Optimizing and Predicting Track Quality in Multilayer Laser Melting of Inconel 718: A Numerical, Experimental and Machine Learning Approach

Optimizing and Predicting Track Quality in Multilayer Laser Melting of Inconel 718: A Numerical, Experimental and Machine Learning Approach | 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 Optimizing and Predicting Track Quality in Multilayer Laser Melting of Inconel 718: A Numerical, Experimental and Machine Learning Approach ANTONY KURIAN, BIBIN JOSE, JINA VARGHESE, ADITHYA BHARATH SHAH, and 3 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-5922562/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract In this paper multilayer fabrication of Inconel 718 using SLM, a widely used additive manufacturing technique was conducted. The quality of the fabricated tracks is dependent on laser power, scanning speed and beam diameter. The effects of these parameters during SLM process were investigated by numerical analysis as well as by experimentation to determine the optimized set of parameters. However, process optimization using experimentation and simulation is expensive as well as time consuming. Therefore, in addition, two machine learning models have been trained to predict the width and height of the fabricated tracks based on the parameters. By numerical analysis, the variation in temperature distribution was determined by varying these parameters and the effect of these parameters was validated experimentally. The effect of these parameters on the geometry, morphology of the track was analyzed by utilizing tools such as macroscope, SEM and EDAX. Among the two ML models, linear regression models demonstrated superior predictive accuracy with less error compared to RNN models. SLM Inconel 718 powder Machine learning SEM FEA RNN Figures Figure 1 Figure 2 Figure 3 1. INTRODUCTION Inconel 718 is widely utilized in the aviation industry for turbine blades, combustion chambers, and other parts due to its exceptional resistance to oxidation and hot corrosion. Inconel 718 industry faces challenges in obtaining intricate and highly accurate parts due to quality standards. Therefore, non-conventional technology is necessary for producing complex, high-accuracy parts with minimal defects [ 1 – 5 ]. Selective Laser Melting is a popular additive manufacturing technique where a thin powder layer is deposited onto a base, and a laser beam scans through the predefined tracks to fuse the powder particles together. This process is repeated until the required part is created [ 6 – 8 ]. The influence of the process parameters on the laser melting process of Inconel 718 which includes the effect of the laser energy on the powder layer and the densification associated with each of the combinations were observed that with high energy density the densification was increased [ 9 – 11 ]. The investigation of the SLMed track's microstructure revealed a fine structure, enhancing the microhardness of the formed part. Numerous experimental studies have examined the mechanical characteristics and microstructure of Inconel 718 components using SLM, but no research has been explored the impact of processing parameters and their optimization on IN718's multilayer laser melting [ 12 ]. Another strategy for parameter optimization in SLM is to investigate the relation between the process parameters and the resulting track using a machine learning algorithm. This paper presents a finite element model for studying temperature distribution during powder melting and the formation of multi-layer laser melted tracks. The model is applied to experimentally vary process parameters using a macroscope to capture surface characteristics. The model is used to predict track width and depth based on unknown parameter combinations, reducing time and cost. 2. METHODOLOGY The material considered for this work was Inconel 718 and numerical analysis was conducted initially to determine the temperature distribution by varying the parameters and laser melting of the multi layered tracks were fabricated. The surface characteristics was observed through the macrosocope for identifying the dimensions of the tracks formed. This data was used as training data for machine learning model for prediction. 2.1 Numerical Analysis The FEA analysis was conducted on ANSYS to determine the temperature distribution and to reduce the complexity and minimize computation time. Certain assumptions were considered such as the flow of the molten pool due to Marangoni effect were neglected and the powder layer were modeled as homogenous. The laser source is modeled as gaussian heat source which were made to travel through x axis through a length of 10 mm and the subsequent temperature distribution was compared with the experimental results. 