International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847

The Effect of Process Parameters on Penetration in Shielded Metal Arc Welding under Magnetic Field using Artificial Neural Networks
R.P. Singh1, R.C. Gupta 2 and S.C. Sarkar 3
1

Mechanical Engineering Department, I.E.T., G.L.A. University Mathura, (U.P.) 2 Mechanical Engineering Department, I.E.T., Lucknow, (U.P)

Mechanical Engineering Department, Kumaon Engineering College, Dwarahat, (Uttarakhand)

*

Corresponding author: R.P. Singh

ABSTRACT
In this study, the effects of various welding parameters on penetration in mild steel having 5 mm thickness welded by shielded metal arc welding were investigated. The welding current, arc voltage and welding speed were chosen as variable parameters. The depths of penetration were measured for each specimen after the welding operations and the effects of these parameters on penetration were researched. Welding currents were chosen as 90, 95,100, 105 and 110 Ampere (A), arc voltages were chosen as 20, 21, 22,23 and 24Volt (V), the welding speeds were chosen as 40, 60 and 80 mm/min and external magnetic field strengths were used as 0, 20,40, 60 and 80 Gauss for all experiments. As a result of this study, it was observed that on increasing welding current, the depth of penetration increased. In addition, arc voltage is another important parameter for penetration. However, its effect is not as much as currents. The highest penetration was observed for 100 A current, 22 V voltage, 40 Gauss magnetic field and 40mm/min welding speed. In this paper, the effect of a longitudinal magnetic field generated by bar magnets on the weld was experimentally investigated. The welding speed was kept constant with the help of a lathe machine. Using the experimental data a multi-layer feed forward artificial neural network with back propagation algorithm was modeled to predict the effects of welding input process parameters on weld bead geometry.

Keywords: Artificial Neural Networks, Back Propagation, Bead Geometry, Input process parameters.

1. INTRODUCTION
Welding is one of the essential and inescapable processes in manufacturing industries. Various types of welding processes like Shielded Metal Arc Welding (SMAW), Gas Tungsten Arc Welding (GTAW), Gas Metal Arc Welding (GMAW), and Flux Cored Arc Welding (FCAW) are being practiced in industrial environment. Welding technology has obtained access virtually to every branch of manufacturing; to name a few, ships, rail road equipments, building construction, boilers, launch vehicles, pipelines, nuclear power plants, aircrafts, automobiles, pipelines. Welding technology needs constant upgrading with the widespread applications of welding, [1]. To consistently produce high quality of welds, arc welding requires experienced welding personnel. One reason for this is the need to properly select welding parameters for a given task to provide a good weld quality which is identified by its micro-structure and the amount of spatter and relied on the correct bead geometry size. Therefore, the use of the control system in arc welding can eliminate much of the guess work often employed by welders to specify welding parameters for a given task [2]. 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 [3]. Many efforts have been carried out for the development of various algorithms in the modeling of arc welding process [4]. In the early days, arc welding was carried out manually so that the weld quality could be totally controlled by the welder ability. Mc Glone and Chadwick have reported a mathematical analysis correlating process variables and bead geometry for the submerged arc welding of square edge close butts [5]. Chandel first applied this technique to the GMA welding process and investigated relationship between process variables and bead geometry [6]. These results showed that arc current has the greatest influence on bead geometry, and that mathematical models derived from experimental results can be used to predict bead geometry accurately. Nearly 90% of welding in world is carried out by one or the other arc welding process; therefore it is imperative to discuss the effects of welding parameters on the weld-ability of the materials during the arc

