Machine learning approach toward near-homogeneous properties of eutectic Aluminum silicon alloy fabricated by a wire arc direct energy deposition

Machine learning approach toward near-homogeneous properties of eutectic Aluminum silicon alloy fabricated by a wire arc direct energy deposition | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Machine learning approach toward near-homogeneous properties of eutectic Aluminum silicon alloy fabricated by a wire arc direct energy deposition M Hemachandra, Ramesh Mamedipaka, Shivraman Thapliyal This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7655300/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 9 You are reading this latest preprint version Abstract This study presents a parameter-based machine learning approach to predict optimal bead geometries in wire arc direct energy deposition (Wa-DED), aiming to reduce the time-consuming and costly trial-and-error procedures typically employed during process development. Decision trees, random forests, and K-nearest neighbors (KNN) models were trained and validated, with all three achieving comparable accuracy. Notably, the random forest model demonstrated superior performance in terms of accuracy, F1 score, and area under the curve (AUC). To further validate the optimized parameters, multilayer thin-walled Al-4047 structures were fabricated and comprehensively evaluated for both microstructural and mechanical properties. Microstructural analysis revealed α-Al dendrites with short, rounded Si in the interlayer and fine, fibrous Si in the melting zone, while EBSD confirmed predominantly equiaxed dendritic grains without a dominant crystallographic orientation, highlighting strong heterogeneity during solidification. Notably, sample-2 exhibited refined grains (92 ± 59 µm) and the highest average misorientation angle (10.14°) owing to its lower heat input. Meanwhile Sample 3 fabricated at equivalent heat input but with higher current amplitude, exhibited elevated KAM values and the highest dislocation density (~ 2.3 x 10¹⁴ m⁻²) due to intensified thermal gradients and cyclic thermal strains. These microstructural features strongly correlated with mechanical behavior, as Sample 2 with its finer grains, higher misorientation, and reduced porosity, achieved superior tensile strength (187.56 ± 11.40 MPa) and ductility (8.32 ± 0.66%), thereby demonstrating the efficacy of combining machine learning optimization with microstructural validation in tailoring Wa-DED components. Wire arc direct energy deposition (Wa-DED) Machine learning Random Forest Heat input Characterization Microstructure Mechanical properties Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 Figure 11 Figure 12 1 Introduction The wire arc direct energy deposition (Wa-DED) process has emerged as a promising alternative to traditional subtractive manufacturing methods for producing large aerospace components. With its low buy-to-fly ratios, Wa-DED offers several advantages, including environmental friendliness, cost-effectiveness, and minimal material wastage [ 1 – 3 ]. This process utilizes beam and arc-based heat sources, where high-power beams offer better forming quality but at the expense of deposition rates and part size. Conversely, the electric arc provides an excellent alternative with enhanced deposition rates, energy efficiency, cost-effectiveness, and safety benefits. [ 2 , 4 , 5 ]. Wa-DED, specifically wire and electric arc-assisted AM (WAAM), has demonstrated its potential in fabricating large-scale structures using various materials such as steel, titanium, nickel, and aluminum. [ 6 , 7 ]. WAAM employs gas metal welding (GMAW), gas tungsten welding (GTAW), and plasma arc welding (PAW) as heat sources. Developing components through Wa-DED involves the optimization of single-track profiles, developing test coupons for optimized bead profiles, and optimizing path planning for fabricating complex geometries. Traditionally, the selection and optimization in the initial steps have relied on trial and error or statistical methods. Various modeling techniques, including Taguchi methods, have been proposed to analyze single-track geometries. [ 3 , 8 ], response surface models [ 9 – 12 ], and regression models [ 13 ]. However, these approaches often have limited process optimization capabilities. In additive manufacturing, the model should be viewed in reverse, as geometric modifications are necessary to account for process-induced variations. Moreover, statistical optimization techniques tend to be time-consuming, expensive, and fail to explain the complex process-microstructure-property relationship in additive manufacturing [ 14 , 15 ]. Therefore, there is a need to integrate machine learning, metallurgy, and mechanistic models to enable effective part design, process planning, fabrication, characterization, and testing of manufactured parts. [ 16 ]. Data-driven machine learning (ML) approaches have gained popularity in additive manufacturing and other production techniques. Previous studies have demonstrated the application of ML techniques for predicting mechanical properties [ 17 ], establishing process-grain structure linkages [ 18 ], studying in-situ melt pool morphology [ 19 ], reducing porosity [ 15 ], and assessing the mechanical properties of joints [ 20 ]. However, most machine learning models in the literature focus on process optimization for powder bed fusion, with limited research reported for Wa-DED. Hence, this study aims to fill this gap by proposing a machine-learning model based on process parameters to predict optimum bead profile geometries in Wa-DED. Implementing this model is expected to reduce the number of steps involved in part building, reducing material waste and overall lead time. To validate the effectiveness of the machine learning model, real-time experimentation is conducted, and the predicted process parameters are used to fabricate test coupons. These coupons are then subjected to microstructural and mechanical characterization, further validating the proposed approach. The work presents a machine learning model to predict the track profile Wa-DED. Subsequently, the model was experimentally validated by comparing the model prediction with the experimental measurements. Additionally, the printing parameters were selected from the validation data for the printing of the multilayer wall printing, and to understand the influence of heat input on the mechanical properties of the aluminum silicon alloys. By integrating machine learning with real-time experimentation and characterization, this work aims to enhance the efficiency and effectiveness of the Wa-DED process, ultimately contributing to the advancement of large-scale component fabrication in aerospace and other industries. 2 Materials and Methods 2.1 Materials Bead-on-plate studies were performed to develop a machine-learning model with ER4047 wire. The composition of the wire is outlined in Table 1 . The weld tracks are laid on mechanically finished and cleaned AA6082 cold rolled + T6 substrate (300 mm × 50 mm × 8 mm) using a 6-axis OTC welding robot equipped with an OTC Daihen synchro feed welding machine (Fig. 1 ( a) ). The robot controller is programmed automatically with deposition instructions before the beginning of each deposit, including welding parameters and the movement of the welding torch. Table 1 Chemical composition of ER 4047 welding wire Material Mg Si Zn Fe Cu Mn Ti Al ER4047 < 0.1 11–13 < 0.2 < 0.6 < 0.3 < 0.15 < 0.15 Bal. 2.2 Data generation The data is generated with the in-house developed Wa-DED facility by depositing 117 single tracks of length 100 mm for different current and travel speeds. The deposition is performed in syncro-feed mode. The deposits are further cross-sectioned, and the weld bead form factor [H/W] is measured at a minimum of three sites for each bead. 2.3 Machine learning modeling The machine learning predictive model uses current and travel speed as the input parameters and the H/W ratio as the output parameter. A machine learning classification approach was used in this work, and for classification, the H/W ratio > 0.8 is assigned as one, and the ratio ≤ 0.8 is given as zero. The classifier model was developed using, i.e., logistic (LR), decision tree (DT), support vector regressor (SVR), and random forest classification models. The model's performance was evaluated based on accuracy, precision, recall, F1-score, and ROC-AUC curve. All models were cross-validated using grid search cross-validation. The best model was also selected based on the performance matrix and used for further work. 2.4 Validating the machine learning model The best-selected model was validated experimentally by depositing a single track for the process parameters selected from the available literature on Wa-DED of eutectic Al-Si alloys [ 21 – 23 ]. It is worth mentioning that the selected parameters were different from the dataset used to develop the machine learning model. 2.5 Test coupon printing and characterization After the validation of the ML model, test coupons were printed for the selected process parameters (Table 2 ). Porosity analysis of the test coupons was conducted using Image J software analysis, considering a total of 60 images, with 20 images captured at the thin walls' top, middle, and bottom sections. The test coupons were subjected to microstructural and mechanical characterization, and the mechanical properties were compared with the previously reported properties. Table 2 Process Parameters for Thin Walls Wall No. Current (I) Voltage (V) TS (cm/min) WFS (cm/min) Heat input (KJ/cm) 1 113 15.63 42 665 2.019 2 86 14.76 36 530 1.693 3 137 15.39 66 757 1.534 3 Results and discussions 3.1 Machine learning model 3.1.1 Performance of the machine learning model The decision tree and random forest model exhibited an accuracy of 83.75%, whereas the KNN model's accuracy was 81.25% (Fig. 2 ). A marginal variation in accuracy was observed for all three models. The confusion matrix and receiver operating characteristic curve (ROC) were plotted to understand the prediction ability of different machine learning models (Figs. 3 and 4 ). Figure 3 exhibited that the KNN model predicts maximum false negatives (5) cases, which means the actual H/W ratio is greater than 0.8 but predicted as ≤ 0.8. However, the decision tree predicted the minimum false negatives. On the contrary, the decision tree and KNN predicted the same false positive, i.e., indicated as > 0.8, whereas the values are below 0.8. The performance criteria of models suggested that the random forest model exhibited the highest F1 score compared to other models. Hence, based on accuracy and F1 score, random forest models outperform both models (Table 3 ). Table 3 Performance matrix for different models S.No Performance criteria (In percentage) KNN DT RF 1 Precision \(\:\frac{\text{T}\text{P}}{(\text{T}\text{P}+\text{F}\text{P})}\times\:100\) 88 88 92 2 Sensitivity \(\:\frac{\text{T}\text{P}}{(\text{T}\text{P}+\text{F}\text{N})}\) ×100 81.5 95.6 92 3 F1score \(\:\:\:\frac{\text{P}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}\times\:\text{S}\text{e}\text{n}\text{s}\text{i}\text{t}\text{i}\text{v}\text{i}\text{t}\text{y}\:}{(\text{P}\text{r}\text{e}\text{c}\text{i}\text{s}\text{i}\text{o}\text{n}+\text{S}\text{e}\text{n}\text{s}\text{i}\text{t}\text{i}\text{v}\text{i}\text{t}\text{y}\:)}\times\:2\times\:100\) 71 84.1 84.64 The area can confirm the better performance of the random forest model under the curve (AUC) of the probabilistic curve (ROC) shows in Fig. 4 . The mean AUC for the random forest model was the highest, i.e., 0.95, compared to KNN and the decision tree model (Fig. 4 ). Hence, it can be concluded that, based on the accuracy, F1 score, and AUC, the random forest model outperforms KNN and decision tree classification models. 3.1.2 Validation of the machine learning model The validation was performed by depositing a single bead for a different set of process parameters, as shown in Table 4 . Comparing the experimental and model results, the model reasonably predicted the bead profile. Table 4 Comparison between experimental and prediction values S.No Process parameters Experimental Criteria value KNN DT RF Current (I) Speed (S) 1 113 42 0.86 0 0 0 0 2 86 36 0.98 1 1 1 1 3 137 66 0.76 0 0 0 0 3.2 Test coupon printing and characterization. 