2.2 Experimentation Inconel 718 was used as the substrate. The substrate was 80 mm x 60 mm x 3mm of size which was cut from IN718 base plate. The substrate was subjected to grinding using a surface grinding machine to make the surface even and emery abrasive paper (grit 150) was used to make the surface of the substrate coarser for facilitating the proper deposition of the powder layer. Sieve shaker machine was used to separate the powder particles for accurate particle separation. IN718 powder was initially spread over the substrate to a layer thickness of 0.3 mm (300 µm). Then the specimen was preheated to a temperature of 600°C using a muffle furnace. Process parameters such as laser power, laser speed and laser diameter were varied to study the effect of these parameters during the multilayer laser melting process. In the experiment Nd-Yag laser was used as the source for the melting of the IN718 powder. Due to certain limitations for conducting full experiments, the L9 Taguchi array shown in Table 1 is used to determine the possible combinations of the parameters. Table 1 L9 Taguchi array Run No. Laser Power Laser Speed Spot Diameter 1 350 40 0.6 2 350 45 0.2 3 350 50 0.4 4 400 40 0.2 5 400 45 0.4 6 400 50 0.6 7 450 40 0.4 8 450 45 0.6 9 450 50 0.2 2.3 Machine learning models The proposed models involve multiple regression and RNN to predict track width and depth based on input dimensions of process parameters. The error metric used is maximum error. Linear regression is a statistical method that models the relationship between independent and continuous variables and aims to find the best fitting line that minimizes the gap between predicted and actual values. RNN are generally used to process the sequential information. The inputs are given sequentially and the input in each timestep uses the processed information from its previous timestep. 2.3.1. Data Augmentation The multivariate normal distribution is a variation of the univariate normal distribution for two or more variables, describing the joint distribution of multiple random variables with different mean and variance and may be correlated with others, ensuring consistency in the distribution of the random vectors. In this distribution, the probability density function is characterized by a mean vector and a covariance matrix, which captures the relationships and variability among the variables. Figure 1 . shows the overall experimental methodology. 3. RESULT AND DISCUSSION 3.1 Numerical analysis and Experimentation The detailed numerical and experimental analysis of all the 9 sets of tracks is presented in Table 2. Figure 2 (a-d) shows the numerical and experimental analysis of optimized track (track 9 in Table 2). Figure 2 (a) shows the FEA Simulation of the optimized track. From the numerical analysis conducted it was observed that when the laser power P = 450 W, scan speed v = 50 mm/s and spot diameter was 0.2 mm, the temperature developed was over 2000°C which was sufficient for the melting of the powder particles. Figure 2 (b) show longitudinal macrostructure and Fig. 2 (c) Cross-sectional macrostructure of the track. The macroscopic analysis of the track is smooth and continuous due to the optimum set of parameters. The SEM analysis conducted as shown in Fig. 2 (d) shows that the track was formed properly, and the powder particles had melted properly without any distortions making the parameters optimum for fabrication of parts during SLM process. EDAX analysis was also conducted on the track and was observed that no major variation from the original chemical composition and the elements which constituted the majority of the composition was Nickel and Chromium which improves the corrosion and oxidation resistance. Composition of Oxygen was less than 5% indicating minor oxidation. 3.2 Prediction of Track width and height The trained models are compared using maximum error. The predicted values of the width are shown in Fig. 3. The error is least in the linear regression model in comparison with Recurrent Neural Network (RNN). In the linear regression model the test data of the width given was 0.637 mm and the model after training predicted 0.581mm. while in the RNN model the test data of the width given was 0.685 mm and the model predicted 0.570 mm. While in the linear regression model the test data of the depth given was 0.691mm and the model after training predicted 0.6971 mm, which is closely accurate. In RNN model the test data of the depth given was 0.722 mm and model predicted 0.193 mm. Maximum error obtained by the linear regression model is 4.27%, while the maximum error obtained by the RNN model is 52.8%. In comparison with the models it has been found that linear regression predicts with less error than RNN. CONCLUSION This study conducted an extensive investigation, utilizing both numerical and experimental methods, to examine the influence of laser parameters—namely scanning speed, laser power, and beam diameter—on the multi-layer laser melting of Inconel 718 powder. The key findings from the study are as follows: A FEA was performed to determine the temperature distribution resulting from various process parameters, and these findings were validated through experimental methods. Optimal parameters were identified: a laser power of 450W, scan