Beruflich Dokumente
Kultur Dokumente
ISSN- 2455-5703
Abstract
This paper investigated the multi-response optimization of shielded metal arc welding (SMAW) process for an optimal
parametric combination to yield favorable bead geometry and mechanical properties of welded joints using artificial neural
networks method. Total twenty five sets of experimental input output data were obtained to train the ANN model and to verify
the prediction made by the model. Cross slide of a lathe machine was used to have constant welding speed. External magnetic
field was obtained with the help of a bar magnet mounted on tailstock side of the lathe machine with a wooden structure. The
main aim was to derive objective functions like welding current, voltage, speed of welding and external magnetic field to be
optimized within experimental domain. The objective functions have been selected in relation to parameters of SMAW welding
bead geometries like bead width, reinforcement height, depth of penetration, and mechanical properties like hardness, impact
strength and tensile strength. The model was trained with the help of eighteen sets of data. Optimal results have been verified
through seven other data sets of experiments. This shows application feasibility of the artificial neural networks for continuous
improvement in product quality in manufacturing industry.
Keywords- Back Propagation, Bead Geometry, External Magnetic Field, Input Process Parameters, Mechanical
Properties, Neurons
I. INTRODUCTION
The advantages of welding, as a joining process, include high joint efficiency, simple set up, flexibility and low fabrication costs
[1]. Shielded metal arc welding is a versatile and flexible process requiring simple equipment, a skilled welder, welder‟s
accessories and electrodes. Welding can be done in all positions, both in shop and at site. Welded joints of sound quality and
adequate mechanical properties can be obtained by using correctly designed electrodes and proper welding procedures. The
process is intermittent, because welding has to be interrupted from time to time to discard the unused stub and to place a fresh
electrode into the holder, and also to deslag the joint. For higher productivity, semi-automatic or fully-automatic welding
processes are preferred [2]. The mechanical strength of weld is influenced not only by the composition of the metal, but also by
the weld bead shape. The current, voltage, welding speed and polarity can influence the bead shape and size [3]. Due to the
effect of external longitudinal magnetic field the weld bead becomes wider. Usually, the wider the nugget diameter is, the better
the mechanical performance of the weld will be [4]. External magnetic field can produce electromagnetic stirring (EMS). EMS
is considered as an effective way to control the weld quality of SMAW at a relatively low cost with high efficiency. The
principle of EMS is the use of Lorentz force, which derives from the interactions of welding current and the externally applied
magnetic field. During the welding process, the molten metal driven by the Lorentz force makes high-speed movement and
eventually affects the melting and solidification process. EMS technique can successfully be applied in arc welding by affecting
the shape of weld pool and refining crystal grains [5]. The weld quality is achievable by meeting quality requirements such as
bead geometry which is highly influenced by various process parameters involved in the process. The weld bead geometry plays
an important role in determining the mechanical properties of the weld. Hence the input welding process variables which
influence the bead geometry must properly be selected to obtain an acceptable high quality joint [6].
Generally welded joints are the locations for the crack initiation due to inherent metallurgical, geometrical defects as
well as heterogeneity in mechanical properties and presence of residual stresses. For maintaining structural integrity of welded
structures throughout the service life of the structure, relationship between welding process, properties and performance of the
structure should be well-understood and established. Several studies and researches have been conducted so far to determine the
effect of welding parameters on weld properties and quality. Heat input affects the weld mechanical properties for SMAW
process [7]. Investigation into the relationship between the welding process parameters and bead geometry began in the mid
1900s and regression analysis was applied to welding geometry research [8]. An ANN model can be developed successfully to
predict weld bead geometry and penetration by considering electrode diameter, current, voltage, travel speed; electrode feed rate,
arc length and arc spread as influential factors for electric arc welding [9]. ANN modeling has been chosen by its capability to
solve complex and difficult problems. It provides better result in comparison to that of multiple regression analysis [10].
1. Multi-mete 2. Battery Eliminator 3. Electric Board 4. Gauss Meter 5. Table 6. Measuring Prob 7. Transformer Welding Set 8. Clamp meter
9. Tail Stock 10. Sleeve 11. Link (Wood) 12. Solenoid 13. Tool post 14. Iron sheet 15. Work piece 16. Electrode 17. Electrode Holder 18.
Metal Strip Connected with head stock 19. Head stock 20. Connecting Wires
Fig. 1: Experimental Setup
Table 1: Data for Training and Prediction
Rockwel
Weld
Welding Depth of l Charpy Tens.