3.2.1 Microstructural characterization Three Al-Si thin walls are fabricated for the parameters outlined in Table 2 , and the transverse sectional views of these walls are presented in Fig. 1 (d) This study has been structured into two parts. The first condition emphasizes the significance of heat input (between samples 1 and 2). The second situation reflects equivalent heat input (between samples 2 and 3) since there is only a marginal difference in heat input between samples 2 and 3. Thus, wherever these effects are prominent, they will be called C1 and C2, respectively. The cross-sectional view of thin walls (Fig. 1 (d) The small bulge can be observed due to slight deviations in the torch path during printing. Therefore, for the better-forming quality of Wa-DED structures, the torch must follow the path of the previously deposited layer laid down. Pores of enormous size were found in the top layer of the thin wall, with most pores being spherical. This is because the heat dissipation rate decreases as the deposition height increases, giving the pores additional time to aggregate and float to the surface [ 24 , 25 ]. Sample 2 displays fewer pores (Fig. 5 ) due to its lower heat input than sample 1, representing the effect of heat input (C1). Higher heat inputs can lead to gas entrapment in the flow due to intensified molten liquid dynamics, i.e., Marangoni flow. Higher heat inputs can also render alloying elements more vulnerable to evaporation, forming pores [ 6 , 26 , 27 ]. It is essential to highlight that samples 3 and 2 demonstrate similar porosity levels because of their comparable heat inputs (C2). The microstructural images of all samples indicate the presence of columnar α-Al dendrites and eutectic Si filling in the inter-dendritic spaces (Fig. 6 and Fig. 7 (a). Nevertheless, due to continuous remelting and reheating of the previously deposited layers, the typical microstructure characteristics, such as size and shape of dendrites, interlayer thickness, eutectic content, and directional growth patterns, are continuously changing along the deposition direction (Fig. 6 ). The high heat input in sample 1 results in a prolonged solidification process, increasing the size of dendrites (Fig. 6 (c)). Figure 6 (a) illustrates that the dendrites above the interlayer are oriented toward the center of the melt pool, which corresponds to the radial heat flow direction. The coarse and long columnar dendrites are more noticeable at the center of the molten pool than the dendrites at the fusion line (Fig. 7 (a,b)). Additionally, directionality was weakened at the center (fusion zone) due to the reduced temperature gradient compared to the fusion boundary (Fig. 6 (b)). In the interlayer, coarse α-Al dendrites with short and round Si were observed, whereas the melting zone exhibits fine and long dendrites (Fig. 6 (a) and Fig. 7 (a)). A porous structure can also be seen near the interlayer, resulting from the entrapment of oxides from the previously deposited layer within the molten pool. The dark regions in Fig. 7 (b) Those that Al essentially contains are called columnar α-Al, which starts to grow as soon as liquid metals solidify until it reaches the eutectic composition. At later stages (below eutectic temperature), the liquid in the two-phase mixture rich in Si will form Al-Si eutectics in the interdendritic regions. This explanation is well supported by the elemental distribution maps in Fig. 7 (b). 3.2.2 Electron backscatter diffraction (EBSD) analysis Electron backscatter diffraction (EBSD) investigations were conducted on Wa-DED fabricated Al-Si alloy thin-wall structures deposited using different process parameters to analyses grain morphology, crystallographic texture, and misorientation behavior. The EBSD analyses were performed on samples extracted from the mid-height section of the deposited structures. Figure 8 (a, b, c) presents the inverse pole Fig. (IPF) maps corresponding to different processing conditions, showing predominantly equiaxed dendritic grain structures in the middle region. Although previous studies have reported the presence of columnar dendritic grain structures near the substrate region, aligned along the direction of the maximum thermal gradient [ 28 ]. The bottom region acts as the primary nucleation zone, where the substrate interface governs grain growth by epitaxial solidification. Dendritic growth generally proceeds along preferred crystallographic directions aligned with the thermal gradient, typically in face-centered cubic (FCC) structures such as Al alloys, due to the lowest resistance to growth in this orientation. [ 29 ]. As the build height increases, the thermal gradient (G) and solidification velocity (R) at the solid–liquid interface decrease. This reduction promotes solute buildup in the melt, leading to the formation of a constitutionally supercooled zone. Consequently, the grain morphology transitions from columnar to equiaxed dendrites at the middle and upper regions of the wall. This transformation is facilitated by heterogeneous nucleation at the center of the molten pool and the suppression of directional growth due to lateral thermal fluctuations. When the number and size of equiaxed dendrites in the melt reach a critical level, the advancing columnar front is blocked, promoting a fully equiaxed grain structure, as shown in Fig. 8 (a, b, c). As a result, no dominant crystallographic orientation was observed. Moreover, the thermal exposure from subsequent layer deposition further alters the crystallographic orientation and influences grain growth. The average grain sizes measured from EBSD data for samples 1, 2, and 3 were approximately 104 ± 32 µm, 92 ± 18 µm, and 90 ± 33 µm, respectively, as shown in Fig. 8 (d, e, f). The higher heat input in Sample 1 led to coarser grains, attributed to the reduced cooling rate and extended time for grain growth, which favored the formation of coarser equiaxed dendrites comparatively. In contrast, Samples 2 and 3, which experienced similar equivalent heat input levels, exhibited comparable grain sizes. The formation of finer grains at lower heat inputs is associated with a higher cooling rate and limited constitutional supercooling, which restricts grain growth. In addition, the investigation assessed the distribution of dislocations in the as-deposited Al-Si thin-wall structures fabricated under different process parameters. The Kernel Average Misorientation (KAM) method was employed using EBSD orientation data to quantify local misorientation. KAM maps (Fig. 9 (a, b, c) reveal localized misorientation gradients within grains, indicative of geometrically necessary dislocations (GNDs) that accommodate plastic deformation and thermal strain induced by cyclic heating and rapid solidification inherent to Wa-DED. This approach enables the estimation of dislocation density using the expression: \(\:{\rho\:}_{GND}\cong\:\frac{{2\theta\:}_{KAM}}{bd}\) Eq. (1) Here \(\:{\theta\:}_{KAM}\:\) is the average KAM value, b is the Burger vector magnitude, and d is the step size. KAM assigns a misorientation value to each point by averaging the angular differences between that point and its immediate neighbours. The calculated geometrically necessary dislocation (GND) densities for different parameters are presented in Table 5 , and are found to be on the order of 10 ¹4 m⁻². Although Samples 2 and 3 were fabricated with the same nominal heat input, their dislocation densities differ markedly due to differences in applied current and associated thermal profiles. Sample 3, processed at a higher current, exhibited the highest dislocation density. The elevated current intensified thermal gradients, increasing localized plastic strain and thermal mismatch stresses during rapid solidification. These conditions promoted extensive dislocation generation, which outweighs the effects of enhanced dynamic recovery associated with its lower misorientation. Sample 1 showed a moderate dislocation density, reflecting a balance between dislocation generation from thermal strains and recovery due to prolonged high-temperature exposure. In contrast, Sample 2 displayed the lowest dislocation density, consistent with its lower peak thermal gradients and reduced accumulation of cyclic thermal strain. Table 5 The calculated Θ KAM and GND density based on KAM maps for Wa-DED-fabricated Al-Si alloy under different process parameters Specimen Θ KAM Dislocation density (m − 2 ) Sample-1 0.86° 2.1 × 10 14 Sample-2 0.68° 1.7 × 10 14 Sample-3 0.98° 2.3 × 10 14 The grain boundary misorientation angle distribution in all three samples (Fig. 10 (a, b, c) demonstrates a predominance of low-angle grain boundaries (LAGBs, < 15° misorientation). The average misorientation angles were measured as approximately 8.14°, 10.14°, and 6.89° for Samples 1, 2, and 3, respectively. Sample 2 exhibited the highest average misorientation despite its finer grain size, which is attributed to rapid cooling and limited dynamic recovery. These conditions preserve a greater fraction of sub-grain boundaries without significant boundary migration. In contrast, Sample 3, produced with an equivalent overall heat input to Sample 2 but under a higher applied current, exhibited a lower average misorientation. The higher applied current likely increased the peak temperatures during deposition, enhancing dynamic recovery and promoting grain boundary migration. The higher heat input sample 1 depicts the intermediate average misorientation, reflecting the partial recovery facilitated by prolonged thermal exposure. The extended high-temperature conditions in Sample 1 allowed adjacent low-angle boundaries to migrate and merge, forming larger grains (as shown in Fig. 8 (a) without necessarily increasing their mutual misorientation. This process resulted in overall grain coarsening rather than a proliferation of high-angle grain boundaries (HAGBs). These observations indicate that reduced thermal energy promotes higher sub-grain misorientation due to rapid cooling, limited recovery, and more pronounced constitutional supercooling conditions that enhance sub-grain formation while suppressing dislocation annihilation. 3.2.3 Mechanical characterization Figure 11 describes the microhardness variation along the building direction of three fabricated samples subjected to different heat inputs. Forty hardness measurements were taken at intervals of 1 mm across the mid-width of the wall from the bottom to the top layer at 100 grams with a dwell time of 15 seconds. There is no noticeable difference in the average Vickers hardness values of the samples (Table 6 ). Nevertheless, the decline in heat input is associated with a small increment in hardness, owing to rapid cooling rates and a consequent increase in dendritic arm spacing [ 27 , 30 ]. Samples 1 and 3 display a wider variability in hardness values from the bottom to the top of the thin wall compared to sample 2, with no evidence of an increasing or decreasing trend. The hardness fluctuations observed in samples 1 and 3 can be attributed to their high currents, porosity content, and interaction with the indenter. The mechanical properties obtained from tensile testing are shown in Fig. 12 and Table 6 . Sample 2 exhibits overall better properties with UTS of 187.56 ± 11.40 and ductility of 8.32 ± 0.66, owing to reduced porosity and fine dendritic arm structure (Fig. 8 (a)). This improvement in sample 2 can be attributed to two key factors. The first is the rapid solidification process, which facilitates the formation of delicate dendritic structures at low heat inputs (C1). A second benefit of adopting low wire feed speeds (accompanied by lower currents and voltages) and travel speeds facilitate the development of delicate dendritic structures due to lower temperature gradients and faster cooling rates in the equivalent heat input condition (C2) [ 22 ]. Also, from the stress-strain plots, it is evident that sample 2 exhibits less variation among the replicate specimens than the other walls. Furthermore, the superior mechanical properties of sample-2 are attributed to its optimal combination of fine grain size (92 ± 59 µm) and the highest average grain boundary misorientation angle (10.14°), resulting in a higher fraction of HAGBs that effectively impede the dislocation motion [ 31 ]. In a similar vein, Sample 3, which exhibits a grain size comparable to that of Sample 2, demonstrates noticeably higher mechanical properties owing to its substantially higher KAM value (0.98°), despite the predominance of LAGBs as indicated by its lower average misorientation angle (6.89°). Additionally, the higher heat input, subsequent cooling rate, and associated grain refinement in Sample-3 may contribute to its higher strength, as reflected in its average tensile strength in the vertical direction UTS V, of 180.73 ± 8.01 (usually lower than the avg. UTS H ). The present results are consistent with the past study and comparable with the as-cast structure (Table 7 ). The higher heat input in wall-1 results in comparatively lower tensile strength, which is attributed to its coarse-grain structure, intermediate misorientation (8.14°), and moderate KAM (0.86). Combining coarser grains and fewer HAGBs underpins its lowest mechanical properties among the three samples. Therefore, these findings indicate that lower heat input in variable heat input conditions (C1), low wire feed speed, and travel speed during equivalent heat input conditions (C2) are the most favorable for achieving superior mechanical properties. Overall, the mechanical properties correlate predominantly with grain size rather than dislocation content (KAM values), and to a lesser extent with LAGB and HAGB fractions. This highlights the dominant strengthening role of grain boundary character over internal strain hardening in the additively manufactured Al 4047 thin walls analyzed. Table 6 The Avg mechanical properties of 3 samples along horizontal and vertical directions Wall No. Sample direction Avg. UTS (MPa) % EL (%) Hardness (HV 0.1 ) 1 H 177.68 ± 12.35 7.71 ± 0.49 74.13 ± 0.9 V 146.79 ± 12.55 5.41 ± 0.17 2 H 187.56 ± 11.40 8.32 ± 0.66 73.39 ± 0.37 V 165.38 ± 7.34 6.19 ± 0.96 3 H 170.48 ± 10.65 9.83 ± 0.55 73.32 ± 0.69 V 180.73 ± 8.01 6.18 ± 0.96 Table 7 Summary of current and previous studies on the mechanical properties of Al-Si alloys. Reference Power Source/Manufacturing Route Alloy Tensile strength (MPa) % Elongation [ 32 ] Welding Wire ER4047 130 5 [ 33 ] As cast Al-12Si 211.62 2 Present Study (S1) Wa-DED (MIG) ER4047 162.23 6.56 Present Study (S2) Wa-DED (MIG) ER4047 176.47 7.25 Present Study (S2) Wa-DED (MIG) ER4047 175.60 8.0 [ 34 ] Wa-DED (GMAW) ER4047 180.46 15.1 [ 23 ] GMAW AA4047 226 14 4 Conclusion In this study, a parameter-based machine learning model was developed and validated to predict the optimal bead profile in Wa-DED of Al-Si alloys. Real-time experiments confirmed the reliability of the model, with the random forest algorithm exhibiting superior classification performance compared to decision tree and KNN models, owing to its higher F1 