simulation of shot peening coverage with the special attention on surface nanocrystallization, Procedia Eng. 10 (2011) 2464–2471.

109

[20] O. Unal, R. Varol, Surface severe plastic deformation of AISI 304 via conventional shot peening, severe shot peening and repeening, Appl. Surf. Sci. 351 (2015) 289–295. [21] S.M. Hassani-Gangaraj, A. Moridi, M. Guagliano, A. Ghidini, Nitriding duration reduction without sacrificing mechanical characteristics and fatigue behavior: the beneficial effect of surface nano-crystallization by prior severe shot peening, Mater. Des. 55 (2014) 492–498. [22] O. Unal, I. Zulcic, R. VAROL, I. Karademir, S. Ates, Novel type shot peening applications on railway axle steel, J. Miner. Met. Mater. Eng. 2 (2016) 1–5. [23] S. Bagherifard, M. Guagliano, Effects of surfaces nanocrystallization induced by shot peening on material properties: a review, Frattura ed Integrità Strutturale 7 (2009) 3–16. [24] X.P. Jiang, X.Y. Wang, J.X. Li, D.Y. Li, C.S. Man, M.J. Shepard, T. Zhai, Enhancement of fatigue and corrosion properties of pure Ti by sandblasting, Mater. Sci. Eng. A 429 (2006) 30–35. [25] J.W. Tian, J.C. Villegas, W. Yuan, D. Fielden, L. Shaw, P.K. Liaw, D.L. Klarstrom, A study of the effect of nanostructured surface layers on the fatigue behaviors of a C-2000 superalloy, Mater. Sci. Eng. A 468–470 (2007) 164–170. [26] Y. Ochi, K. Masaki, T. Matsumura, T. Sekino, Effect of shot-peening treatment on high cycle fatigue property of ductile cast iron, Int. J. Fatigue 23 (2001) 441–448. [27] S. Bagherifard, I. Fernandez-Pariente, R. Ghelichi, M. Guagliano, Effect of severe shot peening on microstructure and fatigue strength of cast iron, Int. J. Fatigue 65 (2014) 64–70. [28] P. Fu, K. Zhan, C. Jiang, Micro-structure and surface layer properties of 18CrNiMo7-6 steel after multistep shot peening, Mater. Des. 51 (2013) 309–314. [29] Y.-S. Nam, Y.-I. Jeong, B.-C. Shin, J.-H. Byun, Enhancing surface layer properties of an aircraft aluminum alloy by shot peening using response surface methodology, Mater. Des. 83 (2015) 566–576. [30] P. Sharma, D. Khanduja, S. Sharma, Parametric study of dry sliding wear of Aluminium metal matrix composites by response surface methodology, Mater. Today: Proc. 2 (2015) 2687–2697. [31] V. Vembu, G. Ganesan, Heat treatment optimization for tensile properties of 8011 Al/15% SiCp metal matrix composite using response surface methodology, Def. Technol. 11 (2015) 390–395. [32] A.R. Hamad, J.H. Abboud, F.M. Shuaeib, K.Y. Benyounis, Surface hardening of commercially pure titanium by laser nitriding: response surface analysis, Adv. Eng. Softw. 41 (2010) 674–679. [33] G.I. Mylonas, G. Labeas, Numerical modelling of shot peening process and corresponding products: residual stress, surface roughness and cold work prediction, Surf. Coat. Technol. 205 (2011) 4480–4494. [34] V. Llaneza, F.J. Belzunce, Study of the effects produced by shot peening on the surface of quenched and tempered steels: roughness, residual stresses and work hardening, Appl. Surf. Sci. 356 (2015) 475–485. [35] S. Bagherifard, R. Ghelichi, M. Guagliano, Mesh sensitivity assessment of shot peening finite element simulation aimed at surface grain refinement, Surf. Coat. Technol. 243 (2014) 58–64. [36] S.M. Hassani-Gangaraj, M. Guagliano, G.H. Farrahi, Finite element simulation of shot peening coverage with the special attention on surface nanocrystallization, Procedia Eng. 10 (2011) 2464–2471. [37] S. Kalpakjian, S.R. Schmid, Manufacturing Engineering & Technology, Pearson Education, Inc., Upper Saddle River, NJ, 2006.