Volume 1, Issue 4, December 2012

Page 12

International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847
welding. Mild steel was selected for work-pieces to be welded because it is the most common form of steel as its price is relatively low while it provides material properties which are acceptable for many applications. Among these fusions welding processes, SMAW is very important. The advantages of this method are that it is the simplest of the all arc welding processes. The equipment is often small in size and can be easily shifted from one place to the other. Cost of the equipment is also low. This process finds numerous applications because of the availability of a wide variety of electrodes which makes it possible to weld a number of metals and their alloys. The welding of the joints may be carried out in any position with highest weld quality and therefore the joints which are difficult to be welded because of their position by automatic welding machines could be easily welded by shielded metal arc welding. Both alternating and direct current power sources could be used effectively. Power sources for this type of welding could be plugged into domestic single phase electric supply, which makes it popular with fabrications of smaller sizes. However, non equilibrium heating and cooling of the weld pool can produce micro-structural changes which may greatly affect mechanical properties of weld metal [7]. Selections of significant input process parameters are very important to obtain good bead geometry, which represents the strength of weld and its quality. Generally input parameters are chosen by the welders and engineers based on their experience and by trial and error method. After conducting trial the welds are inspected whether it meets the joint requirement or not. Since the weld bead quality depends more on the input process parameters, it is essential to study the effects of input process parameters on weld quality. Among the input process parameters welding speed, welding current, arc voltage and external magnetic field were considered as significant parameters in SMAW and the depth of penetration was taken as output value. The effective range of input parameters were decided based on the external look of weld bead such as smooth continuous bead, absence of defects like porosity, undercut etc. by visual inspection. Numerous attempts have been recommended to develop mathematical models relating input process parameters and weld bead geometry. Recently, Artificial Intelligence (AI) methods such as fuzzy logic, Artificial Neural Networks (ANN) and expert system have been deployed as key techniques for monitoring and controlling welding processes. An ANN model was developed by Abdullah Alfaruk et al. 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 successfully [8]. ANN modeling has been chosen by its capability to solve complex and difficult problems. Kim et al. used multiple regression analysis and back propagation neural network in modeling bead height in metal arc welding [9]. They comprehended that the back propagation neural network is considerably more accurate than multiple regression. Nagesh and Datta reported that artificial neural networks are powerful tools for analysis and modeling. They applied back propagation neural network to predict weld bead geometry and penetration in shielded metal arc welding [10]. In present scenario, ANN models have been used by many researchers to understand and predict their targeted information This paper presents the development of neural network model to predict depth of penetration for various input process parameters in mild steel butt welding deposited by SMAW. Multilayer feed forward neural network was developed and it was trained using back propagation algorithm.

2. EXPERIMENTAL WORK
To investigate the weldment characteristics weld beads were obtained by welding two mild steel flat plates of 150 mm x 50 mm x 5 mm dimensions in butt position using mild steel electrodes of 3.15 mm diameter. A manual welding machine was used to weld the plates. A lathe machine was used to provide uniform speed of welding to support electrode holder and bar magnet. The work piece was kept on cross slide with some arrangement. Work-piece moves with cross slide. Bar magnet was connected with tailstock with a wooden structure Since the weldment characteristics depend on welding current, welding voltage, speed of welding and magnetic field, we select different sets of values of these inputs [11]. Welding currents were chosen as 90, 95,100, 105 and 110 A, arc voltages were chosen as 20, 21, 22,23 and 24V, the welding speeds were chosen as 40, 60 and 80 mm/min and external magnetic field strengths were used as 0, 20,40, 60 and 80 Gauss for the experiments. The experimental set-up is shown in figure-1 [12].Current was measured with a clamp meter, voltage was measured with a multi meter and magnetic field was measured with a Gauss meter. To study the bead geometry, each bead was sectioned transversely at two points one near the start (leaving 2 cm from the start) and the other near the end (leaving 2 cm from the end). To get the microstructure, these sectioned beads were ground with emery belt grinder having 0, 2, 3 grade emery papers then polished with a double disk polishing machine. Etching was done with a mixture of 2% nitric acid and 98% ethyl alcohol solution. The average value of depth of penetration was measured. Eighteen sets of values out of twenty five such sets obtained were used for training a network based on back propagation algorithm. Remaining seven sets of the values were used for prediction. These data sets are shown in table-1. A program of back propagation neural network in C++ was used for training and prediction. In this program one input layer having four neurons, two hidden layers, both having five neurons and one output layer having one neuron, were used, which are represented in figure-2 [13].