score and AUC. Microstructural characterization revealed that lower heat input conditions (Sample 2) minimized porosity and favored finer dendritic structures, whereas higher heat input (Sample 1) resulted in coarser grains due to prolonged solidification. EBSD analysis confirmed the predominance of equiaxed dendritic morphologies without a preferred crystallographic orientation, with grain size being strongly dependent on heat input. Misorientation and dislocation characteristics were also highly sensitive to process conditions, where Sample 2 exhibited the highest average misorientation as a result of rapid cooling and limited recovery. While Sample 3, processed with higher current at equivalent heat input, developed the highest dislocation density due to intensified thermal gradients and plastic strain. These microstructural variations directly influenced mechanical performance, with Sample 2 achieving the best combination of strength (187.56 ± 11.40 MPa) and ductility (8.32 ± 0.66) due to its fine dendritic structure and reduced porosity. Notably, superior properties were not governed by overall heat input alone; rather, the correlation of individual parameters such as wire feed speed, current and travel speed also played a decisive role, as demonstrated by the improved performance of Sample 3 under equivalent heat input but altered parameter combinations. Overall, this study establishes that both heat input magnitude and the interplay of individual process parameters must be considered to optimize bead geometry, minimize porosity, refine microstructure, and enhance the mechanical properties of additively manufactured Al-Si thin-wall components. Declarations Funding We acknowledge the support from the Aeronautics Research and Development Board of the Government of India under Grant No. ARDB/01/2031958/M/I. Conflicts of Interests There are no conflicts of interest to declare by the authors. Data Availability Not applicable Code Availability Not applicable Consent to Participate All the authors have given consent to participate. Ethics Approval The authors hereby state that the present work complies with the ethical standards. Author Contribution Shivraman provided the experimental equipment; Hemachandra and Ramesh prepared the figure and wrote the manuscript; Shivraman reviewed and edited the manuscript; Shivraman, Hemachandra and Ramesh reviewed the manuscript. References Ding D, Pan Z, Cuiuri D, Li H (2015) A multi-bead overlapping model for robotic wire and arc additive manufacturing (WAAM). Robot Comput Integr Manuf 31:101–110. https://doi.org/10.1016/j.rcim.2014.08.008 Campatelli G, Campanella D, Barcellona A, et al (2020) Microstructural, mechanical and energy demand characterization of alternative WAAM techniques for Al-alloy parts production. CIRP J Manuf Sci Technol 31:492–499 Dinovitzer M, Chen X, Laliberte J, et al (2019) Effect of wire and arc additive manufacturing (WAAM) process parameters on bead geometry and microstructure. Addit Manuf 26:138–146. https://doi.org/10.1016/j.addma.2018.12.013 Yang D, He C, Zhang G (2016) Forming characteristics of thin-wall steel parts by double electrode GMAW based additive manufacturing. J Mater Process Technol 227:153–160. https://doi.org/10.1016/j.jmatprotec.2015.08.021 Hemachandra M, Thapliyal S, Adepu K (2022) A review on microstructural and tribological performance of additively manufactured parts. J Mater Sci 57:17139–17161. https://doi.org/10.1007/s10853-022-07736-1 Thapliyal S (2019) Challenges associated with the wire arc additive manufacturing (WAAM) of aluminum alloys. Mater Res Express 6:. https://doi.org/10.1088/2053-1591/ab4dd4 Tomar B, Shiva S, Nath T (2022) A review on wire arc additive manufacturing: Processing parameters, defects, quality improvement and recent advances. Mater Today Commun 31:103739 Zavdoveev A, Pozniakov V, Baudin T, et al (2022) Optimization of the pulsed arc welding parameters for wire arc additive manufacturing in austenitic steel applications. Int J Adv Manuf Technol 119:5175–5193 Kumar A, Maji K (2020) Selection of Process Parameters for Near-Net Shape Deposition in Wire Arc Additive Manufacturing by Genetic Algorithm. J Mater Eng Perform 29:3334–3352. https://doi.org/10.1007/s11665-020-04847-1 Mai DS, Doan TK, Paris H, others (2021) Wire and arc additive manufacturing of 308L stainless steel components: Optimization of processing parameters and material properties. Eng Sci Technol an Int J 24:1015–1026 Geng H, Xiong J, Huang D, et al (2017) A prediction model of layer geometrical size in wire and arc additive manufacture using response surface methodology. Int J Adv Manuf Technol 93:175–186. https://doi.org/10.1007/s00170-015-8147-2 Mamedipaka R, Hemachandra M, Thapliyal S (2024) Microstructural and mechanical characterization of direct energy deposited 316L stainless steel developed using an evolutionary computational approach. Proc Inst Mech Eng Part L J Mater Des Appl. https://doi.org/10.1177/14644207241262107 Zhao Y, Li F, Chen S, Lu Z (2019) Unit block--based process planning strategy of WAAM for complex shell--shaped component. Int J Adv Manuf Technol 104:3915–3927 Du Y, Mukherjee T, DebRoy T (2021) Physics-informed machine learning and mechanistic modeling of additive manufacturing to reduce defects. Appl Mater Today 24:. https://doi.org/10.1016/j.apmt.2021.101123 Raju KL, Thapliyal S, Sigatapu S, et al (2022) Process Parameter Dependent Machine Learning Model for Densification Prediction of Selective Laser Melted Al-50Si Alloy and its Validation. J Mater Eng Perform 0–7. https://doi.org/10.1007/s11665-022-06831-3 DebRoy T, Mukherjee T, Wei HL, et al (2021) Metallurgy, mechanistic models and machine learning in metal printing. Nat Rev Mater 6:48–68. https://doi.org/10.1038/s41578-020-00236-1 Xie X, Bennett J, Saha S, et al (2021) Mechanistic data-driven prediction of as-built mechanical properties in metal additive manufacturing. npj Comput Mater 7:86 Popova E, Rodgers TM, Gong X, et al (2017) Process-structure linkages using a data science approach: application to simulated additive manufacturing data. Integr Mater Manuf Innov 6:54–68 Scime L, Beuth J (2019) Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process. Addit Manuf 25:151–165 Thapliyal S, Mishra A (2021) Machine learning classification-based approach for mechanical properties of friction stir welding of copper. Manuf Lett 29:52–55. https://doi.org/10.1016/j.mfglet.2021.05.010 Guo Y, Han Q, Hu J, et al (2022) Comparative study on wire-arc additive manufacturing and conventional casting of Al--Si alloys: porosity, microstructure and mechanical property. Acta Metall Sin (English Lett 35:475–485 Chen C, Sun G, Du W, et al (2023) Effect of equivalent heat input on WAAM Al-Si alloy. Int J Mech Sci 238:107831 Langelandsvik G, Horgar A, Furu T, et al (2020) Comparative study of eutectic Al-Si alloys manufactured by WAAM and casting. Int J Adv Manuf Technol 110:935–947 Pramod R, Kumar SM, Girinath B, et al (2020) Fabrication, characterisation, and finite element analysis of cold metal transfer--based wire and arc additive--manufactured aluminium alloy 4043 cylinder. Weld World 64:1905–1919 Gu J, Gao M, Yang S, et al (2019) Pore formation and evolution in wire+ arc additively manufactured 2319 Al alloy. Addit Manuf 30:100900 Utyaganova VR, Filippov A V, Shamarin NN, et al (2020) Controlling the porosity using exponential decay heat input regimes during electron beam wire-feed additive manufacturing of Al-Mg alloy. Int J Adv Manuf Technol 108:2823–2838 Rosli NA, Alkahari MR, bin Abdollah MF, et al (2021) Review on effect of heat input for wire arc additive manufacturing process. J Mater Res Technol 11:2127–2145. https://doi.org/10.1016/j.jmrt.2021.02.002 Dinda GP, Dasgupta AK, Bhattacharya S, et al (2013) Microstructural characterization of laser-deposited Al 4047 alloy. Metall Mater Trans A Phys Metall Mater Sci 44:2233–2242. https://doi.org/10.1007/s11661-012-1560-3 Chu Q, Bai R, Jian H, et al (2018) Microstructure, texture and mechanical properties of 6061 aluminum laser beam welded joints. Mater Charact 137:269–276. https://doi.org/10.1016/j.matchar.2018.01.030 Vandersluis E, Ravindran C (2017) Comparison of measurement methods for secondary dendrite arm spacing. Metallogr Microstruct Anal 6:89–94 Najafi S, Eivani AR, Samaee M, et al (2018) A comprehensive investigation of the strengthening effects of dislocations, texture and low and high angle grain boundaries in ultrafine grained AA6063 aluminum alloy. Mater Charact 136:60–68. https://doi.org/https://doi.org/10.1016/j.matchar.2017.12.004 Heard DW, Brophy S, Brochu M (2012) Solid freeform fabrication of Al--Si components via the CSC-MIG process. Can Metall Q 51:302–312 Dwivedi DK, Sharma R, Kumar A (2006) Influence of silicon content and heat treatment parameters on mechanical properties of cast Al--Si--Mg alloys. Int J Cast Met Res 19:275–282 Haselhuhn AS, Buhr MW, Wijnen B, et al (2016) Structure-property relationships of common aluminum weld alloys utilized as feedstock for GMAW-based 3-D metal printing. Mater Sci Eng A 673:511–523 Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 11 Jan, 2026 Reviews received at journal 05 Dec, 2025 Reviewers agreed at journal 26 Nov, 2025 Reviewers agreed at journal 25 Nov, 2025 Reviewers agreed at journal 25 Nov, 2025 Reviewers invited by journal 25 Nov, 2025 Editor assigned by journal 22 Sep, 2025 Submission checks completed at journal 22 Sep, 2025 First submitted to journal 19 Sep, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7655300","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":550963343,"identity":"1e8c5d8c-51ed-4eae-b047-d8d5625b761a","order_by":0,"name":"M Hemachandra","email":"","orcid":"","institution":"National Institute of Technology Warangal","correspondingAuthor":false,"prefix":"","firstName":"M","middleName":"","lastName":"Hemachandra","suffix":""},{"id":550963348,"identity":"624e0ca6-948a-4a05-965b-c05b57e74904","order_by":1,"name":"Ramesh Mamedipaka","email":"","orcid":"","institution":"National Institute of Technology Warangal","correspondingAuthor":false,"prefix":"","firstName":"Ramesh","middleName":"","lastName":"Mamedipaka","suffix":""},{"id":550963350,"identity":"412defe3-6903-412f-8a45-1dd83b06ec38","order_by":2,"name":"Shivraman Thapliyal","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA2UlEQVRIiWNgGAWjYDACCcYGCIO9geEAhAUTIaiF5wDDgQPEaYEzEhhg1uAH8rOb26QrarbJm0u+PXj4YxuDPH8Dc9sDfFoM7hxskzxz7Lbhztl5CQcOtjEYzjjA2G6AV4tEYrNhA9ttxg23cwwOHNzGwLiBgbFNAp8W+RkgLf9u22+4eQasxZ6gFoYbiY0PG9tuJ264wQPWkkhQiwFYS9/t5A1ngA47+08iecZhgg5Lf3Cw4dtt2w3Hzxh/qDhjY9vf3v4Mv8PQAFAxMynqR8EoGAWjYBRgBQBKUFN8Bs14mgAAAABJRU5ErkJggg==","orcid":"","institution":"National Institute of Technology Warangal","correspondingAuthor":true,"prefix":"","firstName":"Shivraman","middleName":"","lastName":"Thapliyal","suffix":""}],"badges":[],"createdAt":"2025-09-19 06:53:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-7655300/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7655300/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":97366592,"identity":"f46866a1-cf93-497b-880a-c05f9bc9eacc","added_by":"auto","created_at":"2025-12-03 15:41:09","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11505957,"visible":true,"origin":"","legend":"","description":"","filename":"Manuscript.docx","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/c497fa4a6b66fd751a035f96.docx"},{"id":97366599,"identity":"9eaecc42-c050-4d5e-8f78-855a1db77316","added_by":"auto","created_at":"2025-12-03 15:41:12","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5293,"visible":true,"origin":"","legend":"","description":"","filename":"f38807da844b4f22a567f7403826148e.json","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/136d43c2fa43bf87e0803177.json"},{"id":97248239,"identity":"5c82962b-6f49-45e5-9a26-63ddf7450f3b","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":108912,"visible":true,"origin":"","legend":"","description":"","filename":"f38807da844b4f22a567f7403826148e1enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/0b08c201ce21b3d056edbee4.xml"},{"id":97248242,"identity":"8445fcbf-b280-48af-a42d-2683b1530dc8","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1742704,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/ed52bb11933f75902916c471.jpeg"},{"id":97369306,"identity":"736b45a1-c1ca-40b1-8175-f57dcc89da80","added_by":"auto","created_at":"2025-12-03 16:24:14","extension":"jpeg","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1755702,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/bfe7714d45ca99f13eccb886.jpeg"},{"id":97248246,"identity":"e3518c29-23d5-4ad8-bb8b-17d5cc9f1b0d","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":235520,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage11.