This study proposes an integrated approach for optimizing the friction stir welding (FSW) process by combining sensor-based data acquisition, machine learning (ML), and digital twin (DT) technologies. Real-time sensor data including rotational speed, welding speed, axial force, torque, and temperature were collected during FSW operations. These parameters were correlated with weld quality indicators, such as surface appearance, internal defects, and tensile strength. A dataset of 132 weld samples was used to train supervised and unsupervised ML models, achieving a defect classification accuracy of 95%. In parallel, a COMSOL-based digital twin was developed to simulate thermo-mechanical aspects of the welding process. The model incorporated temperature-dependent material properties, frictional heat generation, and plastic deformation behavior to predict stress, strain, and temperature distributions. Model predictions were validated against experimental sensor data, confirming accuracy in peak temperature and torque estimation. The integrated ML-DT system functioned as a decision-support tool, enabling real-time process monitoring, virtual experimentation, and predictive defect detection. When implemented in an industrial environment, the system dynamically adapted welding parameters to maintain optimal conditions. This approach enhances process stability, reduces material waste, and improves weld integrity, offering a scalable solution for intelligent manufacturing and Industry 4.0 applications.

High-speed friction stir welding (HSFSW) has emerged as a promising technique for improving the manufacturing efficiency of aluminum alloy structures by enabling faster welding while maintaining the quality of welded joints. This study investigates the mechanical properties and microstructural characteristics of AA 7020-T651 aluminum alloy joints welded using a novel multi-pin tool at high feed rates ranging from 2500 to 6000 mm/min under a constant rotational speed of 4000 rpm. Defect-free welds were successfully fabricated, as confirmed by metallographic analysis and micro-computed tomography (µ-CT). The multi-pin tool facilitated consistent material flow and heat distribution, which contributed to reliable joint formation across all feed rates. At the highest feed rate, the tensile strength reached 76% of the base material. A consistent softening in the nugget zone (NZ) was observed, and electron backscatter diffraction (EBSD) analysis showed a more than 70% grain size reduction in this zone, averaging ~3 µm, due to dynamic recrystallization. These findings underscore the suitability of HSFSW with multi-pin tools for high-speed industrial applications, offering enhanced productivity without compromising structural integrity.

In this study, the electrical resistivity measurements as an accessible technique for NDT inspection of FSW_ed sections were investigated. For this purpose, 1080 Aluminum plates were welded by the FSW technique. The welding parameters were optimized using the Taguchi L₉ orthogonal design of experiments. The rotational speed, tool shape, tilt angle, and welding speed were taken into consideration as welding parameters. The soundness of welds was inspected by visual tests, radiography, and metallography observations. The electrical resistivity of welded sections was compared with the area of bulk defects. It revealed that there is a linear relationship between the electrical resistivity and the area of bulk defects. The acquired intercept of the linear equation was introduced as a reference point. The results showed that the area of bulk defects could be predicted by $d=0.0001R-C$ an equation with good accuracy, and the joints with lower electrical resistivity than the reference point could be considered as a sound weld.

This paper aims to investigate the effects of different pin shapes on the corrosion behaviour of the friction stir welded (FSW_ed) joints. Friction stir welding (FSW) of 1080 aluminium alloy was done with the Taguchi L₉ orthogonal design using three-pin shapes including cylindrical, square, and triangle. The corrosion behaviour of welded samples was studied using the weight loss and potentiodynamic polarisation methods. Results revealed that the corrosion resistance of samples welded by cylindrical pin shape is higher than that of the samples welded by square and triangle pin shapes. Optical microscopy and scanning electron microscopy (SEM) were employed to investigate the corroded surface. The observations showed that the samples welded by cylindrical pin have better surface quality and less corrosion features than the samples welded by square and triangle pins in 3.5 wt.% NaCl solution. Furthermore, the microstructure analysis approved that decrease in grain size leads to higher corrosion.

Friction stir welding (FSW) of high-strength aluminum alloys, including EN AW 7020-T651, encounters significant challenges under weld line gap conditions, leading to compromised joint integrity. This study develops a predictive, data-driven framework to assess and optimize the gap bridgeability performance of FSW joints with weld line gaps ranging from 0 to 4 mm in 2 mm thick plates. A structured experimental matrix was implemented, systematically varying rotational speed, welding speed, axial force, and tool shoulder diameter. To promote stable material flow and consistent weld quality under varying gap conditions, a multi-pin tool was employed throughout the welding trials. This configuration supported defect-free weld formation across a broad process window and contributed to improved weld soundness under gap conditions. Weld quality was evaluated using a comprehensive, multi-criteria approach that required (i) defect-free joints verified by visual and cross-sectional (metallographic) inspection, (ii) an ultimate tensile strength of at least 230 MPa, and (iii) a novel metric termed weak area percentage (WAP). Derived from micro-hardness mapping, WAP quantified the proportion of the heat-affected zone (HAZ) exhibiting hardness below 96 HV, providing a more robust and spatially sensitive measure of mechanical integrity than conventional average hardness values. Two machine learning models, Logistic Regression and Random Forest, were trained to classify weld acceptability. The Random Forest model demonstrated superior performance, achieving 92.5% classification accuracy and an F1-score of 0.90. Feature importance analysis identified the interaction terms “welding speed × gap size” and “rotational speed × gap size” as the most influential predictors of weld quality.