Optics and Laser Technology 94 (2017) 296–309
Contents lists available at ScienceDirect
Optics and Laser Technology journal homepage: www.elsevier.com/locate/jolt
Full length article
Study on laser welding of austenitic stainless steel by varying incident angle of pulsed laser beam Nikhil Kumar a,⇑, Manidipto Mukherjee b, Asish Bandyopadhyay a a b
Mechanical Engineering Department, Jadavpur University, Kolkata 700032, India Mechanical Engineering Department, C.V. Raman College of Engineering, Bhubaneswar 752 054, India
a r t i c l e
i n f o
Article history: Received 14 August 2016 Received in revised form 2 April 2017 Accepted 9 April 2017 Available online 22 April 2017 Keywords: Laser welding Response surface methodology Modeling and optimization Microstructure Mechanical properties
a b s t r a c t In the present work, AISI 304 stainless steel sheets are laser welded in butt joint configuration using a robotic control 600 W pulsed Nd:YAG laser system. The objective of the work is of twofold. Firstly, the study aims to find out the effect of incident angle on the weld pool geometry, microstructure and tensile property of the welded joints. Secondly, a set of experiments are conducted, according to response surface design, to investigate the effects of process parameters, namely, incident angle of laser beam, laser power and welding speed, on ultimate tensile strength by developing a second order polynomial equation. Study with three different incident angle of laser beam 89.7 deg, 85.5 deg and 83 deg has been presented in this work. It is observed that the weld pool geometry has been significantly altered with the deviation in incident angle. The weld pool shape at the top surface has been altered from semispherical or nearly spherical shape to tear drop shape with decrease in incident angle. Simultaneously, planer, fine columnar dendritic and coarse columnar dendritic structures have been observed at 89.7 deg, 85.5 deg and 83 deg incident angle respectively. Weld metals with 85.5 deg incident angle has higher fraction of carbide and d-ferrite precipitation in the austenitic matrix compared to other weld conditions. Hence, weld metal of 85.5 deg incident angle achieved higher micro-hardness of 280 HV and tensile strength of 579.26 MPa followed by 89.7 deg and 83 deg incident angle welds. Furthermore, the predicted maximum value of ultimate tensile strength of 580.50 MPa has been achieved for 85.95 deg incident angle using the developed equation where other two optimum parameter settings have been obtained as laser power of 455.52 W and welding speed of 4.95 mm/s. This observation has been satisfactorily validated by three confirmatory tests. Ó 2017 Elsevier Ltd. All rights reserved.
1. Introduction Austenitic stainless steel has a wide range of applications in nuclear structural fabrication, valve bodies and vessel internals because of their excellent mechanical properties. Joining process is required for this and laser welding is such a joining process. Laser welding has several advantages when compared to the conventional welding. It is non-contact type and its localized and narrow heat zone can create high quality result. Common re-working and after-work procedures are no more required. Laser welding has been widely applied in various industries including automotive, microelectronics, aerospace, medical, optoelectronics, microsystems etc. Kuryntsev and Gilmutdinov [1] have studied the laser welding of type 321 stainless steel and have found that the defo-
⇑ Corresponding author. E-mail addresses:
[email protected] (N. Kumar), m.mukherjee.ju@gmail. com (M. Mukherjee),
[email protected] (A. Bandyopadhyay). http://dx.doi.org/10.1016/j.optlastec.2017.04.008 0030-3992/Ó 2017 Elsevier Ltd. All rights reserved.