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/997803259402f98d8e1558e2.jpeg"},{"id":97248248,"identity":"4142da9d-02c2-4d1d-80ad-a877048f55d7","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"emf","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":96912,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage12.emf","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/a6fee6a6a716b367f417b04d.emf"},{"id":97366593,"identity":"c320dec6-5326-452d-b850-cdda96d1e521","added_by":"auto","created_at":"2025-12-03 15:41:09","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":459667,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/cc169e3b73292c9063c042be.jpeg"},{"id":97248253,"identity":"e56addfd-5939-4ca2-9f1c-28d2b2c3e564","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":92026,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/113d7700d435eec354454ecc.jpeg"},{"id":97370028,"identity":"06f3f833-3c2d-4064-acf9-e9ef458c1481","added_by":"auto","created_at":"2025-12-03 16:26:32","extension":"jpeg","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":366834,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/1f018ea56d89f4a9df496205.jpeg"},{"id":97248250,"identity":"327199e0-a2a1-4a14-b218-d4ff142a83f5","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"jpeg","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":175050,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/1f9508e57818cd78d2fba39d.jpeg"},{"id":97366600,"identity":"7d60c6f3-3564-46db-8321-8933b36e0568","added_by":"auto","created_at":"2025-12-03 15:41:13","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1398027,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/1d381964ec72f82756af2020.png"},{"id":97366597,"identity":"f969279f-7e4d-4b04-a289-9e737486dbe7","added_by":"auto","created_at":"2025-12-03 15:41:11","extension":"jpeg","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1302058,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/0992afd8be32957782723ea0.jpeg"},{"id":97664451,"identity":"37c6be19-3862-4d63-a5a3-2975d1d277aa","added_by":"auto","created_at":"2025-12-08 08:36:02","extension":"jpeg","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1408450,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/da9fb7590a2c73b7a24b613d.jpeg"},{"id":97248260,"identity":"63b8a5c6-5209-4875-a04d-27370c439e0c","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"jpeg","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1955528,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/33d13210c44b5b969c7f5281.jpeg"},{"id":97366595,"identity":"0b2f2bcb-bd8b-439d-a5b7-094f60e1fa91","added_by":"auto","created_at":"2025-12-03 15:41:10","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":129644,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/bebe634ea88157e5a9a5ca82.png"},{"id":97664439,"identity":"89953706-2276-4530-ba73-34faf0033f88","added_by":"auto","created_at":"2025-12-08 08:35:46","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":319648,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/58a9df4ecf932f256cb3a7a0.png"},{"id":97248265,"identity":"cb384a59-d26d-4dcc-bd0c-19cfc85eb812","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":183335,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage11.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/9745f1292ef12a48f8320d0e.png"},{"id":97248263,"identity":"7aafc434-537c-428e-af32-940fd34b9b72","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":15230,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage12.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/21d4b8576ec2d700cca865a9.png"},{"id":97248266,"identity":"26f182f1-b974-4d02-a1fd-017c9a63f8df","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48866,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/279f369ed57bc1490e53aa99.png"},{"id":97248258,"identity":"adb18813-6972-46de-b303-2accd815fc63","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":27753,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/b2f1f0cb2663e950c53d0e52.png"},{"id":97369525,"identity":"ef28510c-817b-4907-ab5f-f00a78b769cb","added_by":"auto","created_at":"2025-12-03 16:25:05","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59874,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/a4bf2acce0e1e29080b079f8.png"},{"id":97366601,"identity":"a6256b64-dbec-4b2b-8497-9fa96d955274","added_by":"auto","created_at":"2025-12-03 15:41:13","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":139671,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/b4bb4ec58f0e950f175d2688.png"},{"id":97664446,"identity":"76292710-f18f-4ed0-a384-0b77b319dcce","added_by":"auto","created_at":"2025-12-08 08:35:58","extension":"png","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":253824,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/4a369d6f388037d810abb82d.png"},{"id":97248261,"identity":"b8f22350-65b6-4e49-9b36-be3a69b4eb0a","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":169978,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/590a7e58a623efa233494b63.png"},{"id":97248267,"identity":"8e697c3d-ceee-42f6-bf32-c7a1e0c777c7","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":25,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":262469,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/52c278d69412397153b74eab.png"},{"id":97248270,"identity":"68cc3e89-47e3-4b4c-ad84-6db63baf946c","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"png","order_by":26,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":351974,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/4b8b561cc1f94e1ada371352.png"},{"id":97366596,"identity":"bea1e6d8-92c6-4bb3-b3af-27287f101bff","added_by":"auto","created_at":"2025-12-03 15:41:11","extension":"xml","order_by":27,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":107728,"visible":true,"origin":"","legend":"","description":"","filename":"f38807da844b4f22a567f7403826148e1structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/54d1b8d923c31ddd487dda0c.xml"},{"id":97248269,"identity":"ee72ae30-515d-4a9a-82e1-b154537bb420","added_by":"auto","created_at":"2025-12-02 12:41:57","extension":"html","order_by":28,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":117794,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/2d968900af90d3210eb1e24b.html"},{"id":97248234,"identity":"6074466c-1952-45ee-9f91-5d0a7ce575a0","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":1149316,"visible":true,"origin":"","legend":"\u003cp\u003eSchematic displaying (a) OTC synchro feed robotic Wa-DED setup, (b) Sample locations and other features, (c) Thin walls with different parameter sets, (d) Stereomicroscope images of transverse profiles of three thin wall deposits\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/f1509feb114dfdeb3c89eb77.png"},{"id":97248231,"identity":"ebc852e8-040f-43a5-96df-cfe7904cc9ec","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":993674,"visible":true,"origin":"","legend":"\u003cp\u003eBar chart showing the accuracy of the model\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/f03b1f29d0e9c94741e30e3b.png"},{"id":97248232,"identity":"7f2a9475-1619-4f07-b232-48bf260253fb","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":177796,"visible":true,"origin":"","legend":"\u003cp\u003eConfusion matrix for (a) KNN, (b) decision tree, and (c) random forest model\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/9834491bcb48137637ae1a26.png"},{"id":97248233,"identity":"00a26e58-c5e4-406f-8702-99a5ed6f4357","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":199601,"visible":true,"origin":"","legend":"\u003cp\u003eROC curve for (a) KNN, (b) Decision tree, and (c) Random Forest model\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/ac9fc1a6e0e48b87850a1510.png"},{"id":97664443,"identity":"cd011ac5-a6b2-416c-9020-55b06caaf346","added_by":"auto","created_at":"2025-12-08 08:35:48","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":661685,"visible":true,"origin":"","legend":"\u003cp\u003eImage displaying the effect of heat input on % porosity\u003c/p\u003e","description":"","filename":"5.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/077b919e1241fb4e1e34a432.png"},{"id":97248237,"identity":"53fb4f93-7f0b-4ecc-bfc1-20127e172e9f","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":1653488,"visible":true,"origin":"","legend":"\u003cp\u003eOptical images showing (a) Interlayer, directionality at the fusion boundary, and pores of Sample 2, (b) Reduction in the directional growth in the fusion zone of sample 2, (c) Coarse α-Al columnar dendrites in sample 1\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/fef6515ee1846c2a966c1f5c.png"},{"id":97366587,"identity":"89ebc0d3-166a-47a7-ae37-2097522b7d34","added_by":"auto","created_at":"2025-12-03 15:41:09","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":1279753,"visible":true,"origin":"","legend":"\u003cp\u003eSEM image of sample 2 displaying (a) Dendrites growth, fusion boundary, and size and shape of eutectic Si, (b) with EDS to show α-Al and eutectic Si and their distribution\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/35330d5b6a41049164bf4016.png"},{"id":97248236,"identity":"8a57e4d5-7802-403a-aebe-34030dfd1152","added_by":"auto","created_at":"2025-12-02 12:41:56","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":1240623,"visible":true,"origin":"","legend":"\u003cp\u003eInverse pole figure and grain size distribution chart of (a, d) sample-1, (b, e) sample-2, and (c, f) sample-3, respectively.\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/53938543df31c40bfca9dec3.png"},{"id":97369962,"identity":"d92a8924-dc6b-4005-aeae-0d1465726991","added_by":"auto","created_at":"2025-12-03 16:26:12","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":1805754,"visible":true,"origin":"","legend":"\u003cp\u003eKAM maps of Wa-DED -fabricated Al-Si alloy under different process parameters: (a) Sample 1, (b) Sample 2, and (c) Sample 3\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/4d7c19b4dc92fe61de454e7c.png"},{"id":97366603,"identity":"76c650c2-c1fb-49ee-b939-f6b1ae6d8e38","added_by":"auto","created_at":"2025-12-03 15:41:15","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":118477,"visible":true,"origin":"","legend":"\u003cp\u003eGrain boundary misorientation angle distributions of Wa-DED -fabricated Al-Si alloy under different process parameters: (a) Sample 1, (b) Sample 2, and (c) Sample 3\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/dfe41b006cb044c2dc5a3a65.png"},{"id":97366602,"identity":"9966966b-34b4-44db-9e78-dc15bfbf9b90","added_by":"auto","created_at":"2025-12-03 15:41:14","extension":"png","order_by":11,"title":"Figure 11","display":"","copyAsset":false,"role":"figure","size":934100,"visible":true,"origin":"","legend":"\u003cp\u003eHardness variation along the building direction of 3 samples\u003c/p\u003e","description":"","filename":"11.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/55b7c8230088b1b711c49602.png"},{"id":97366594,"identity":"8a691ce6-cb5e-495b-a99b-760d5d0b7594","added_by":"auto","created_at":"2025-12-03 15:41:09","extension":"png","order_by":12,"title":"Figure 12","display":"","copyAsset":false,"role":"figure","size":86107,"visible":true,"origin":"","legend":"\u003cp\u003eAverage tensile test results of 3 samples in horizontal (H) and vertical (V) directions at three different locations\u003c/p\u003e","description":"","filename":"12.png","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/449637c543433daaa1eabe6b.png"},{"id":97893340,"identity":"062f8c0f-6929-4daf-910c-03b24ac6390f","added_by":"auto","created_at":"2025-12-10 15:30:09","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":10662449,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7655300/v1/ce322d94-85db-4e02-8185-a4bedd2878fa.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Machine learning approach toward near-homogeneous properties of eutectic Aluminum silicon alloy fabricated by a wire arc direct energy deposition","fulltext":[{"header":"1 Introduction","content":"\u003cp\u003eThe wire arc direct energy deposition (Wa-DED) process has emerged as a promising alternative to traditional subtractive manufacturing methods for producing large aerospace components. With its low buy-to-fly ratios, Wa-DED offers several advantages, including environmental friendliness, cost-effectiveness, and minimal material wastage [\u003cspan additionalcitationids=\"CR2\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. This process utilizes beam and arc-based heat sources, where high-power beams offer better forming quality but at the expense of deposition rates and part size. Conversely, the electric arc provides an excellent alternative with enhanced deposition rates, energy efficiency, cost-effectiveness, and safety benefits. [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e, \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Wa-DED, specifically wire and electric arc-assisted AM (WAAM), has demonstrated its potential in fabricating large-scale structures using various materials such as steel, titanium, nickel, and aluminum. [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. WAAM employs gas metal welding (GMAW), gas tungsten welding (GTAW), and plasma arc welding (PAW) as heat sources.