Machine Learning Based Prediction of Surface Quality in Friction Stir Welding

In this study, Al-5083 and St-12 alloy sheets were friction stir lap welded at different travel (7–23 cm min−1) and rotation (750–1125 rev min−1) speeds of the welding tool. The welded joints were characterised by various methods including shear tensile and Vickers microhardness tests, optical and scanning electron microscopies and X-ray diffraction analysis. The results showed that the weld zone defects decreased, and the joint strength also improved significantly with the reducing tool travel speed from 23 to 7 cm min−1. The travel speed of 11 cm min−1 was an optimum speed within the travel speeds investigated in this work. Additionally, raising the tool rotation speed enhanced the joint strength slightly. The microhardness results confirmed the formation of intermetallic phase layer with the hardness of ∼335 HV at the swirl layered structure.

• Accurate prediction of weld quality in friction stir welding process of high-density polyethylene. • Combining Coupled Eulerian-Lagrangian (CEL) simulations with machine learning (ML) models. • Both CEL and ML models closely match experimental results. • Simulations accurately replicated flow mixing patterns observed in experiments. Friction stir welding (FSW) has gained recognition as an effective technique for joining thermoplastic materials. This study presents a comprehensive investigation to predict FSW characteristics such as weld temperatures, vertical axial forces, material flow and defect formation in FSW of high-density polyethylene (HDPE) using both computational and machine learning approaches, validated by experimental results. A three-dimensional Coupled Eulerian-Lagrangian (CEL) numerical model is developed to simulate the HDPE welding process under varying process parameters of rotational and welding speeds, and bottom preheating conditions. The CEL numerical model is validated against experimental data, showing markedly good agreement in predicting void formation, thermal and mechanical responses. Results indicate that preheating the bottom surface proved effective in reducing thermal gradients and stabilizing vertical force fluctuations. Quantitatively, bottom-surface preheating at 60 °C reduced peak plunge force from 1600 N close to 1100 N and raised peak temperatures from 70 °C (without preheat) to about 100 °C, hence smoothing through-thickness gradients and stabilizing the axial force distribution. In parallel, two multilayer perceptron artificial neural networks (ANNs) are trained on experimental data to predict temporal temperature and axial force responses with R 2 between 0.91 and 0.99, demonstrating high accuracy and strong generalization across the tested FSW parameter space. The computational CEL simulations accurately predicted weld zone morphology, including stir zone, thermo-mechanically affected zone and heat-affected zone, however The combined CEL–ANN framework enables rapid generation of process maps that highlights defective and defect-free welding windows, with optimal conditions obtained at 1500 rpm and 30 mm/min with 60 °C preheating, and poor weldability observed at 1000 rpm without preheating due to wormhole formation. These findings provide process-driven design guidelines for developing robust HDPE joints and demonstrate how integrating computational modeling with machine learning enables efficient optimization, offering valuable insights into the key characteristics of the FSW process in thermoplastics and supporting its broader adoption in industrial applications.

Friction stir welding process has been studied extensively in the last decades since its early stage. Most of the research done so far is related to the process development including tool design, material weldability, post-weld mechanical behavior, and microstructural properties. More recently, in-line process monitoring and artificial intelligence algorithms are introduced into this process, but mainly to specific material configuration and joint thicknesses. This study will focus on the evaluation of different machine learning approaches including principle component analysis, K-nearest neighbor, multilayer perceptron, single vector machine, and random forest methods on a friction stir welding cell environment. The input variables provided from this cell environment are namely divided into two groups: one group refers to the application variables and the other group is related to the friction stir welding process variables. The application variables target the aluminum alloys, joint configuration, sheet thicknesses, initial mechanical properties, and their chemical composition. The friction stir welding process variables dictate the rotational speed, travel speed, forging force, longitudinal and transverse forces, torque, and specific energy. The output response to model from these machine learning algorithms is the defect index, which has been quantified using high-resolution immersed bath ultrasounds. This nondestructive evaluation technique has been described previously, which can detect defects ≥150 µm in thin sheets. The defect index has been classified into five classes, which is distinguished by the nature of defect, cold weld, or hot weld, as well as the width of the internal volumetric defect upon ultrasound C-scan result. The dataset, which is composed of around 500 various process conditions, has been generated over the last few years and the variables were taken exclusively in constant weld regime and in the force control mode using the output average values. This paper compares the best resulting machine learning methods applied on a friction stir welding cell basis, which is the K-nearest neighbor and multilayer perceptron algorithms. The K-nearest neighbor model reaches a deviation of 0.55 on the defect index in comparison with the experimental values, which is slightly better than the multilayer perceptron model, which obtains a score of 0.69. Over the initial 59 available model parameters, 10 and 15 of them were retained in the final algorithm using these techniques. The main predictors include the material thickness, base material ultimate tensile stress, rotational speed, travel speed, weld forces, and specific energy. The K-nearest neighbor model was able to provide a map of defect indices with regard to rotational speed and travel speed but was only possible when a higher density of data was found within the prediction area. A data density score was also included within the model to inform the end-user about the prediction reliability. The machine learning models are mainly about differentiating various cases rather than representing the physical phenomena as determined using the finite element analysis. That being said, in order to improve the prediction reliability as well as the machine learning models, the data twinning concept, which consists of generating simulated friction stir welding process conditions by finite element analysis, is briefly discussed.