Serial Magneti Width(m Reinforceme
Curre Voltag Speed Penetrati Hardne Imp.Strengt Strengt
Numbe c Field m) nt Height
nt (A) e (V) (mm/mi on (mm) ss (B) h. (J) h.
r (Gauss) (mm)
n) (MPa)
Tensile Strength(MPa)
Rockwell Hardness(B)
Rockwell Hardness(B)
Weld Wedth(mm)
Weld Wedth(mm)
Measured
Measured
Measured
Predicted
Predicted
Predicted
Predicted
Current
S.N.
S.N.
(A)
1 90 23 40 0 6.92 6.54 1.14 1.10 0.76 0.74 91 85.6 132 131.8 268 274.5
2 95 22 60 40 6.05 6.42 1.11 1.08 0.72 0.71 86 85.1 135 132.1 278 275.2
3 95 21 80 60 6.04 6.44 1.04 1.06 0.66 0.70 89 85.4 137 132.3 284 276.1
4 100 24 40 40 6.99 6.58 1.16 1.14 0.78 0.73 89 85.2 131 131.7 252 273.3
5 105 21 60 40 5.98 6.41 1.14 1.11 0.77 0.74 81 84.8 128 130.8 272 274.1
6 105 22 60 20 5.96 6.40 1.13 1.09 0.73 0.72 78 84.6 127 130.6 270 273.3
7 110 21 60 20 5.97 6.39 1.10 1.08 0.75 0.74 79 83.9 126 130.9 270 273.6
Voltage (V)
% age
%age
%age
S. N.
S.N.
% age
9
1 23 40 0 -5.49 -3.51 -2.63 -5.53 -0.15 2.43
0
9
2 22 60 40 +6.12 -2.70 -1.39 -1.05 -2.15 -1.01
5
9
3 21 80 60 +6.62 +1.92 +6.06 -4.04 -3.43 -2.78
5
1
4 0 24 40 40 -5.87 -1.72 -6.41 -4.27 0.53 8.45
0
1
5 0 21 60 40 +7.20 -2.63 -3.90 4.44 2.19 0.77
5
1
6 0 22 60 20 +7.38 -3.54 -1.37 8.46 2.63 1.22
5
1
7 1 21 60 20 +7.04 -1.82 -1.33 6.20 3.68 1.33
0
A. Weld Width
The weld width of the welded joints was almost unaffected if the magnetic field was changed from 0 to 20 gauss or from 20 to
40 gauss. If the field was increased from 40 gauss to 60 gauss, the weld width increased from 6.97 mm to 6.99 mm. and if it was
increased from 60 gauss to 80 gauss, the weld width increased from 6.99 mm to 7.03 mm. If the speed of welding was increased
from 40 mm/min to 60 mm/ min, the weld width decreased from 5.94 mm to 5.90 mm, and if it was increased from 60 mm/min
to 80 mm/min, the weld width of the weld decreased from 5.90 mm to 5.86 mm. The effect of voltage was positive for weld
width i.e. if voltage was increased from 20 V to 24 V, the weld width increased from 6.01 mm to 6.25 mm. The increment in
current, increased the weld width for all the investigated values. If the current was increased from 90 A to 110 A the weld width
increased from 5.91 mm to 5.97 mm. The variation of weld width with magnetic field, voltage, welding speed and current were
shown in figures 3, 4, 5 and 6 respectively.
B. Reinforcement Height
Reinforcement heights of all the joints were evaluated and they were presented in table 1. The magnetic field had almost no
effect on reinforcement height if it was changed in between 0 and 40 gauss, and after this the reinforcement height decreased if
magnetic field was increased upto 80 gauss which was our investigation range. If the magnetic field was increased from 40 gauss
to 60 gauss the reinforcement height decreased from 1.14 mm to 1.11 mm and if it was increased from 60 gauss to 80 gauss the
reinforcement height decreased from 1.11 mm to 1.09 mm. If the speed of welding was increased from 40 mm/ min to 80
mm/min the reinforcement height continuously decreased. Increment in voltage from 20 to 24V, increased the reinforcement
height from 1.06 mm to1.12 mm. if the increment in current was from 90 A to 110 A, the reinforcement height of weld generally.
The variation of reinforcement height with magnetic field, voltage, welding speed and current were shown clearly in figures 7, 8,
9, and 10 respectively.
C. Depth of Penetration
The depth of penetration of the weld cross-section was measured and the results were displayed in table 1. There was generally
no effect of magnetic field on depth of penetration if the strength of the field was less than 40 gauss and if it was increased from
40 gauss to 80 gauss the depth of penetration decreased from 0.80 mm to 0.76 mm. If the speed of welding was increased from
40 mm /min to 80 mm/ min the depth of penetration decreased from 0.83 mm to 0.76 mm. If the voltage was increased from 20
V to 24 V the depth of penetration decreased from 0.78 mm to 0.72 mm. If the current was increased from 90 V to 110 V, the
depth of penetration increased from 0.70 mm to 0.75 mm. The variation of depth of penetration with magnetic field, voltage,
welding speed and current were shown in figures 11, 12, 13 and 14 respectively.