\u003c/p\u003e\u003cp\u003eDeveloping components through Wa-DED involves the optimization of single-track profiles, developing test coupons for optimized bead profiles, and optimizing path planning for fabricating complex geometries. Traditionally, the selection and optimization in the initial steps have relied on trial and error or statistical methods. Various modeling techniques, including Taguchi methods, have been proposed to analyze single-track geometries. [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e, \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], response surface models [\u003cspan additionalcitationids=\"CR10 CR11\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e], and regression models [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. However, these approaches often have limited process optimization capabilities. In additive manufacturing, the model should be viewed in reverse, as geometric modifications are necessary to account for process-induced variations. Moreover, statistical optimization techniques tend to be time-consuming, expensive, and fail to explain the complex process-microstructure-property relationship in additive manufacturing [\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e, \u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. Therefore, there is a need to integrate machine learning, metallurgy, and mechanistic models to enable effective part design, process planning, fabrication, characterization, and testing of manufactured parts. [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eData-driven machine learning (ML) approaches have gained popularity in additive manufacturing and other production techniques. Previous studies have demonstrated the application of ML techniques for predicting mechanical properties [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e], establishing process-grain structure linkages [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e], studying in-situ melt pool morphology [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e], reducing porosity [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e], and assessing the mechanical properties of joints [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. However, most machine learning models in the literature focus on process optimization for powder bed fusion, with limited research reported for Wa-DED. Hence, this study aims to fill this gap by proposing a machine-learning model based on process parameters to predict optimum bead profile geometries in Wa-DED. Implementing this model is expected to reduce the number of steps involved in part building, reducing material waste and overall lead time. To validate the effectiveness of the machine learning model, real-time experimentation is conducted, and the predicted process parameters are used to fabricate test coupons. These coupons are then subjected to microstructural and mechanical characterization, further validating the proposed approach.\u003c/p\u003e\u003cp\u003eThe work presents a machine learning model to predict the track profile Wa-DED. Subsequently, the model was experimentally validated by comparing the model prediction with the experimental measurements. Additionally, the printing parameters were selected from the validation data for the printing of the multilayer wall printing, and to understand the influence of heat input on the mechanical properties of the aluminum silicon alloys. By integrating machine learning with real-time experimentation and characterization, this work aims to enhance the efficiency and effectiveness of the Wa-DED process, ultimately contributing to the advancement of large-scale component fabrication in aerospace and other industries.\u003c/p\u003e"},{"header":"2 Materials and Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e\u003ch2\u003e2.1 Materials\u003c/h2\u003e\u003cp\u003eBead-on-plate studies were performed to develop a machine-learning model with ER4047 wire. The composition of the wire is outlined in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. The weld tracks are laid on mechanically finished and cleaned AA6082 cold rolled\u0026thinsp;+\u0026thinsp;T6 substrate (300 mm \u0026times; 50 mm \u0026times; 8 mm) using a 6-axis OTC welding robot equipped with an OTC Daihen synchro feed welding machine (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e (\u003cb\u003ea)\u003c/b\u003e). The robot controller is programmed automatically with deposition instructions before the beginning of each deposit, including welding parameters and the movement of the welding torch.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eChemical composition of ER 4047 welding wire\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"9\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eMaterial\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eMg\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSi\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eZn\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eFe\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eCu\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\"\u003e\u003cp\u003eMn\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\"\u003e\u003cp\u003eTi\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c9\"\u003e\u003cp\u003eAl\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003eER4047\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e11\u0026ndash;13\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c7\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c8\"\u003e\u003cp\u003e\u0026lt;\u0026thinsp;0.15\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c9\"\u003e\u003cp\u003eBal.\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec4\" class=\"Section2\"\u003e\u003ch2\u003e2.2 Data generation\u003c/h2\u003e\u003cp\u003eThe data is generated with the in-house developed Wa-DED facility by depositing 117 single tracks of length 100 mm for different current and travel speeds. The deposition is performed in syncro-feed mode. The deposits are further cross-sectioned, and the weld bead form factor [H/W] is measured at a minimum of three sites for each bead.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e\u003ch2\u003e2.3 Machine learning modeling\u003c/h2\u003e\u003cp\u003eThe machine learning predictive model uses current and travel speed as the input parameters and the H/W ratio as the output parameter. A machine learning classification approach was used in this work, and for classification, the H/W ratio\u0026thinsp;\u0026gt;\u0026thinsp;0.8 is assigned as one, and the ratio\u0026thinsp;\u0026le;\u0026thinsp;0.8 is given as zero. The classifier model was developed using, i.e., logistic (LR), decision tree (DT), support vector regressor (SVR), and random forest classification models. The model's performance was evaluated based on accuracy, precision, recall, F1-score, and ROC-AUC curve. All models were cross-validated using grid search cross-validation. The best model was also selected based on the performance matrix and used for further work.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec6\" class=\"Section2\"\u003e\u003ch2\u003e2.4 Validating the machine learning model\u003c/h2\u003e\u003cp\u003eThe best-selected model was validated experimentally by depositing a single track for the process parameters selected from the available literature on Wa-DED of eutectic Al-Si alloys [\u003cspan additionalcitationids=\"CR22\" citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]. It is worth mentioning that the selected parameters were different from the dataset used to develop the machine learning model.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec7\" class=\"Section2\"\u003e\u003ch2\u003e2.5 Test coupon printing and characterization\u003c/h2\u003e\u003cp\u003eAfter the validation of the ML model, test coupons were printed for the selected process parameters (Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). Porosity analysis of the test coupons was conducted using Image J software analysis, considering a total of 60 images, with 20 images captured at the thin walls' top, middle, and bottom sections. The test coupons were subjected to microstructural and mechanical characterization, and the mechanical properties were compared with the previously reported properties.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eProcess Parameters for Thin Walls\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWall No.\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCurrent (I)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eVoltage (V)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTS (cm/min)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eWFS\u003c/p\u003e\u003cp\u003e(cm/min)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e\u003cp\u003eHeat input (KJ/cm)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e113\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e15.63\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e42\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e665\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e2.019\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e86\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e14.76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e36\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e530\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.693\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e137\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e15.39\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e757\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1.534\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e"},{"header":"3 Results and discussions","content":"\u003cdiv id=\"Sec9\" class=\"Section2\"\u003e\u003ch2\u003e3.1 Machine learning model\u003c/h2\u003e\u003cdiv id=\"Sec10\" class=\"Section3\"\u003e\u003ch2\u003e3.1.1 Performance of the machine learning model\u003c/h2\u003e\u003cp\u003eThe decision tree and random forest model exhibited an accuracy of 83.75%, whereas the KNN model's accuracy was 81.25% (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e). A marginal variation in accuracy was observed for all three models.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe confusion matrix and receiver operating characteristic curve (ROC) were plotted to understand the prediction ability of different machine learning models (Figs.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and \u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e exhibited that the KNN model predicts maximum false negatives (5) cases, which means the actual H/W ratio is greater than 0.8 but predicted as \u0026le;\u0026thinsp;0.8. However, the decision tree predicted the minimum false negatives. On the contrary, the decision tree and KNN predicted the same false positive, i.e., indicated as \u0026gt;\u0026thinsp;0.8, whereas the values are below 0.8. The performance criteria of models suggested that the random forest model exhibited the highest F1 score compared to other models. Hence, based on accuracy and F1 score, random forest models outperform both models (Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003ePerformance matrix for different models\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eS.No\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePerformance criteria\u003c/p\u003e\u003cp\u003e(In percentage)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eDT\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eRF\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePrecision \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\text{T}\\text{P}}{(\\text{T}\\text{P}+\\text{F}\\text{P})}\\times\\:100\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e88\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e88\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e92\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSensitivity \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\frac{\\text{T}\\text{P}}{(\\text{T}\\text{P}+\\text{F}\\text{N})}\\)\u003c/span\u003e\u003c/span\u003e\u0026times;100\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e81.5\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e95.6\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e92\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eF1score\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\:\\:\\frac{\\text{P}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}\\times\\:\\text{S}\\text{e}\\text{n}\\text{s}\\text{i}\\text{t}\\text{i}\\text{v}\\text{i}\\text{t}\\text{y}\\:}{(\\text{P}\\text{r}\\text{e}\\text{c}\\text{i}\\text{s}\\text{i}\\text{o}\\text{n}+\\text{S}\\text{e}\\text{n}\\text{s}\\text{i}\\text{t}\\text{i}\\text{v}\\text{i}\\text{t}\\text{y}\\:)}\\times\\:2\\times\\:100\\)\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003e71\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e84.1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e84.64\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eThe area can confirm the better performance of the random forest model under the curve (AUC) of the probabilistic curve (ROC) shows in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. The mean AUC for the random forest model was the highest, i.e., 0.95, compared to KNN and the decision tree model (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e). Hence, it can be concluded that, based on the accuracy, F1 score, and AUC, the random forest model outperforms KNN and decision tree classification models.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec11\" class=\"Section3\"\u003e\u003ch2\u003e3.1.2 Validation of the machine learning model\u003c/h2\u003e\u003cp\u003eThe validation was performed by depositing a single bead for a different set of process parameters, as shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e. Comparing the experimental and model results, the model reasonably predicted the bead profile.\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eComparison between experimental and prediction values\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"8\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eS.No\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colspan=\"2\" nameend=\"c3\" namest=\"c2\"\u003e\u003cp\u003eProcess parameters\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eExperimental\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eCriteria value\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eKNN\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c7\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eDT\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c8\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003eRF\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eCurrent (I)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eSpeed (S)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e113\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e42\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.86\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e86\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e36\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.98\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e137\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e\u003cp\u003e66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e\u003cp\u003e0.76\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c5\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c6\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c7\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c8\"\u003e\u003cp\u003e0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e\u003cdiv id=\"Sec12\" class=\"Section2\"\u003e\u003ch2\u003e3.2 Test coupon printing and characterization.