Prediction of Tensile Strength in Friction Stir Welded Aluminium Alloy Using Artificial Neural Network

Process optimisation and mechanical properties of friction stir lap welds of 7075-T6 stringers on 2024-T3 skin

Invented in 1991, friction stir welding (FSW) is a new solid state joining technique. This process has many advantages over fusion welding techniques including absence of filler material, shielding gas, fumes and intensive light, solid state joining, better microstructure, better strength and fatigue life, and etc. The difficulty with FSW is in the high forces involved especially in axial direction which requires use of robust fixturing and very stiff FSW machines. Reduction of FSW force would simplify implementation of the process on less stiff CNC machines and industrial robots. In this paper axial welding force reduction is investigated by use of tool design and welding parameters in FSW of 3.07 mm thick AA6061-T6 sheets at right angle. Attempt is made to reduce the required axial force while having acceptable ultimate tensile strength (UTS). It is found that shoulder working diameter and shoulder angle are the most important parameters in the axial force determination yet pin angle has minor effect. According to the developed artificial neural network (ANN) model, proper selection of shoulder diameter and angle can lead to approximately 40% force reduction with acceptable UTS. Regions of tool design and welding parameters are found which result in reduced axial force along with acceptable UTS.

Prediction of multiple characteristics of Friction-Stir welded joints by Levenberg Marquardt algorithm based artificial neural network

Friction stir welding is an alternative to traditional welding processes, that can be considered both environmentally friendly and a tool to face weldability problems, specially in aluminium alloys. This process uses a particularly design tool that rotates and travels along the join, using equipment similar to a milling machine. Control in friction stir welding helps to assure weld soundness if done correctly. Different control approaches have been developed using force, torque and temperature sensors as input variables. Depending on the application this control strategies are successful while using dedicated FSW equipment, looking into local implementation it is more viable and cost effective to employ an adapted milling machine for the job. This type of mechanism only allows setting parameters such as travel speed, rotational speed and depth of penetration, which can be insufficient to obtain sound welds. Based on the set outcome of local implementation model-based control is an alternative to classic control and it uses a set of equations that describe the process as the input. The proposed model consists of a set of algebraic balance equations describing how mass and energy, in terms of power consumption, transform along the process predicting the soundness of the joint in every step. This model serves as the base for a control system that can predict the soundness in the next step and react accordingly, modifying basic process parameters such as plunge depth as well as travel and rotational speed minimizing mass loss. Experimental data has been compared to the model output in order to validate it.

Investigating the robust bonding between dissimilar materials such as aluminium alloys and polymers is vital in welding processes. This research delves into the structural interface and mechanical attributes achieved through aluminium alloy 6061 (AA6061) and Polypropylene (PP) friction stir welding by varying traverse and rotational speeds. A travel speed was varied with a range of 50–80 mm/min, and the tool rotational speed between 1300 and 1900 rpm was investigated to analyse their impact on mechanical properties. The lap joint configuration in which the PP is at the top and AA6061 is at the bottom was chosen from trial experiments. A scanning electron microscope observed that many locations had a tight bond in most interface areas. The average hardness values at the stir zone in the joints at different rotational speeds initially increased up to 1700 rpm due to fine fragmentation and proper distribution of these aluminium chips, and then reduced after rising to 1900 rpm. Evaluation of weld quality encompassed assessments of weld nugget hardness and lap tensile shear strength, elucidating the mechanical interlocking mechanism between AA6061 particles and PP as a critical factor contributing to joint stability. The intersection of aluminium chips and solidified polymer contributes to enhancement in the shear strength due to the phenomenon known as mechanical interlocking. The highest tensile shear strength of 5.4 MPa was observed at a rotational speed of 1700 rpm and a traverse speed of 60 mm/min, signifying optimal process conditions for enhancing mechanical integrity.

'Friction stir welding' (FSW) is an eco-friendly, energy-efficient solid-state welding procedure invented for high strength alloys as well as materials that are otherwise difficult for welding through the traditional fusion welding methods. This research paper elaborates on a hybrid Taguchi-genetic algorithm that optimises the FSW process parameters to yield favourable electrical conductivity of aluminium alloy AA1350. The optimisation parameters considered are travel speed, rotational speed and tool tilt angle. First, the Taguchi method is applied to reduce the number of design experiments and find the optimal set for quality parameters of the system. Subsequently, the genetic algorithm is employed to search for the optimum set of design parameters by using Taguchi solution as the initial population. The results obtained by the proposed method were compared to the outcomes of the conventional Taguchi L9 method and the simple genetic algorithm. The best performance was obtained by the proposed method in 51 iterations, whereas simple genetic algorithm used 182 iterations for the same.