cused laser beam has increased the volume of weld pool that in turn to reduce the requirement for preparation of edge and gap between workpieces. Yan et al. [2] have investigated the microstructure and mechanical properties of tungsten inert gas, laser and laser-TIG hybrid welded 304 stainless steel. They have found that laser welded sample has highest tensile strength and smallest dendrite size than all other. Experimental investigation on dissimilar pulsed Nd:YAG laser welding of AISI 420 stainless steel to kovar alloy has been reported in [3] and they have found that the start of solidification in the kovar side of weld zone has occurred by means of epitaxial growth. In the work of Ai et al. [4] a defect-responsive optimization method for the fiber laser butt welding of dissimilar materials has been investigated. The genetic algorithm (GA) is applied to solve the model. The dissimilar laser welding of AISI 316L stainless steel to Ti6-Al4-6V alloy via pure vanadium interlayer has been studied by Tomashchuk et al. [5]. The effects of laser power, scanning speed, defocus distance, beam incident angle and line energy on weld bead geometry and
297
N. Kumar et al. / Optics and Laser Technology 94 (2017) 296–309
shearing force of laser welded dissimilar AISI 304L and AISI 430 stainless steel has been investigated in [6]. An investigation has been made by Keskitalo et al. [7] to study the laser welding of duplex stainless steel with nitrogen gas as shielding gas. The result suggests that nitrogen increases austenite levels in the weld metal and improved toughness levels. A study of simulation of laser butt welding of AISI 316L stainless steel sheet using various heat sources has been performed and experimental validation also has been done in [8]. The simulated thermal cycles, residual stress and distortion has been validated by experiments. In the research of Chen et al. [9], the influence of processing parameters on the characteristic of stainless steel/copper laser welding has been studied. Hao et al. [10] have investigated the effects of beam oscillating parameters on the weld morphologies. They have found that the difference in cross-section width from top to the lower gradually has reduced to disappear with the increase in oscillating frequency. An attempt has been made to improve the quality of the weldment between nickel titanium (NiTi) and AISI 316L stainless steel wires in [11]. A pulsed wave Nd:YAG laser system has been used for the welding of CP Ti and stainless steel sheets and the effect of pulse profiles used in laser welding on weld appearance, weld geometry, microstructure, hardness variation, joint strength and failure mode of weld have been investigated in [12]. Tan and Shin [13] have studied the multi-scale modeling of solidification and microstructure development in laser keyhole welding process for austenitic stainless steel. The model predictions are validated with the experimental results and the effects of the welding parameters are analyzed based on numerical and experimental results. Optimization of CO2 laser welding of DP/TRIP steel sheets using statistical approach has been conducted in [14]. In the article of Matsunawa et al. [15] the observation of keyhole as well as weld pool dynamics and their related phenomena to reveal the mechanism of porosity formation and its suppression methods have been studied. A numerical simulation model has been developed by Cho et al. [16] to study the temperature profile characteristics of weld bead and molten pool dynamics of high power disk laser welding process. Numerical and experimental study of molten pool formation during continuous laser welding of AZ91 magnesium alloy has been reported in [17]. A mathematical model has been developed by Zhou et al. [18] to analyze the heat transfer, fluid flow and keyhole dynamics during pulsed keyhole laser welding. A numerical and experimental investigation of laser welding of titanium alloy (Ti6Al4V) for modeling the temperature distribution to predict the heat affected zone, depth and width of the molten pool has been analyzed in [19]. Shanmugarajan et al. [20] have studied the effect of process parameters such as laser power, welding speed, shielding gas and laser beam mode on microstructure and mechanical properties of laser welded sample of type 304B4 borated stainless steel. Torkamany et al. [21] have analyzed the pulsed Nd:YAG laser welding of pure niobium plate to titanium alloy Ti-6AL-4V sheet in butt joint. The effect of pulsed Nd:YAG laser welding parameters and subsequent post-weld heat treatment on microstructure and hardness of AISI 420 stainless steel have been studied by Baghjari and Mousavi [22]. In the research of Chen et al. [23] the effect of laser-beam offsetting on microstructural characteristics and fracture behaviour of the laser butt joint of titanium alloy have been studied. An experimental procedure has been developed by Atabaki et al. [24] to join thick advance high strength steel plates by using the hybrid laser/arc welding (HLAW) process. An investigation has been made by Sun et al. [25] to analyze the laser butt joint of Al/steel dissimilar materials. Within scope of literature review, it has been observed that almost limited or no information is available on the effect of laser incident angle on the mechanical and microstructural properties of pulsed laser welding of AISI 304 stainless steel sheets in a butt joint configuration. It is one of the important parameter that may be co-
related with responses. This research aims to find the optimum incident angle for which the laser optic lens will be protected and also to understand the physical mechanisms responsible for the joint quality of the laser beam butt welding process of stainless steel plates. In the present work, 3 factors-5 levels experiments have been planned using response surface methodology (RSM) design matrix and analyzing the responses of interest by developed mathematical models based on experimental results. The second order mathematical equations have been developed for predicting the desired weld quality. In addition to statistical evaluation of the welded joints, metallurgical and mechanical analyses have been carried out on laser welded three specimens with incident angle 89.7 deg, 85.5 deg and 83 deg incident angle. A 3-D responses surface and contour plots have been developed to find the combined effect of input parameters on responses.