\u003c/h2\u003e\u003cdiv id=\"Sec13\" class=\"Section3\"\u003e\u003ch2\u003e3.2.1 Microstructural characterization\u003c/h2\u003e\u003cp\u003eThree Al-Si thin walls are fabricated for the parameters outlined in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, and the transverse sectional views of these walls are presented in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e (d) This study has been structured into two parts. The first condition emphasizes the significance of heat input (between samples 1 and 2). The second situation reflects equivalent heat input (between samples 2 and 3) since there is only a marginal difference in heat input between samples 2 and 3. Thus, wherever these effects are prominent, they will be called C1 and C2, respectively.\u003c/p\u003e\u003cp\u003eThe cross-sectional view of thin walls (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e (d) The small bulge can be observed due to slight deviations in the torch path during printing. Therefore, for the better-forming quality of Wa-DED structures, the torch must follow the path of the previously deposited layer laid down. Pores of enormous size were found in the top layer of the thin wall, with most pores being spherical. This is because the heat dissipation rate decreases as the deposition height increases, giving the pores additional time to aggregate and float to the surface [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e, \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Sample 2 displays fewer pores (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e) due to its lower heat input than sample 1, representing the effect of heat input (C1). Higher heat inputs can lead to gas entrapment in the flow due to intensified molten liquid dynamics, i.e., Marangoni flow. Higher heat inputs can also render alloying elements more vulnerable to evaporation, forming pores [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e]. It is essential to highlight that samples 3 and 2 demonstrate similar porosity levels because of their comparable heat inputs (C2).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe microstructural images of all samples indicate the presence of columnar α-Al dendrites and eutectic Si filling in the inter-dendritic spaces (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e and Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (a). Nevertheless, due to continuous remelting and reheating of the previously deposited layers, the typical microstructure characteristics, such as size and shape of dendrites, interlayer thickness, eutectic content, and directional growth patterns, are continuously changing along the deposition direction (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). The high heat input in sample 1 results in a prolonged solidification process, increasing the size of dendrites (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e (c)). Figure\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e (a) illustrates that the dendrites above the interlayer are oriented toward the center of the melt pool, which corresponds to the radial heat flow direction. The coarse and long columnar dendrites are more noticeable at the center of the molten pool than the dendrites at the fusion line (Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (a,b)).\u003c/p\u003e\u003cp\u003eAdditionally, directionality was weakened at the center (fusion zone) due to the reduced temperature gradient compared to the fusion boundary (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e (b)). In the interlayer, coarse α-Al dendrites with short and round Si were observed, whereas the melting zone exhibits fine and long dendrites (Fig.\u0026nbsp;\u003cspan refid=\"Fig6\" class=\"InternalRef\"\u003e6\u003c/span\u003e (a) and Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (a)). A porous structure can also be seen near the interlayer, resulting from the entrapment of oxides from the previously deposited layer within the molten pool. The dark regions in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (b) Those that Al essentially contains are called columnar α-Al, which starts to grow as soon as liquid metals solidify until it reaches the eutectic composition. At later stages (below eutectic temperature), the liquid in the two-phase mixture rich in Si will form Al-Si eutectics in the interdendritic regions. This explanation is well supported by the elemental distribution maps in Fig.\u0026nbsp;\u003cspan refid=\"Fig7\" class=\"InternalRef\"\u003e7\u003c/span\u003e (b).\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec14\" class=\"Section3\"\u003e\u003ch2\u003e3.2.2 Electron backscatter diffraction (EBSD) analysis\u003c/h2\u003e\u003cp\u003eElectron backscatter diffraction (EBSD) investigations were conducted on Wa-DED fabricated Al-Si alloy thin-wall structures deposited using different process parameters to analyses grain morphology, crystallographic texture, and misorientation behavior. The EBSD analyses were performed on samples extracted from the mid-height section of the deposited structures. Figure\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e (a, b, c) presents the inverse pole Fig. (IPF) maps corresponding to different processing conditions, showing predominantly equiaxed dendritic grain structures in the middle region. Although previous studies have reported the presence of columnar dendritic grain structures near the substrate region, aligned along the direction of the maximum thermal gradient [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. The bottom region acts as the primary nucleation zone, where the substrate interface governs grain growth by epitaxial solidification. Dendritic growth generally proceeds along preferred crystallographic directions aligned with the thermal gradient, typically\u0026thinsp;\u0026lt;\u0026thinsp;100\u0026thinsp;\u0026gt;\u0026thinsp;in face-centered cubic (FCC) structures such as Al alloys, due to the lowest resistance to growth in this orientation. [\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e].\u003c/p\u003e\u003cp\u003eAs the build height increases, the thermal gradient (G) and solidification velocity (R) at the solid\u0026ndash;liquid interface decrease. This reduction promotes solute buildup in the melt, leading to the formation of a constitutionally supercooled zone. Consequently, the grain morphology transitions from columnar to equiaxed dendrites at the middle and upper regions of the wall. This transformation is facilitated by heterogeneous nucleation at the center of the molten pool and the suppression of directional growth due to lateral thermal fluctuations. When the number and size of equiaxed dendrites in the melt reach a critical level, the advancing columnar front is blocked, promoting a fully equiaxed grain structure, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e (a, b, c). As a result, no dominant crystallographic orientation was observed. Moreover, the thermal exposure from subsequent layer deposition further alters the crystallographic orientation and influences grain growth. The average grain sizes measured from EBSD data for samples 1, 2, and 3 were approximately 104\u0026thinsp;\u0026plusmn;\u0026thinsp;32 \u0026micro;m, 92\u0026thinsp;\u0026plusmn;\u0026thinsp;18 \u0026micro;m, and 90\u0026thinsp;\u0026plusmn;\u0026thinsp;33 \u0026micro;m, respectively, as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e (d, e, f). The higher heat input in Sample 1 led to coarser grains, attributed to the reduced cooling rate and extended time for grain growth, which favored the formation of coarser equiaxed dendrites comparatively. In contrast, Samples 2 and 3, which experienced similar equivalent heat input levels, exhibited comparable grain sizes. The formation of finer grains at lower heat inputs is associated with a higher cooling rate and limited constitutional supercooling, which restricts grain growth.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eIn addition, the investigation assessed the distribution of dislocations in the as-deposited Al-Si thin-wall structures fabricated under different process parameters. The Kernel Average Misorientation (KAM) method was employed using EBSD orientation data to quantify local misorientation. KAM maps (Fig.\u0026nbsp;\u003cspan refid=\"Fig9\" class=\"InternalRef\"\u003e9\u003c/span\u003e (a, b, c) reveal localized misorientation gradients within grains, indicative of geometrically necessary dislocations (GNDs) that accommodate plastic deformation and thermal strain induced by cyclic heating and rapid solidification inherent to Wa-DED. This approach enables the estimation of dislocation density using the expression:\u003cdiv class=\"BlockQuote\"\u003e\u003cp\u003e\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\rho\\:}_{GND}\\cong\\:\\frac{{2\\theta\\:}_{KAM}}{bd}\\)\u003c/span\u003e\u003c/span\u003e \u003cb\u003eEq.\u0026nbsp;(1)\u003c/b\u003e\u003c/p\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003eHere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\theta\\:}_{KAM}\\:\\)\u003c/span\u003e\u003c/span\u003eis the average KAM value, \u003cem\u003eb\u003c/em\u003e is the Burger vector magnitude, and \u003cem\u003ed\u003c/em\u003e is the step size. KAM assigns a misorientation value to each point by averaging the angular differences between that point and its immediate neighbours. The calculated geometrically necessary dislocation (GND) densities for different parameters are presented in Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e, and are found to be on the order of 10\u003csup\u003e\u0026sup1;4\u003c/sup\u003e m⁻\u0026sup2;. Although Samples 2 and 3 were fabricated with the same nominal heat input, their dislocation densities differ markedly due to differences in applied current and associated thermal profiles. Sample 3, processed at a higher current, exhibited the highest dislocation density. The elevated current intensified thermal gradients, increasing localized plastic strain and thermal mismatch stresses during rapid solidification. These conditions promoted extensive dislocation generation, which outweighs the effects of enhanced dynamic recovery associated with its lower misorientation. Sample 1 showed a moderate dislocation density, reflecting a balance between dislocation generation from thermal strains and recovery due to prolonged high-temperature exposure. In contrast, Sample 2 displayed the lowest dislocation density, consistent with its lower peak thermal gradients and reduced accumulation of cyclic thermal strain.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eThe calculated Θ\u003csub\u003eKAM\u003c/sub\u003e and GND density based on KAM maps for Wa-DED-fabricated Al-Si alloy under different process parameters\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"3\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026times;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSpecimen\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eΘ\u003csub\u003eKAM\u003c/sub\u003e\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eDislocation density (m\u003csup\u003e\u0026minus;\u0026thinsp;2\u003c/sup\u003e)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSample-1\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.86\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e\u003cp\u003e2.1 \u0026times; 10\u003csup\u003e14\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSample-2\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.68\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e\u003cp\u003e1.7 \u0026times; 10\u003csup\u003e14\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003eSample-3\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e\u003cp\u003e0.98\u0026deg;\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026times;\" colname=\"c3\"\u003e\u003cp\u003e2.3 \u0026times; 10\u003csup\u003e14\u003c/sup\u003e\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe grain boundary misorientation angle distribution in all three samples (Fig.\u0026nbsp;\u003cspan refid=\"Fig10\" class=\"InternalRef\"\u003e10\u003c/span\u003e (a, b, c) demonstrates a predominance of low-angle grain boundaries (LAGBs, \u0026lt; 15\u0026deg; misorientation). The average misorientation angles were measured as approximately 8.14\u0026deg;, 10.14\u0026deg;, and 6.89\u0026deg; for Samples 1, 2, and 3, respectively. Sample 2 exhibited the highest average misorientation despite its finer grain size, which is attributed to rapid cooling and limited dynamic recovery. These conditions preserve a greater fraction of sub-grain boundaries without significant boundary migration. In contrast, Sample 3, produced with an equivalent overall heat input to Sample 2 but under a higher applied current, exhibited a lower average misorientation. The higher applied current likely increased the peak temperatures during deposition, enhancing dynamic recovery and promoting grain boundary migration. The higher heat input sample 1 depicts the intermediate average misorientation, reflecting the partial recovery facilitated by prolonged thermal exposure. The extended high-temperature conditions in Sample 1 allowed adjacent low-angle boundaries to migrate and merge, forming larger grains (as shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e (a) without necessarily increasing their mutual misorientation. This process resulted in overall grain coarsening rather than a proliferation of high-angle grain boundaries (HAGBs). These observations indicate that reduced thermal energy promotes higher sub-grain misorientation due to rapid cooling, limited recovery, and more pronounced constitutional supercooling conditions that enhance sub-grain formation while suppressing dislocation annihilation.