STUDY THE EFFECT OF CO2 MAG WELDING PROCESS PARAMETERS ON THE HEAT INPUT AND JOINT GEOMETRY DIMENSIONS USING EXPERIMENTAL AND COMPUTATIONAL METHODS

-In this paper, predicted models for heat input and joint geometry dimensions after CO2MAG welding process have been developed. Before welding, steel specimens were first prepared and then butt welded using electrode wire melted and supplied into the molten pool by applying heat input continuously. Weld bead dimensions were first measured, and then the results were analyzed to check the adequacy of the models by Response Surface Method using DOE technique. These models were found capable of predicting the optimum performance dimensions required for the joint geometry in terms of weld bead width, reinforcement height and penetration. The obtained results indicated that the heat input depends on voltage, wire feed speed and gas flow rate, while for the weld bead dimensions; the gas flow rate has less effect. A comparison between the experimental and predicted results was made, and a good agreement was found between them.


INTRODUCTION
CO2-MAG is an arc welding process where heat is generated for arc between the workpiece and a consumable metal electrode with an externally supplied gaseous shield of gas either inert, such as CO2. Itis a versatile process, gives very little loss of alloying elements and can be operated as semi as well as fully automated. A bare solid wire called electrode is continuously fed to the weld zone, it becomes filler metal as it is consumed.
Electrical energy is supplied from the welding generator for melting wire and workpiece to be welded. The weld is made by falling successive drops on the weld puddle. The arc and the molten puddle are protected from contamination by the atmosphere (i.e., oxygen and Diyala Journal of Engineering Sciences flow rates. In addition, gas density, or the weight of the gas relative to air, has a major influence on the minimum flow rate required to effectively shield the weld (3) .
Welding with the recommended heat input results in good mechanical properties in the heat affected zone (HAZ). The heat supplied by the welding process affects the mechanical properties of the welded joint. Heat input can be referred to as "the electrical energy supplied by the welding arc to the workpiece. The most important characteristic of heat input is that it governs the cooling rates in welds and thereby affects the microstructure of the weld metal and the heat affected zone. A change in microstructure directly affects the mechanical properties of welds. Therefore, the control of heat input is very important in arc welding in terms of quality control (4) .
Quality of the welded joint in CO2-MAG welding process depends on number of parameters, like type and thickness of base metal, design type, welding position, etc., but the proper selection of welding parameters is also very important. Due to that, the weld bead geometry in CO2-MAG welding process and heat input with regard to weld voltage, wire feeding speed and gas flow rate were experimentally investigated in the present work, since the proper selection of gas flow and heat input will provide a weld joint with satisfactory geometrical characteristics (5) .

49
A large amount of research works have been carried out to find out the most suitable combination of input process parameters for a desired output using different welding processes and various computer software as tools for modeling and optimization the weld bead geometry, such as Taguchi (6) , Artificial neural networks (ANN) (7) , and Response surface methodology (RSM) (8) . Das et al. (9) studied the effect of arc voltage, current and welding speed on the weld joint geometry, while Shoeb et al. (10) considered also the influence of gas flow rate. In addition, Patel and Patel (11) investigated also the wire diameter and wire feed rate during CO2-MAG welding process. But, there is a little work about modeling and computational optimization of the closed butt weld bead geometry by using Design of Experiment (DOE) with (RSM) technique to predict mathematical models that can be used to obtain the optimum responses for any given input parameters. Therefore, the aim of this paper is to study the influence of main welding parameters (voltage, wire feeding speed and gas flow rate) on the heat input and final weld pool geometry during CO2-MAG welding using DOE and RSM method.

Material and Specimens Preparation
The material used in the present work is low carbon steel (LCS) plate with 5 mm thickness in the hot rolled condition. This material was chemically analyzed in State Company for Inspection and Engineering Rehabilitation (SIER) in Baghdad, and its chemical composition is given in Table (1), showing that the experimental material conforms to the standard low carbon steel type AISI 1010 (12) . The plate was cut to provide specimens with size 50 mm× 25 mm×5 mm to be welded in a closed Butt weld joint design by CO2-MAG process. Specimens from the as-received material were tensile tested according to ASTM E8 in Strength Laboratory / University of Technology-Baghdad, and the results are given in Table (2).The results in this table represent the average of three readings (three samples).

Selection of Welding Parameters
Despite the use of CO2-MAG welding process is influenced by number of Parameters, three of them were only selected in this investigation: voltage, wire feeding speed and gas flow rate in two levels (input parameters), as shown in Table (3). These parameters were chosen according to the capacity of CO2-MAG welding machine and practical experience of the welder skill.

Welding Procedure
Twenty specimens were welded by CO2-MAG process at different values of voltage, wire feeding speed and gas flow rate according to design matrix established by Design of (1) together with the specimens before and after welding process.