D. Rockwell Hardness
The hardness across the weld cross-section was measured using a Rockwell hardness testing machine, and the readings were
displayed in table 1. The hardness of weld metal (WM) region was found greater than the HAZ region, but lower than the base
metal (BM) region, irrespective of filler metals used. There was no effect of magnetic field on hardness if the strength of the
field was less than 40 gauss and if it was increased from 40 gauss to 80 gauss the hardness increased from 90 RHB to 94 RHB. If
the speed of welding was increased from 40 mm /min to 80 mm/ min the hardness increased from 88RHB to 93 RHB. If the
voltage was increased from 20 V to 24 V the hardness decreased from 91 RHB to 82 RHB. If the current was increased from 90
V to 110 V, the hardness decreased from 89 RHB to 80 RHB. The variation of hardness properties with magnetic field, voltage,
welding speed and current were shown in figures 15, 16, 17 and 18 respectively.
E. Impact Strength
Charpy impact strength (toughness) values of all the joints were evaluated and they were presented in table 1. The magnetic
field had no effect on impact strength if it was changed in between 0 and 40 gauss, the impact strength remained constant at 131
J, and after this the impact strength increased if magnetic field was increased upto 80 gauss which was our investigation range. If
the magnetic field was increased from 40 gauss to 60 gauss the impact strength increased from 131 J to 136 J and if it was
increased from 60 gauss to 80 gauss the impact strength increased from 136 J to 138 J. If the speed of welding was increased
from 40 mm/ min to 80 mm/min the impact strength continuously increased. Increment in voltage from 20 to 24V, decreased the
impact strength from 138 J to 130 J., if the increment in current was from 90 A to 110 A, the impact strength of weld decreased
from 134 J to 126 J. The variation of toughness (impact strength) properties with magnetic field, voltage, welding speed and
current were shown clearly in figures 19, 20, 21, and 22 respectively.
F. Tensile Properties
Transverse tensile property of the joints was evaluated. The specimens were tested, and the results were presented in table 1. The
yield strength and tensile strength of unwelded base metal were measured as 359 and 524 M Pa, respectively. But the yield
strength and tensile strength of mild steel (fabricated using E-6013, rutile electrode filler metal) joints were reduced by about
50% in both the cases. The tensile strength of the welded joints was unaffected if the magnetic field was changed from 0 to 20
gauss or from 20 to 40 gauss. If the field was increased from 40 gauss to 60 gauss, the tensile strength increased from 266 M Pa
to 268 M Pa. and if it was increased from 60 gauss to 80 gauss, the tensile strength increased from 268 M Pa to 272 M Pa. If the
speed of welding was increased from 40 mm/min to 60 mm/ min, the tensile strength increased from 254 M Pa to 258 M Pa and
if it was increased from 60 mm/min to 80 mm/min, the tensile strength of the weld increased from 258 M Pa to 262 M Pa. The
effect of voltage was adverse for tensile strength i.e. if voltage was increased from 20 V to 24 V, the tensile strength decreased
continuously from 284 M Pa to 276 M Pa. The increment in current also decreased the tensile strength for all the investigated
values. If the current was increased from 90 A to 110 A the tensile strength decreased from 282 M Pa to 272 M Pa. The variation
of tensile properties with magnetic field, voltage, welding speed and current were shown in figures 23, 24, 25 and 26
respectively.
G. Prediction of Weld Bead Geometry and Mechanical Properties using Artificial Neural Networks
The developed neural network architecture was trained with help of back propagation algorithm using 18 data sets. The
developed network was tested out of 7 datasets. The training data sets and testing data sets are shown in table 1, the testing data
were not used for training the network. The % error was calculated between the experimental and predicted values as shown in
figure-2. The % error is ranging between -6.41 to 8.46. The other predictions are in between the above ranges and hence are very
close to the practical values, which indicate the super predicting capacity of the artificial neural network model.
V. DISCUSSION
In this investigation, an attempt was made to find out the best set of values of current, voltage, speed of welding and external
magnetic field to produce the best quality of weld in respect of weld width, reinforcement height, depth of penetration, hardness,
tensile strength and impact strength. Shielded metal arc welding is a universally used process for joining several metals.