\u003c/p\u003e\u003c/div\u003e\u003cdiv id=\"Sec15\" class=\"Section3\"\u003e\u003ch2\u003e3.2.3 Mechanical characterization\u003c/h2\u003e\u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig11\" class=\"InternalRef\"\u003e11\u003c/span\u003e describes the microhardness variation along the building direction of three fabricated samples subjected to different heat inputs. Forty hardness measurements were taken at intervals of 1 mm across the mid-width of the wall from the bottom to the top layer at 100 grams with a dwell time of 15 seconds. There is no noticeable difference in the average Vickers hardness values of the samples (Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e). Nevertheless, the decline in heat input is associated with a small increment in hardness, owing to rapid cooling rates and a consequent increase in dendritic arm spacing [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e, \u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Samples 1 and 3 display a wider variability in hardness values from the bottom to the top of the thin wall compared to sample 2, with no evidence of an increasing or decreasing trend. The hardness fluctuations observed in samples 1 and 3 can be attributed to their high currents, porosity content, and interaction with the indenter.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003eThe mechanical properties obtained from tensile testing are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig12\" class=\"InternalRef\"\u003e12\u003c/span\u003e and Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e. Sample 2 exhibits overall better properties with UTS of 187.56\u0026thinsp;\u0026plusmn;\u0026thinsp;11.40 and ductility of 8.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.66, owing to reduced porosity and fine dendritic arm structure (Fig.\u0026nbsp;\u003cspan refid=\"Fig8\" class=\"InternalRef\"\u003e8\u003c/span\u003e(a)). This improvement in sample 2 can be attributed to two key factors. The first is the rapid solidification process, which facilitates the formation of delicate dendritic structures at low heat inputs (C1). A second benefit of adopting low wire feed speeds (accompanied by lower currents and voltages) and travel speeds facilitate the development of delicate dendritic structures due to lower temperature gradients and faster cooling rates in the equivalent heat input condition (C2) [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Also, from the stress-strain plots, it is evident that sample 2 exhibits less variation among the replicate specimens than the other walls. Furthermore, the superior mechanical properties of sample-2 are attributed to its optimal combination of fine grain size (92\u0026thinsp;\u0026plusmn;\u0026thinsp;59 \u0026micro;m) and the highest average grain boundary misorientation angle (10.14\u0026deg;), resulting in a higher fraction of HAGBs that effectively impede the dislocation motion [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]. In a similar vein, Sample 3, which exhibits a grain size comparable to that of Sample 2, demonstrates noticeably higher mechanical properties owing to its substantially higher KAM value (0.98\u0026deg;), despite the predominance of LAGBs as indicated by its lower average misorientation angle (6.89\u0026deg;). Additionally, the higher heat input, subsequent cooling rate, and associated grain refinement in Sample-3 may contribute to its higher strength, as reflected in its average tensile strength in the vertical direction UTS\u003csub\u003eV,\u003c/sub\u003e of 180.73\u0026thinsp;\u0026plusmn;\u0026thinsp;8.01 (usually lower than the avg. UTS\u003csub\u003eH\u003c/sub\u003e). The present results are consistent with the past study and comparable with the as-cast structure (Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e).\u003c/p\u003e\u003cp\u003eThe higher heat input in wall-1 results in comparatively lower tensile strength, which is attributed to its coarse-grain structure, intermediate misorientation (8.14\u0026deg;), and moderate KAM (0.86). Combining coarser grains and fewer HAGBs underpins its lowest mechanical properties among the three samples. Therefore, these findings indicate that lower heat input in variable heat input conditions (C1), low wire feed speed, and travel speed during equivalent heat input conditions (C2) are the most favorable for achieving superior mechanical properties. Overall, the mechanical properties correlate predominantly with grain size rather than dislocation content (KAM values), and to a lesser extent with LAGB and HAGB fractions. This highlights the dominant strengthening role of grain boundary character over internal strain hardening in the additively manufactured Al 4047 thin walls analyzed.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eThe Avg mechanical properties of 3 samples along horizontal and vertical directions\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\"\u0026plusmn;\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eWall No.\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003eSample direction\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAvg. UTS\u003c/p\u003e\u003cp\u003e(MPa)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003e% EL\u003c/p\u003e\u003cp\u003e(%)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003eHardness (HV\u003csub\u003e0.1\u003c/sub\u003e)\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003e1\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e177.68\u0026thinsp;\u0026plusmn;\u0026thinsp;12.35\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e7.71\u0026thinsp;\u0026plusmn;\u0026thinsp;0.49\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e74.13\u0026thinsp;\u0026plusmn;\u0026thinsp;0.9\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eV\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e146.79\u0026thinsp;\u0026plusmn;\u0026thinsp;12.55\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e5.41\u0026thinsp;\u0026plusmn;\u0026thinsp;0.17\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003e2\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e187.56\u0026thinsp;\u0026plusmn;\u0026thinsp;11.40\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e8.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.66\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e73.39\u0026thinsp;\u0026plusmn;\u0026thinsp;0.37\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eV\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e165.38\u0026thinsp;\u0026plusmn;\u0026thinsp;7.34\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e6.19\u0026thinsp;\u0026plusmn;\u0026thinsp;0.96\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e\u003cb\u003e3\u003c/b\u003e\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eH\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e170.48\u0026thinsp;\u0026plusmn;\u0026thinsp;10.65\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e9.83\u0026thinsp;\u0026plusmn;\u0026thinsp;0.55\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c5\" morerows=\"1\" rowspan=\"2\"\u003e\u003cp\u003e73.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.69\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eV\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c3\"\u003e\u003cp\u003e180.73\u0026thinsp;\u0026plusmn;\u0026thinsp;8.01\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"char\" char=\"\u0026plusmn;\" colname=\"c4\"\u003e\u003cp\u003e6.18\u0026thinsp;\u0026plusmn;\u0026thinsp;0.96\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003cp\u003e\u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e\u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e\u003cdiv class=\"CaptionContent\"\u003e\u003cp\u003eSummary of current and previous studies on the mechanical properties of Al-Si alloys.\u003c/p\u003e\u003c/div\u003e\u003c/caption\u003e\u003ccolgroup cols=\"5\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e\u003cp\u003eReference\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e\u003cp\u003ePower Source/Manufacturing Route\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAlloy\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e\u003cp\u003eTensile strength (MPa)\u003c/p\u003e\u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e\u003cp\u003e% Elongation\u003c/p\u003e\u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWelding Wire\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eER4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e130\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e5\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eAs cast\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAl-12Si\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e211.62\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e2\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePresent Study (S1)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWa-DED (MIG)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eER4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e162.23\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e6.56\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePresent Study (S2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWa-DED (MIG)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eER4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e176.47\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e7.25\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003ePresent Study (S2)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWa-DED (MIG)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eER4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e175.60\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e8.0\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eWa-DED (GMAW)\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eER4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e180.46\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e15.1\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e\u003cp\u003e[\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e]\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c2\"\u003e\u003cp\u003eGMAW\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c3\"\u003e\u003cp\u003eAA4047\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c4\"\u003e\u003cp\u003e226\u003c/p\u003e\u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e\u003cp\u003e14\u003c/p\u003e\u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/colgroup\u003e\u003c/table\u003e\u003c/div\u003e\u003c/p\u003e\u003c/div\u003e\u003c/div\u003e"},{"header":"4 Conclusion","content":"\u003cp\u003eIn this study, a parameter-based machine learning model was developed and validated to predict the optimal bead profile in Wa-DED of Al-Si alloys. Real-time experiments confirmed the reliability of the model, with the random forest algorithm exhibiting superior classification performance compared to decision tree and KNN models, owing to its higher F1 score and AUC. Microstructural characterization revealed that lower heat input conditions (Sample 2) minimized porosity and favored finer dendritic structures, whereas higher heat input (Sample 1) resulted in coarser grains due to prolonged solidification. EBSD analysis confirmed the predominance of equiaxed dendritic morphologies without a preferred crystallographic orientation, with grain size being strongly dependent on heat input. Misorientation and dislocation characteristics were also highly sensitive to process conditions, where Sample 2 exhibited the highest average misorientation as a result of rapid cooling and limited recovery. While Sample 3, processed with higher current at equivalent heat input, developed the highest dislocation density due to intensified thermal gradients and plastic strain. These microstructural variations directly influenced mechanical performance, with Sample 2 achieving the best combination of strength (187.56\u0026thinsp;\u0026plusmn;\u0026thinsp;11.40 MPa) and ductility (8.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.66) due to its fine dendritic structure and reduced porosity. Notably, superior properties were not governed by overall heat input alone; rather, the correlation of individual parameters such as wire feed speed, current and travel speed also played a decisive role, as demonstrated by the improved performance of Sample 3 under equivalent heat input but altered parameter combinations. Overall, this study establishes that both heat input magnitude and the interplay of individual process parameters must be considered to optimize bead geometry, minimize porosity, refine microstructure, and enhance the mechanical properties of additively manufactured Al-Si thin-wall components.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWe acknowledge the support from the Aeronautics Research and Development Board of the Government of India under Grant No. ARDB/01/2031958/M/I.