2. Response surface methodology Response surface methodology is a useful design of experiment method that is gaining popularity. This includes a review of basic experimental designs for fitting linear response surface models, in addition to a description of methods for the determination of optimum operating conditions. The steps of response surface methodology are: (i) Developing experimental strategy for selecting independent variables. (ii) Statistical modeling to build an approximate relationship between the response and process variables. (iii) Optimization for finding values of process variables producing desirable values of the response. When all the independent variables are measurable, controllable and continuous during experiments, response surface, y can be expressed with negligible error by:
y ¼ f ðxÞb þ 0
ð1Þ
where x = (x1, x2, . . ., xk). 0 f ðxÞ = a vector function of p elements. b = a vector of p unknown constant coefficients. = a random experimental error assumed zero mean. In RSM, an approximate model is needed to develop for the true response surface. The approximated model is constructed utilizing observed data from the process or system. Multiple regression analysis is commonly used for this. Usually, a second-order polynomial equation is used in RSM, which is given by
y ¼ b0 þ
k k X XX X bi xi þ bij xi xj þ bii x2i þ e i¼1
ð2Þ
i¼1
where parameters b0 bi, bij, bii are called regression coefficient for i = 0, 1, . . ., k and j = 0, 1. . .k. 2.1. Desirability function analysis It is an approach in which, individual responses are transformed to corresponding desirability values. Desirability value depends on acceptable tolerance range as well as target of the response. Unity is assigned, as the response reaches its target value, which is most desired situation. Beyond acceptable limit, desirability value assumes zero. In this study, individual desirability function posses one of the following two characteristics:
298
N. Kumar et al. / Optics and Laser Technology 94 (2017) 296–309
For goal of maximum, the desirability ðdi Þ will be defined by
di ¼
8 0; > > > :
yi Li Hi Li
wi
if responseðyi Þ 6 low v alueðLi Þ ; as responseðyi Þ v aries from lowðLi Þ to highðHi Þ if responseðyi Þ P high v alueðHi Þ
1;
ð3Þ For goal of minimum, the desirability will be defined by
di ¼
8 1; > > > :
Hi yi Hi Li
wi
if responseðyi Þ 6 low v alueðLi Þ ; as responseðyi Þv aries from lowðLi Þ to highðHi Þ if responseðyi Þ P high v alueðHi Þ
0;
ð4Þ A weight (w) can be assigned to a goal to emphasize the particular desirability function. Weights can be varied between 0.1 and 10. A weight greater than 1 gives more emphasis to the goal, while weights less than 1 give less emphasis. The simultaneous objective function, D, is a geometric mean of all transformed responses: 1 r r r D ¼ d11 d12 . . . d1n Rri ¼
n Y
Fig. 1. Schematic diagram of laser welding process and joint configuration.
