This paper presents an experimental investigation of laser overlap welding of low carbon galvanized steel. Based on a structured experimental design using the Taguchi method, the investigation is focused on the evaluation of various laser welding parameters effects on the welds quality. Welding experiments are conducted using a 3 kW Nd:YAG laser source. The selected laser welding parameters (laser power, welding speed, laser fiber diameter, gap between sheets and sheets thickness) are combined and used to evaluate the variation of three geometrical characteristics of the weld (penetration depth, bead width at the surface and bead width at the interface). Various improved statistical tools are used to analyze the effects of welding parameters on the variation of the weld quality and to identify the possible relationship between these parameters and the geometrical characteristics of the weld. The results reveal that the reached hardness values are similar for all the experimental tests and all welding parameters are relevant to the weld quality with a relative predominance of laser power and welding speed. The effect of the gap is relatively limited. The investigation results reveal also that there are many options to consider for building an efficient welds quality prediction model. Results achieved using an artificial neural network based simplified model provide an indication of the prediction model performances.

Laser welding (LW) becomes one of the most economical high quality joining processes. LW offers the advantage of very controlled heat input resulting in low distortion and the ability to weld heat sensitive components. To exploit efficiently the benefits presented by LW, it is necessary to develop an integrated approach to identify and control the welding process variables in order to produce the desired weld characteristics without being forced to use the traditional and fastidious trial and error procedures. The paper presents a study of weld bead geometry characteristics prediction for laser overlap welding of low carbon galvanized steel using 3D numerical modelling and experimental validation. The temperature dependent material properties, metallurgical transformations and enthalpy method constitute the foundation of the proposed modelling approach. An adaptive 3D heat source is adopted to simulate both keyhole and conduction mode of the LW process. The simulations are performed using 3D finite element model on commercial software. The model is used to estimate the weld bead geometry characteristics for various LW parameters, such as laser power, welding speed and laser beam diameter. The calibration and validation of the 3D numerical model are based on experimental data achieved using a 3 kW Nd:Yag laser system, a structured experimental design and confirmed statistical analysis tools. The results reveal that the modelling approach can provide not only a consistent and accurate prediction of the weld characteristics under variable welding parameters and conditions but also a comprehensive and quantitative analysis of process parameters effects on the weld quality. The results show great concordance between predicted and measured values for weld bead geometry characteristics, such as depth of penetration, bead width at the top surface How to cite this paper: Oussaid, K. and El Ouafi, A.

Predictive modelling for quality analysis becomes one of the most critical requirements for a continuous improvement of reliability, efficiency and safety of laser welding process. Accurate and effective model to perform non-destructive quality estimation is an essential part of this assessment. This paper presents a structured approach developed to design an effective artificial neural network based model for predicting the weld bead dimensional characteristic in laser overlap welding of low carbon galvanized steel. The modelling approach is based on the analysis of direct and interaction effects of laser welding parameters such as laser power, welding speed, laser beam diameter and gap on weld bead dimensional characteristics such as depth of penetration, width at top surface and width at interface. The data used in this analysis was derived from structured experimental investigations according to Taguchi method and exhaustive FEM based 3D modelling and simulation efforts. Using a factorial design, different neural network based prediction models were developed, implemented and evaluated. The models were trained and tested using experimental data, supported with the data generated by the 3D simulation. Hold-out test and k-fold cross validation combined to various statistical tools were used to evaluate the influence of the laser welding parameters on the performances of the models. The results demonstrated that the proposed approach resulted successfully in a consistent model providing accurate and reliable predictions of weld bead dimensional characteristics under variable welding conditions. The best model presents prediction errors lower than 7% for the three weld quality characteristics.