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConflicts of Interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThere are no conflicts of interest to declare by the authors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCode Availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eNot applicable\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eConsent to Participate\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll the authors have given consent to participate.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics Approval\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors hereby state that the present work complies with the ethical standards.\u003c/p\u003e\n\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eShivraman provided the experimental equipment; Hemachandra and Ramesh prepared the figure and wrote the manuscript; Shivraman reviewed and edited the manuscript; Shivraman, Hemachandra and Ramesh reviewed the manuscript.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eDing D, Pan Z, Cuiuri D, Li H (2015) A multi-bead overlapping model for robotic wire and arc additive manufacturing (WAAM). Robot Comput Integr Manuf 31:101\u0026ndash;110. https://doi.org/10.1016/j.rcim.2014.08.008\u003c/li\u003e\n\u003cli\u003eCampatelli G, Campanella D, Barcellona A, et al (2020) Microstructural, mechanical and energy demand characterization of alternative WAAM techniques for Al-alloy parts production. CIRP J Manuf Sci Technol 31:492\u0026ndash;499\u003c/li\u003e\n\u003cli\u003eDinovitzer M, Chen X, Laliberte J, et al (2019) Effect of wire and arc additive manufacturing (WAAM) process parameters on bead geometry and microstructure. Addit Manuf 26:138\u0026ndash;146. https://doi.org/10.1016/j.addma.2018.12.013\u003c/li\u003e\n\u003cli\u003eYang D, He C, Zhang G (2016) Forming characteristics of thin-wall steel parts by double electrode GMAW based additive manufacturing. J Mater Process Technol 227:153\u0026ndash;160. https://doi.org/10.1016/j.jmatprotec.2015.08.021\u003c/li\u003e\n\u003cli\u003eHemachandra M, Thapliyal S, Adepu K (2022) A review on microstructural and tribological performance of additively manufactured parts. J Mater Sci 57:17139\u0026ndash;17161. https://doi.org/10.1007/s10853-022-07736-1\u003c/li\u003e\n\u003cli\u003eThapliyal S (2019) Challenges associated with the wire arc additive manufacturing (WAAM) of aluminum alloys. Mater Res Express 6:. https://doi.org/10.1088/2053-1591/ab4dd4\u003c/li\u003e\n\u003cli\u003eTomar B, Shiva S, Nath T (2022) A review on wire arc additive manufacturing: Processing parameters, defects, quality improvement and recent advances. Mater Today Commun 31:103739\u003c/li\u003e\n\u003cli\u003eZavdoveev A, Pozniakov V, Baudin T, et al (2022) Optimization of the pulsed arc welding parameters for wire arc additive manufacturing in austenitic steel applications. Int J Adv Manuf Technol 119:5175\u0026ndash;5193\u003c/li\u003e\n\u003cli\u003eKumar A, Maji K (2020) Selection of Process Parameters for Near-Net Shape Deposition in Wire Arc Additive Manufacturing by Genetic Algorithm. J Mater Eng Perform 29:3334\u0026ndash;3352. https://doi.org/10.1007/s11665-020-04847-1\u003c/li\u003e\n\u003cli\u003eMai DS, Doan TK, Paris H, others (2021) Wire and arc additive manufacturing of 308L stainless steel components: Optimization of processing parameters and material properties. Eng Sci Technol an Int J 24:1015\u0026ndash;1026\u003c/li\u003e\n\u003cli\u003eGeng H, Xiong J, Huang D, et al (2017) A prediction model of layer geometrical size in wire and arc additive manufacture using response surface methodology. Int J Adv Manuf Technol 93:175\u0026ndash;186. https://doi.org/10.1007/s00170-015-8147-2\u003c/li\u003e\n\u003cli\u003eMamedipaka R, Hemachandra M, Thapliyal S (2024) Microstructural and mechanical characterization of direct energy deposited 316L stainless steel developed using an evolutionary computational approach. Proc Inst Mech Eng Part L J Mater Des Appl. https://doi.org/10.1177/14644207241262107\u003c/li\u003e\n\u003cli\u003eZhao Y, Li F, Chen S, Lu Z (2019) Unit block--based process planning strategy of WAAM for complex shell--shaped component. Int J Adv Manuf Technol 104:3915\u0026ndash;3927\u003c/li\u003e\n\u003cli\u003eDu Y, Mukherjee T, DebRoy T (2021) Physics-informed machine learning and mechanistic modeling of additive manufacturing to reduce defects. Appl Mater Today 24:. https://doi.org/10.1016/j.apmt.2021.101123\u003c/li\u003e\n\u003cli\u003eRaju KL, Thapliyal S, Sigatapu S, et al (2022) Process Parameter Dependent Machine Learning Model for Densification Prediction of Selective Laser Melted Al-50Si Alloy and its Validation. J Mater Eng Perform 0\u0026ndash;7. https://doi.org/10.1007/s11665-022-06831-3\u003c/li\u003e\n\u003cli\u003eDebRoy T, Mukherjee T, Wei HL, et al (2021) Metallurgy, mechanistic models and machine learning in metal printing. Nat Rev Mater 6:48\u0026ndash;68. https://doi.org/10.1038/s41578-020-00236-1\u003c/li\u003e\n\u003cli\u003eXie X, Bennett J, Saha S, et al (2021) Mechanistic data-driven prediction of as-built mechanical properties in metal additive manufacturing. npj Comput Mater 7:86\u003c/li\u003e\n\u003cli\u003ePopova E, Rodgers TM, Gong X, et al (2017) Process-structure linkages using a data science approach: application to simulated additive manufacturing data. Integr Mater Manuf Innov 6:54\u0026ndash;68\u003c/li\u003e\n\u003cli\u003eScime L, Beuth J (2019) Using machine learning to identify in-situ melt pool signatures indicative of flaw formation in a laser powder bed fusion additive manufacturing process. Addit Manuf 25:151\u0026ndash;165\u003c/li\u003e\n\u003cli\u003eThapliyal S, Mishra A (2021) Machine learning classification-based approach for mechanical properties of friction stir welding of copper. Manuf Lett 29:52\u0026ndash;55. https://doi.org/10.1016/j.mfglet.2021.05.010\u003c/li\u003e\n\u003cli\u003eGuo Y, Han Q, Hu J, et al (2022) Comparative study on wire-arc additive manufacturing and conventional casting of Al--Si alloys: porosity, microstructure and mechanical property. Acta Metall Sin (English Lett 35:475\u0026ndash;485\u003c/li\u003e\n\u003cli\u003eChen C, Sun G, Du W, et al (2023) Effect of equivalent heat input on WAAM Al-Si alloy. Int J Mech Sci 238:107831\u003c/li\u003e\n\u003cli\u003eLangelandsvik G, Horgar A, Furu T, et al (2020) Comparative study of eutectic Al-Si alloys manufactured by WAAM and casting. Int J Adv Manuf Technol 110:935\u0026ndash;947\u003c/li\u003e\n\u003cli\u003ePramod R, Kumar SM, Girinath B, et al (2020) Fabrication, characterisation, and finite element analysis of cold metal transfer--based wire and arc additive--manufactured aluminium alloy 4043 cylinder. Weld World 64:1905\u0026ndash;1919\u003c/li\u003e\n\u003cli\u003eGu J, Gao M, Yang S, et al (2019) Pore formation and evolution in wire+ arc additively manufactured 2319 Al alloy. Addit Manuf 30:100900\u003c/li\u003e\n\u003cli\u003eUtyaganova VR, Filippov A V, Shamarin NN, et al (2020) Controlling the porosity using exponential decay heat input regimes during electron beam wire-feed additive manufacturing of Al-Mg alloy. Int J Adv Manuf Technol 108:2823\u0026ndash;2838\u003c/li\u003e\n\u003cli\u003eRosli NA, Alkahari MR, bin Abdollah MF, et al (2021) Review on effect of heat input for wire arc additive manufacturing process. J Mater Res Technol 11:2127\u0026ndash;2145. https://doi.org/10.1016/j.jmrt.2021.02.002\u003c/li\u003e\n\u003cli\u003eDinda GP, Dasgupta AK, Bhattacharya S, et al (2013) Microstructural characterization of laser-deposited Al 4047 alloy. Metall Mater Trans A Phys Metall Mater Sci 44:2233\u0026ndash;2242. https://doi.org/10.1007/s11661-012-1560-3\u003c/li\u003e\n\u003cli\u003eChu Q, Bai R, Jian H, et al (2018) Microstructure, texture and mechanical properties of 6061 aluminum laser beam welded joints. Mater Charact 137:269\u0026ndash;276. https://doi.org/10.1016/j.matchar.2018.01.030\u003c/li\u003e\n\u003cli\u003eVandersluis E, Ravindran C (2017) Comparison of measurement methods for secondary dendrite arm spacing. Metallogr Microstruct Anal 6:89\u0026ndash;94\u003c/li\u003e\n\u003cli\u003eNajafi S, Eivani AR, Samaee M, et al (2018) A comprehensive investigation of the strengthening effects of dislocations, texture and low and high angle grain boundaries in ultrafine grained AA6063 aluminum alloy. Mater Charact 136:60\u0026ndash;68. https://doi.org/https://doi.org/10.1016/j.matchar.2017.12.004\u003c/li\u003e\n\u003cli\u003eHeard DW, Brophy S, Brochu M (2012) Solid freeform fabrication of Al--Si components via the CSC-MIG process. Can Metall Q 51:302\u0026ndash;312\u003c/li\u003e\n\u003cli\u003eDwivedi DK, Sharma R, Kumar A (2006) Influence of silicon content and heat treatment parameters on mechanical properties of cast Al--Si--Mg alloys. Int J Cast Met Res 19:275\u0026ndash;282\u003c/li\u003e\n\u003cli\u003eHaselhuhn AS, Buhr MW, Wijnen B, et al (2016) Structure-property relationships of common aluminum weld alloys utilized as feedstock for GMAW-based 3-D metal printing. Mater Sci Eng A 673:511\u0026ndash;523\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"silicon","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scon","sideBox":"Learn more about [Silicon](https://www.springer.com/journal/12633)","snPcode":"12633","submissionUrl":"https://submission.nature.com/new-submission/12633/3","title":"Silicon","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"Wire arc direct energy deposition (Wa-DED), Machine learning, Random Forest, Heat input, Characterization, Microstructure, Mechanical properties","lastPublishedDoi":"10.21203/rs.3.rs-7655300/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7655300/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThis study presents a parameter-based machine learning approach to predict optimal bead geometries in wire arc direct energy deposition (Wa-DED), aiming to reduce the time-consuming and costly trial-and-error procedures typically employed during process development. Decision trees, random forests, and K-nearest neighbors (KNN) models were trained and validated, with all three achieving comparable accuracy. Notably, the random forest model demonstrated superior performance in terms of accuracy, F1 score, and area under the curve (AUC). To further validate the optimized parameters, multilayer thin-walled Al-4047 structures were fabricated and comprehensively evaluated for both microstructural and mechanical properties. Microstructural analysis revealed α-Al dendrites with short, rounded Si in the interlayer and fine, fibrous Si in the melting zone, while EBSD confirmed predominantly equiaxed dendritic grains without a dominant crystallographic orientation, highlighting strong heterogeneity during solidification. Notably, sample-2 exhibited refined grains (92\u0026thinsp;\u0026plusmn;\u0026thinsp;59 \u0026micro;m) and the highest average misorientation angle (10.14\u0026deg;) owing to its lower heat input. Meanwhile Sample 3 fabricated at equivalent heat input but with higher current amplitude, exhibited elevated KAM values and the highest dislocation density (~\u0026thinsp;2.3 x 10\u0026sup1;⁴ m⁻\u0026sup2;) due to intensified thermal gradients and cyclic thermal strains. These microstructural features strongly correlated with mechanical behavior, as Sample 2 with its finer grains, higher misorientation, and reduced porosity, achieved superior tensile strength (187.56\u0026thinsp;\u0026plusmn;\u0026thinsp;11.40 MPa) and ductility (8.32\u0026thinsp;\u0026plusmn;\u0026thinsp;0.66%), thereby demonstrating the efficacy of combining machine learning optimization with microstructural validation in tailoring Wa-DED components.\u003c/p\u003e","manuscriptTitle":"Machine learning approach toward near-homogeneous properties of eutectic Aluminum silicon alloy fabricated by a wire arc direct energy deposition","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-02 12:41:52","doi":"10.21203/rs.3.rs-7655300/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2026-01-11T07:22:40+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-12-05T08:29:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"226822666063187613743607426920403805932","date":"2025-11-26T06:46:52+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"147881308736571112049949542054193060510","date":"2025-11-26T04:07:40+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"60574895970034740815759239335458535654","date":"2025-11-26T03:28:55+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-11-26T03:24:56+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-09-22T09:11:44+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-09-22T09:10:46+00:00","index":"","fulltext":""},{"type":"submitted","content":"Silicon","date":"2025-09-19T06:39:49+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"silicon","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"scon","sideBox":"Learn more about [Silicon](https://www.springer.com/journal/12633)","snPcode":"12633","submissionUrl":"https://submission.nature.com/new-submission/12633/3","title":"Silicon","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"em","reportingPortfolio":"Springer Hybrid","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"6c540ac5-80e9-4f5e-997d-6108c0379d74","owner":[],"postedDate":"December 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[],"tags":[],"updatedAt":"2026-02-15T03:53:17+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-02 12:41:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7655300","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7655300","identity":"rs-7655300","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}