! 1r r
di i
R
i
ð5Þ
top to bottom in the weld using Vickers’s microhardness testing machine (Make: LECO Co., USA; Model: LM248AT) at 50 gf load with 15 s dwell time. The measurement of tensile strength of welded samples has been conducted on Instron (Model-8801) as per ASTM E8 standard with strain rate of 2.8 104 s1. The schematic view of the tensile test specimen as per ASTM E8 with dimension is shown in Fig. 4.
i¼1
where n is the number of responses in the measure. Each response can be assigned an importance relative to the other responses. Importance ðr i Þ values varies from 1, the least important, to 5, the most important [26]. 3. Experimental set up and procedure
4. Results and discussions
AISI 304 type stainless steel has been chosen for experimental work, the dimensions of workpieces before welding 100 mm 20 mm 1.5 mm. The chemical composition of the material is shown in Table 1. The schematic diagram of laser welding process is shown in Fig. 1 and h indicates the incident angle. All the experiments have been conducted on JK600HP Nd:YAG laser generator (GSI, UK) integrated with ABB IRB 1410 robotic control. The welding of work piece has been conducted in pulse width of 5 ms and repetition rate of 25 Hz. Experimental set-up is given in Fig. 2. Samples are butt jointed and during welding technically zero gap between two sheets is maintained in each case. An argon gas jet emerges from the side nozzle which makes a fixed angle with the laser beam to avoid any external atmospheric contamination during welding. The laser beam has a spot size 0.75 mm. The ranges of input parameters are selected on the basis of trial experiments conducted by using one factor at a time approach. The chosen process parameters and their limits are given in Table 2. Eighteen experiments have been conducted as per central composite rotatable design (CCD) including 4 center points. Statistical software Design-Expert v10 has been applied to establish the design matrix. Fig. 3 shows a weld sample in butt joint configuration. Samples for the metallographic examinations have been prepared by polishing successively in 80, 120, 220, 320, 400, 1200, 1600, 2000 grade emery papers to remove the scratches. The compositions of the etchant are 2.4 gm. of CuCl2, 10 ml of 99% C2H5OH and 10 ml of 40% HCl. Furthermore, micro-hardness survey has been made on flat metallographic specimen across the joints and
The measured response is listed in Table 3. Design-Expert v10 software has been applied for analyzing the measured response and determining the mathematical model with best fit. The fitted quadratic polynomial model for response is statistically significant for the prediction within working range of welding parameters. Therefore, they will be used for further analysis. The maximum and minimum ultimate tensile strength are observed for sample no. 5 (P = 450 W, S = 5 mm/s and A = 85.5 deg) and sample no. 10 (P = 425 W, S = 5.5 mm/s and A = 83 deg) respectively. The metallurgical characteristics along with mechanical properties like hardness and tensile strength of weld sample no. 5, 10 and 16 have been presented in the following section. In general the laser welding is being performed with an incident angle of 89.7 deg, hence a comparison has been made between the sample no. 5 and 16 and sample no. 10 and 16. Since sample no 16 has been welded with an incident angle 89.7 (90 deg). 4.1. Development of mathematical model The adequacy of the developed model is tested using the sequential f-test, lack-of-fit test and analysis-of-variance (ANOVA) technique using the Design-Expert v10 software to obtain the bestfit model. The ANOVA tables also show the other adequacy measure R2, adjusted R2, adequacy precision R2 and predicted R2 for response is given in Table 3. The adequate precision compares
Table 1 Chemical compositions (wt.%) of AISI 304 stainless steel. Type
304 SS
Chemical composition C%
Si%
Mn%
P%
S%
Cr%
Ni%
Mo%
Cu%
Nb%
Al%
N%
0.079
0.2858
1.8
0.032
0.0194
18.56
8.20
0.265
0.292
0.0281
0.0063
–
299
N. Kumar et al. / Optics and Laser Technology 94 (2017) 296–309
Fig. 2. Robotic control laser welding set-up.