Volume 1, Issue 4, December 2012

Page 13

International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847
Table 1: Data for Training and Prediction
Magnetic Field (Gauss) 0 20 40 60 80 60 60 60 60 60 40 40 40 20 20 20 20 20 0 40 60 40 40 20 20 0.79 0.79 0.80 0.77 0.76 0.78 0.76 0.75 0.74 0.72 0.83 0.79 0.76 0.70 0.71 0.74 0.77 0.75 0.76 0.72 0.66 0.78 0.77 073 0.75 Depth of Penetration (mm)

Serial Number

Current (A)

Voltage (V)

Welding Speed (mm/min) 40 40 40 40 40 60 60 60 60 60 40 60 80 60 60 60 60 60 40 60 80 40 60 60 60

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

90 90 90 90 90 95 95 95 95 95 100 100 100 90 95 100 105 110 90 95 95 100 105 105 110

24 24 24 24 24 20 21 22 23 24 22 22 22 20 20 20 20 20 23 22 21 24 21 22 21

Table 2: Measured and Predicted Values with percentage Error

S.N.

Current (A)

Voltage (V)

Welding Speed (mm/min)

Magnetic Field (Gauss)

Depth of Penetration (mm) Measured 0.76 0.72 0.66 0.78 0.77 0.73 0.75

Depth of Penetration (mm) Predicted 0.74 0.71 0.70 0.73 0.74 0.72 0.74

Error in Depth of penetration % age -2.63 -1.39 +6.06 -6.41 -3.90 -1.37 -1.33

1 2 3 4 5 6 7

90 95 95 100 105 105 110

23 22 21 24 21 22 21

40 60 80 40 60 60 60

0 40 60 40 40 20 20

Volume 1, Issue 4, December 2012

Page 14

International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847 3. RESULTS
Table-2, depicted depth of penetration from the experiment and predicted output values using artificial neural feed forward network. The measured and predicted output values are close to each other. The aim of this paper shows the possibility of the use of neural networks to predict the weld bead geometry.

1. Multi-meter 5. Table 9. Tail Stock 13. Tool post 17. Electrode Holder 20. Connecting Wires

2. Battery Eliminator 3. Electric Board 6. Measuring Prob 7. Transformer Welding Se 10. Sleeve 11. Link (Wood) 14. Iron sheet 15. Workpiece 18. Metal Strip Connected with head stock Figure 1 Experimental Setup

4. Gauss Meter 8. Clamp meter 12. Solenoid 16. Electrode 19. Head stock

I/P La yer

Hidden La yer

Hidden La ye r

O/P La ye r

Curren t

Voltage Dept h of Pen etratio n Travel s peed

Magnetic field

Figure 2 Feed-forward neural network (4-5-5-1) architecture 3.1 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.

Volume 1, Issue 4, December 2012

Page 15

International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847
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 3, 4, 5 and 6 respectively
0.805 ? 0.8 ) m 0.795 m ( n0.79 o i 0.785 t a r t0.78 e n 0.775 e P f 0.77 o h 0.765 t p e0.76 D 0.755 0 20 40 60 80 100

0.79
? 0.78 ) m 0.77 m ( n o i0.76 t a r 0.75 t e n 0.74 e P f 0.73 o h t 0.72 p e 0.71 D 19 20 21 22 23 24 25

Magnetic Field (Gauss) ? Figure 3: Magnetic Field vs Depth of Penetration At 90 A, 24 V and 40 mm/min.

Voltage(V) ? Figure 4: Voltage vs Depth of Penetration At 95 A, 60 mm/min and 60 Gauss

0.84 ? ) 0.83 m m ( 0.82 n o i0.81 t a r 0.8 t e n 0.79 e P f 0.78 o h 0.77 t p e 0.76 D 0.75 20 40 60 80 100 Welding Speed (mm/min) ? Figure .5: Welding Speed vs Depth of Penetration At 100 A, 22 V and 40 Gauss 0

0.78 0.77 ?0.76 ) m 0.75 m ( 0.74 n o i t0.73 a r t e 0.72 n e P f0.71 o h 0.7 t p e 0.69 D

90

100

110

Current (A) ? Figure 6: Current vs Depth of Penetration at 20 V, 60 mm/min and 20 Gauss

3.1 Prediction of Depth of Penetration 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 6.06. 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.