Generally in this process speed of welding and feed rate of electrode both are controlled manually but in the present work the
speed of welding was controlled with the help of cross slide of a lathe machine hence only feed rate of electrode was controlled
manually which ensures better weld quality. In the present work external magnetic field was utilized to distribute the electrode
metal and heat produced to larger area of weld which improves several mechanical properties of the weld. The welding process
is a very complicated process in which no mathematical accurate relationship among different parameters can be developed. In
present work back propagation artificial neural network was used efficiently in which random weights were assigned to co-relate
different parameters which were rectified during several iterations of training. Finally the improved weights were used for
prediction which provided the results very near to the experimental values.
VI. CONCLUSION
The experimental analysis confirms that, artificial neural networks are power tools for analysis and modeling. Results revealed
that an artificial neural network is one of the alternatives methods to predict the weldbead geometry. Hence it can be proposed
for real time work environment. Based on the experimental work and the neural network modeling the following conclusions are
drawn:
1) A strong joint of mild steel is found to be produced in this work by using the SMAW technique.
2) If amperage is increased, weld width, depth of penetration and reinforcement height generally increase but hardness, tensile
strength and impact strength of weld generally decrease.
3) If voltage of the arc is increased, weld width and reinforcement height generally increase but depth of penetration hardness,
tensile strength and impact strength of weld, generally decrease.
4) If travel speed is increased, weld width, reinforcement height and depth of penetration of weld generally decrease but
hardness, tensile strength and impact strength of weld and generally increase.
5) If magnetic field is increased, weld width, hardness, tensile strength and impact strength of weld; generally increase but
reinforcement height and depth of penetration of weld, decrease.
6) Artificial neural networks based approaches can be used successfully for predicting the output parameters like weld width,
reinforcement height, depth of penetration, hardness, strength and impact strength of weld as shown in table 2. However the
error is rather high as in some cases in predicting depth of penetration it is more than 7 percent and in predicting hardness
and tensile strength it is more than 8 percent. Increasing the number of hidden layers and iterations can minimize this error.
REFERENCES
[1] Armentani. ER, Esposito, Sepe. R, "The Effect of Thermal Properties and Weld Efficiency on Residual Stresses in Welding". Journal of Achievements in
Materials and Manufacturing Engineering, 2007; Vol 20:319-322.
[2] Nadkarni,SV, Modern Arc Welding Technology", IBH Publication, New Delhi, India, 1988, pp-70.
[3] Apps. R, Gourd. L, Nelson. K, "Effect of Welding Variables upon Bead Shape and Size in Submerged Arc Welding", Welding Metal Fabrication, 1963,
31(11): pp. 453.
[4] Podlogar. V, Klopcic. B, Stumberger. G, Dolinar. D “Magnetic Core Model of a Mid Frequency Resistance Spot Welding Transformer”. IEEE Trans.
Magn. Feb. 2010; Vol. 46, pp. 602–605.
[5] Li. YB, Lin. ZQ, Chen. GL, Wang. YS, Xi. SY, “Study on Moving GTA Weld Pool in an Externally Applied Longitudinal Magnetic Field with
Experimental and Finite Element Methods”, Model. Simul. Mater. Sci. Eng. Oct. 2002, Vol. 10, pp. 781–798.
[6] Tarang, YS, Yang. WH, “Optimization of Weld Bead Geometry in GTAW by Taguchi method”, Int J Adv Manuf Technol, 1998, Vol. 14, pp.549 – 554.
[7] Funderburk SR, "Key Concepts in Welding Engineering", Welding Innovation, 1999, Vol. 16, No. 1.
[8] Lee. JI, Um. KW, “A Prediction of welding Process Parameters by prediction of Back-Bead Geometry”. J Mater Process Technol, 2000.
[9] Abdullah. AF, Hasib. MA, “Prediction of Weld Bead Geometry and Penetration in Electric Arc Welding using Artificial Neural Networks". Int. Jour. of
Mech. & Mechatronics Engg., 2010, Vol. 10, No: 04.
[10] Kim. LS, Lee. SH, Yarlagadda. PKDV, 2003. “Comparison of Multiple Regression and Back Propagation Neural Network Approaches in Modeling top
Bead Height of Multi-pass Gas Metal Arc Welds”, Sci. & Tech.of Welding and Joining, 2003, 8(5), pp. 347-352.
[11] Khan. MI, “Welding Science and Technology” New Age International (P) Limited Publishers, 2007.
[12] O.P. Khanna. OP, “A Text Book of Welding Technology”, Dhanpat Rai Publications Ltd., 2006, pp-3.
[13] Yousif. K. M. Daws. KM, Kazem. BI, 2008. “Prediction of Friction Stir Weldingm Characteristic Using Neural Network”. Jourdan J, 2008, Mech &
Indus. Engg. 2, pp. 151-155.