Measurements of Joint Geometry Dimensions
After each welding test, the weldment were cut, sectioned, ground, polished and finally etched to see the profile of the joint geometry with necessary dimensions for measuring purpose, which is schematically similar to that was shown in reference (13) , see  Table (4).
Since the heat input parameter has a significant effect on the quality of the joint geometry, therefore it was decided to calculate the values of heat input for all weldments by using the following equation For modeling and optimization the heat input at the same levels of used voltage, wire feed speed and gas flow rate, the current reading was taken during the welding process from the machine. Also, the welding speed was calculated for each test. Therefore, the heat input value was calculated for each welding test taking into account that the thermal efficiency is equal to 0.8 for MAG welding type (5) . The results of calculated welding speed and heat input are listed in Table (

RESULTS & DISCUSSION
The response surface methodology was achieved using the Design of Expert version 8 software to determine the predicted models for the dimensions of the weld joint geometry 51 very small (< 0.5). This means that these two parameters contributed the highest effect on the weld joint geometry, while the gas flow rate has no influence on the bead width and reinforcement height [Tables (6 and 7)], since the gas flow rate term (C) is not in the model, except that it affects the penetration depth due to the appearance of this term in the model, as shown in Table ( 8).
The ANOVA analyses also pointed out that the quadratic effect was useful to incorporate into bead width and reinforcement models, since the second order terms were highly significant with a P-value lower than 0.05. In addition, it was noticed in Table ( other factors held constant at the reference value. Accordingly, the perturbation plots for these three models are illustrated in Fig. (6). So, this figure indicates that, individually, both voltage and wire feeding speed largely affect the bead width and reinforcement, but they have a slight influence on the penetration. This is likely due to the higher heat input that increased the fusion of the material at the top surface of the joint. While, the gas flow rate has no effect on the bead width and reinforcement but slightly affect the penetration and this is may be due to the chemical affinity of the CO2 gas with the molten material of the joint.
Since the diagnosis of the residuals reveals no statistical problems with the models, so the design of experiment generates the response surface plots in form of 2D contour, 3D surface and cube plots. Figures (7 and 8) show the 2D contour plots for the bead width and reinforcement, respectively as a function of voltage and wire feeding speed at gas flow rate of 10 L/min. It was found that welding at a gas flow rate of 8 and 12 L/min had no effect on these responses. It can be noticed from

Modeling of Heat Input
Similarly for the heat input, the analysis of variance (ANOVA) for response quadratic model was constructed by DOE software as given in Table ( 53 the predicted and actual values for heat input, as depicted in Fig. (10). so, the final predicted equation for the heat input in terms of the actual input factors is: Heat input = -7.06339 -0.090188 * Voltage + 0.080047 * Wire feeding speed+ 0.74389 2 .. . (5) 0.039065 * Gas flow rate -2 004 * Wire feeding speed -2.42014E -* Gas flow rate In order to diagnose the statistical properties of this model, it was found that the residuals that falling on a straight line implying errors are normally distributed. Also, the residuals versus predicted actual for heat input data exhibited no obvious pattern or unusual structure implying models are accurate.
To gain perspective on the model, it is necessary to present the perturbation of the predicted response resulted by varying only one parameter at a time from the center point of the investigated region. Fig. (11) demonstrates the perturbation plot for the heat input model, indicating that, individually, all input parameters affect the heat input response. The wire feeding speed largely increased the heat input because of more molten material accumulated in the weld joint at higher feeding speed, whereas both voltage and gas flow rate slightly reduced the heat input due to the higher wire speed and higher and more chemical reaction of CO2 gas with the higher accumulated molten metal at the weld joint.
Because the diagnosis of the residuals manifested no statistical problems, the response surface plots were generated in terms of 3D surface plot, since all input parameters are significant in this model. Fig. (12) Depicts the 3D surface plot for the heat input response as a function of voltage and wire feeding speed at various gas flow rates. This figure shows the wire feeding speed is more effective on the heat input response at 10 L/min gas flow rate [ Fig. (12b)] because of the higher molten material accumulated in the weld joint at higher feeding speed. Whereas both voltage and gas flow rate have a slight influence on heat input, and this is possibly ascribed to the higher wire speed and more chemical reaction of CO2 with the more accumulated molten material in the weld joint. Finally, these observations are confirmed by the cube plot for penetration, as shown in Fig. (12d) for the heat input response.

Computational Optimization
A computational optimization method was used in this work by selecting the desired goals for each factor and response. This computational optimization is provided by the Design of Experiment software to find out the optimum combinations of parameters in order to fulfill the requirements as desired. Therefore, this software used for the optimization purpose; based on the data from the predictive models for four responses, weld bead width, reinforcement height, penetration and heat input as a function of three factors: arc voltage, wire feeding speed and gas flow rate. The computational optimization process involves combining the goals into an overall desirability function. To develop the new predicted models, a new objective function, named 'Desirability' which allows to properly combining all the goals, was evaluated. Desirability is an objective function, to be maximized through a computational optimization, which ranges from zero to one at the goal. A higher value for desirability indicates the response value is more desirable. If it is equal to zero, this means a completely undesired response (14) .
Adjusting its weight or importance may alter the characteristics of a goal, and the aim of the optimization is to find a good set of conditions that will meet all the goals. Usually, the weights are used to establish an evaluation of the goal's 3Dimportance when maximizing desirability function; in this work, weights are not changed since the four responses have the same importance and are not in conflict within each other.
The ultimate goal of this optimization was to obtain the maximum response that simultaneously satisfied all the variable properties.  Fig. (13) shows the 3D surface plot for desirability as a function of voltage and wire feeding speed at 8 L/min gas flow rate.

RSM achieved by DOE technique has shown its effectiveness and usefulness as a tool to
predict the responses in MAG-CO2 welding technique for any given input parameters.
2. Quadratic models were obtained by RSM achieved by DOE technique for the optimum heat input with the optimum dimensions of the weld joint geometry of the welded parts by the CO2-MAG process.
3. The arc voltage and wire feeding speed are found the most effective welding parameters in the predicted quadratic models of weld bead width and reinforcement height, while gas flow rate is only influential in the predicted models of bead penetration and heat input.