Table 2 Process control parameters and their limits. Parameters with units
Notation
Power, W Scanning speed, mm/s Incident angle, deg
P S A
Levels 2
1
0
+1
+2
407 4.16 81.29
425 4.50 83.00
450 5.00 85.50
475 5.50 88.00
492 5.84 89.70
Fig. 3. Top view of welded sample in butt joint configuration.
model terms are significant. The ‘‘lack-of-fit F-value” of 6.88 implies there is a 7.10% chance that a ‘‘lack-of-fit F-value” this large could occur due to noise. The ‘‘Predicted R2” of 0.9235 is in reasonable agreement with the ‘‘adjusted R2” of 0.9724 (i.e., the difference is less than 0.2). ‘‘Adequate precision” measures the signal to noise ratio. A ratio greater than 4 is desirable. The value of adequate precision of 25.13 indicates an adequate model. The model can be used to navigate the design space. The mathematical models for ultimate tensile strength, which can be used for prediction within same design space, are shown below: (a) In term of coded factors
UTS ¼ 575:434 þ 28:67P 15:135S þ 15:28A 26:10PA þ 13:52SA 55:71P2 83:83S2 22:69A2 Fig. 4. Laser welded sample for tensile test as per ASTM E8.
ð6Þ
(b) In term of actual factors
UTS ¼ 64732:16 þ 117:08P þ 2398:25S þ 760:81A the range of predicted value at the design points to the average predicted error [27]. The associated p-value of less than 0.05 for the model (i.e., p-value < 0.05, at 95% confidence level) indicates that the model terms are statistically significant. The lack-of-fit value of the model indicates non-significant, as this desirable. The ANOVA indicates that for the ultimate tensile strength model (Table 4), the laser power (P), welding speed (S), incident angle (A), interaction effect of laser power and incident angle (P A), welding speed and incident angle (S A), the quadratic effect of the laser power (P2 ), welding speed (S2 ), incident angle (A2 ) are the significant model terms. The interaction effect of laser power and welding speed (P S) is not significant and thus, eliminated by backward elimination process to improve model adequacy. The ANOVA result for reduced quadratic model is shown in Table 5. The model F-value of 75.98 implies the model is significant. There is only a 0.01% chance that an F-value this large could occur due to noise. Values of ‘‘prob > F” less than 0.05 indicate
0:41PA þ 10:81SA 0:08P2 335:31S2 3:63A2
ð7Þ
4.2. Validation of the developed model The developed response surface equation, derived from multiple regression analysis has been validated by conducting confirmatory tests. Three confirmatory experiments have been conducted and welding conditions have been chosen randomly. The tested results of experiments are presented in Table 6. It is obtained from Table 6 that there is a small error percentage between experimental and the predicted values form developed regression equation, which shows that the developed model can yield nearly accurate results. Fig. 5 shows the relationship between the actual and predicted values of responses. This figure also indicates that the developed model is adequate and predicted results are in good agreement with measured data.
300
N. Kumar et al. / Optics and Laser Technology 94 (2017) 296–309
Table 3 Central composite design for actual factors and measured experimental results. Experiment no.
Power, W
Welding speed, mm/s
Incident angle, deg
Ultimate tensile strength, MPa
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
450 425 492 450 450 475 475 450 425 425 450 425 475 475 450 450 450 407
5.00 4.50 5.00 5.84 5.00 4.50 5.50 5.00 5.50 5.50 5.00 4.50 5.50 4.50 5.00 5.00 4.15 5.00
85.5 88.0 85.5 85.5 85.5 88.0 83.0 85.5 88.0 83.0 85.5 83.0 88.0 83.0 81.2 89.7 85.5 85.5
577.81 443.82 466.94 334.00 579.26 429.66 408.23 578.42 404.30 308.64 565.23 369.48 432.23 492.48 491.16 537.28 348.60 374.70
Table 4 ANOVA for the fitted quadratic polynomial model for UTS of welded samples (before elimination). Source
Sum of squares
Df
Mean square
F value
p-value Prob > F
Remark
Model P S A PS PA SA P2 S2 A2 Residual Lack of fit Pure error Cor total
1.327E+005 11222.45 3125.26 3190.65 43.62 5450.72 1461.78 39262.43 88886.75 6512.97 1920.05 1786.99 133.06 1.346E+005
9 1 1 1 1 1 1 1 1 1 8 5 3 17
14741.39 11222.45 3125.26 3190.65 43.62 5450.72 1461.78 39262.43 88886.75 6512.97 240.01 357.40 44.35
61.42 46.76 13.02 13.29 0.18 22.71 6.09 163.59 370.35 27.14