This paper is to study of dissimilar welding AA6061 aluminium alloy and AZ31B magnesium alloy with ER5356 filler using friction stir welding. 2 mm thick plates of aluminium and magnesium were used. Friction stir welding operations were performed at different rotation and travel speeds and used the fixed tilt angle which is 3°. The rotation speeds varied from 800 to 1100 rpm, and the travel speed varied from 80 to 100 mm/min. In the range rotation speed of 800 to 1000 rpm and welding speed of 80 to 100 mm/min there are no defect at the weld. Tensile test show the higher tensile strength is 198 MPa and the welding efficiency is about 76%.

This paper reports on the employment of the machine learning (ML) techniques, namely support vector machine (SVM), artificial neural networks (ANN), and random forest (RF), for predicting the tensile behavior of friction stir processed (FSP) dissimilar aluminium alloys joints (6083-T651 and 8011-H14). The dissimilar aluminium joints are fabricated using the friction stir welding (FSW) process. After that, the friction-stir welded joints are subjected to the FSP procedure at different combinations of process parameters. The rotational speed, traverse speed, and tilt angle are used as the input parameters, while tensile strength and grain size are used as the output parameters. In addition, three performance characteristics (i.e., coefficient of correlation (CC), mean absolute error (MAE), and root mean square error (RMSE)) are used to check the adequacy of the developed model of ML techniques. It is observed that support vector machine_radial basis function kernel is the most accurate modeling technique for predicting the tensile behavior of processes samples. Furthermore, the optical microscope is also utilized to check the grain size of the nugget zone (NZ) of the weld bead for FSP. It is found that the minimum grain size (i.e., 5.06 µm) is obtained for the FSP sample and this grain size corresponded to the high ultimate tensile strength (UTS). Moreover, the fractographic analysis showed the ductile behavior of FSW and FSP samples.

Mechanical characterization of friction stir welded dissimilar aluminium alloy using Taguchi approach

Friction stir welding (FSW) is a solid state welding process in which a non-consumable rotating tool with a specially designed pin and shoulder is inserted into the abutting edges of sheets or plates to be joined and subsequently traversed along the joint line. In FSW, a pin tool with different shapes spins against the edges of two metal pieces of same or different thickness positioned next to each other. As the pin travels along, it creates friction that generates heat, mixes, and joins the alloys without melting them. To optimize the process, several researchers created pins of different shape, and geometry, and used them in FSW but varied the depth, rotational speed, and tilt angle of the pins. Statistical analysis has been used to identify the most optimum combination of tool and weld parameters that could support high-speed production. Many studies support that the faster FSW is carried out, the stronger (better weld quality) the resulting welds will be. The objective of this paper is to predict the effects of some of the process parameters on the performance of the aluminum alloy components joined using ANSYS simulation tool. Although not reported in this paper, the mechanical and metallurgical properties of the welded members have been measured in the laboratory. The goal is to gain an understanding of how FSW can be used to successfully join aluminum alloys and to study the effect of the various process parameters on the process. The material used is AA6061 as it is one of the popular choices for automotive applications. Experiments have been conducted to validate some of the simulation results from ANSYS software.

Friction stir welding (FSW) is gaining traction as a preferred technique due to its potential to reduce heat input and enhance the mechanical properties of welded joints. However, the path to commercializing FSW for flange joints is not without challenges. Two primary obstacles are the complexity of the welding path and the intricate design requirements for the fixtures. These factors contribute to the difficulty in determining the ideal weld settings and process parameters, which are critical for achieving optimal results. The current study addresses these challenges by applying FSW to flange joints using custom-engineered fixtures. These fixtures are meticulously designed to hold the pipes and plates securely during the welding process. The focus of the research is on optimizing the multi-performance characteristics of FSW for Al 6063 flange joints through the hybrid Grey-based Taguchi method. The integrity of the weld joint is assessed by examining various mechanical properties within the weld zone, including rotation speed, travel speed, tool profile, and shoulder diameter. The study identifies the optimal parameter settings for the FSW process: a rotation speed of 3000 rpm, a travel speed of 3 mm/min2, a shoulder diameter of 20 mm, and a conical tool profile. Under these ideal conditions, the welded material exhibited a tensile strength of 170.169 MPa, a hardness of 63.7709 HV, and a corrosion rate of 0.022 mm/year. These findings underscore the effectiveness of the optimized FSW process in producing robust and durable flange joints.