4. DISCUSSION
In this investigation, an attempt was made to obtain 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 depth of penetration. 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

Volume 1, Issue 4, December 2012

Page 16

International Journal of Application or Innovation in Engineering & Management (IJAIEM)

Web Site: www.ijaiem.org Email: editor@ijaiem.org, editorijaiem@gmail.com Volume 1, Issue 4, December 2012 ISSN 2319 - 4847
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.

5. CONCLUSION
The experimental analysis confirms that, artificial neural networks are powerful tools for analysis and modeling. Results revealed that an artificial neural network is one of the alternatives methods to predict the weld bead geometry in terms of depth of penetration. 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, depth of penetration generally increases. 3. If voltage of the arc is increased, depth of penetration decreases. 4. If travel speed is increased, depth of penetration of weld decreases. 5. If magnetic field is increased, depth of penetration of weld decreases. 6. Artificial neural networks based approaches can be used successfully for predicting the output parameters like weld width, reinforcement height and depth of penetration of weld, however the error is rather high as in some cases in predicting the depth of penetration it is more than 6 percent. Increasing the number of hidden layers and iterations can be used to minimize this error.

REFERENCES
[1] O.P. Khanna, A text book of welding technology, Dhanpat Rai Publications Ltd., 2006, P3. [2] I.S. Kim et al, An investigation into an intelligent system for predicting bead geometry in GMA welding process. J Mater Process Technol 2005. [3] J.I. Lee and K.W. Um, A prediction of welding process parameters by prediction of weld-bead geometry. J Mater Process Technol, 2000. [4] P.J. Modenesi and R.C. Avelar, The influence of small variations of wire characteristics on gas metal arc welding process stability. J Mater Process Technol, 1999. [5] J.C. Mc Glone and D.B. Chadwick, The submerged arc butt welding of mild steel. Part 2. The prediction of weld bead geometry from the procedure parameters. Welding Institute Report, 80/1978/PE; 1978. [6] R.S. Chandel, Mathematical modeling of gas metal arc weld features, Proceedings of the forth international conference on modeling of casting and welding processes, Palm Coast (FL); 17 22 April, 1988. pp. 109120. [7] R.L. Little, Welding and Welding Technology, McGraw Hill Book Company, New York, 1973. [8] Al-faruk Abdullah et al, Prediction of Weld Bead Geometry and Penetration in Electric Arc Welding using Artificial Neural Networks. Int. Jour. of Mech. & Mechatronics Engg 10 No: 04. [9] L.S. Kim, S.H Lee, and P.K.D.V Yarlagadda, 2003. Comparison of multiple regression and back propagation neural network approaches in modeling top bead height of multipass gas metal arc welds. Sci. & Tech.of welding and joining, 8(5), pp. 347-352. [10] D.S. Nagesh and G.L. Datta, 2002. Prediction of weld bead geometry and penetration in shielded arc welding using artificial neural networks. J. of Matr. Proc. Tech, 123(2) pp. 303-312. [11] Md. Ibrahim Khan, Welding Science and Technology New Age International (P) Limited Publishers, 2007. [12] R.P. Singh et al., Prediction of Weld Bead Geometry in Shielded Metal Arc Welding under External Magnetic Field using Artificial Neural Networks , International Journal of Manufacturing Technology and Research, Vol. 8 number 1, pp. 9-15, 2012. [13] R.P. Singh et al., Application of Artificial Neural Network to Analyze and Predict the Mechanical Properties of Shielded Metal Arc Welded Joints under the Influence of External Magnetic Field, International Journal of Engineering Research & Technology (IJERT), pp. 1-12, Vol. 8, Issue 1, October, 2012. AUTHOR Rudra pratap singh received the B.E. degree in Mechanical Engineering from MMMEC Gorakhpur in 1992 and the M.Tech. degree in mechanical Engineering in 2009 from UPTU Lucknow. During 1992 to 1999 he worked in Jindal Group as a quality control engineer, from 1999 to till date he is working in GLA group (now GLA University) Mathura as a faculty in Mechanical Engineering Department. He is pursuing Ph. D. (Registered in, March, 2010) from Uttarakhand Technical

Volume 1, Issue 4, December 2012

Page 17