speed of 50 mm/s, and beam diameter of 0.2 mm produced a smooth and continuous track without distortions. SEM and EDAX were used to analyze this track. ML models were also explored, with two models trained to predict track width and depth based on process parameters. The Linear regression model outperformed the RNN model, yielding more accurate predictions. Declarations Funding Declaration The authors declare that no funding for conducting this research work. Ethics and Consent to Publish declarations : Not applicable Data Availability Declaration The authors declare that the data supporting the findings of this work are available within the manuscript Competing Interest Declaration The authors declare that no competing Interest for this research work Author Contribution The authors Dr Antony Kurian and Dr Bibin Jose wrote the main content of the manuscript. The Machine Learning approach and programming was done by Dr Jina Varghese. The experimental runs were performed by Mr Adithya bharath shah, George Antony,V S Chris and Yaseen Ahammad References Patalas-Maliszewska, J., Feldshtein, E., Devojno, O., Śliwa, M., Kardapolava, M., & Lutsko, N. (2020). Single tracks as a key factor in additive manufacturing technology—analysis of research trends and metal deposition behavior. Materials , 13 (5), 1115. Razavykia, A., Brusa, E., Delprete, C., & Yavari, R. (2020). An overview of additive manufacturing technologies—a review to technical synthesis in numerical study of selective laser melting. Materials , 13 (17), 3895. Soujon, M., Kallien, Z., Roos, A., Zeller-Plumhoff, B., & Klusemann, B. (2022). Fundamental study of multi-track friction surfacing deposits for dissimilar aluminum alloys with application to additive manufacturing. Materials & Design , 219 , 110786. Jose, B., Manoharan, M., Natarajan, A., Muktinutalapati, N. R., Madhusudhan Reddy, G., & Meshram, S. D. (2023). Current research and developments in welding of 18% nickel maraging steel. Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and Applications , 237 (6), 1295-1318. Jia, Qingbo & Gu, Dongdong. (2014). Selective laser melting additive manufacturing of Inconel 718 superalloy parts: Densification, microstructure and properties. Journal of Alloys and Compounds. 585. 713-721. Zhong, C., Gasser, A., Backes, G., Fu, J., & Schleifenbaum, J. H. (2022). Laser additive manufacturing of Inconel 718 at increased deposition rates. Materials Science and Engineering: A , 844 , 143196. K. Moussaoui, W. Rubio, M. Mousseigne, T. Sultan, F. Rezai. (2018). Effects of Selective Laser Melting additive manufacturing parameters of Inconel 718 on porosity, microstructure and mechanical properties. Materials Science and Engineering: A. 735. Antony, K., Arivazhagan, N., & Senthilkumaran, K. (2014). Numerical and experimental investigations on laser melting of stainless steel 316L metal powders. Journal of Manufacturing Processes , 16 (3), 345-355. Kumar, P., Chakravadhhanula, V.S.K., Manwatkar, S.K. et al. (2021) Establishing the Qualitative Relationship Between Process Parameters: Microstructure, Phases and Defects in SLM-PBF Manufactured and Heat-Treated Inconel 718 Alloy. Trans Indian Natl. Acad. Eng. 6, 1083–1097 Jose, B., Manikandan, M., Arivazhagan, N., Muktinutalapati, N. R., & Madhusudhan Reddy, G. (2023). Development of a novel welding technique with reduced heat-input by employing double-pulsed gas metal arc welding for aerospace grade 18% Ni maraging steel. Journal of Manufacturing Science and Engineering , 145 (2), 021002. Xu J, Wu Z, Niu J, Song Y, Liang C, Yang K, Chen Y, Liu Y. Effect of Laser Energy Density on the Microstructure and Microhardness of Inconel 718 Alloy Fabricated by Selective Laser Melting. Crystals. 2022; 12(9):1243. Nayak, S. K., Mishra, S. K., Jinoop, A. N., Paul, C. P., & Bindra, K. S. (2020). Experimental studies on laser additive manufacturing of Inconel-625 structures using powder bed fusion at 100 µm layer thickness. Journal of Materials Engineering and Performance , 29 , 7636-7647. Tables Tables 2 is available in the Supplementary Files section. Additional Declarations No competing interests reported. Supplementary Files TABLE2.docx Cite Share Download PDF Status: Posted Version 1 posted 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-5922562","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":411490762,"identity":"4f86de6f-674b-45fc-9148-736245ed53e8","order_by":0,"name":"ANTONY KURIAN","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABA0lEQVRIiWNgGAWjYBACCSA+AOd9YGDgQZIgoAWklHEGkOYhRgsDTAszDwPCGpxAckbuwwM/d9jl2bP3HpO2zamVsec/wPjhB4NFHi4t0hLpBgd7zyQX8/CcS5PO3Xach0cigVmyh0GiGJcWOYk0hgO8bcyJPRI5ZkAtx4BagAYBXZzYgEfLwb9t9RAtliAt/AeYf+PTIg3Ucpi37TBEC+O2GmCIJbDhtUWy5xnDYdm244k9Z84YW/ZuO8DDcyOxzbLHALcWieNpzB/ftlUntrf3GN74ua3Onr3/8OEbPyrqcGpBBixAjx8G0oxAxQZEqAcCZmB6qSNO6SgYBaNgFIwoAADgw07wcV6GHQAAAABJRU5ErkJggg==","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":true,"prefix":"","firstName":"ANTONY","middleName":"","lastName":"KURIAN","suffix":""},{"id":411490763,"identity":"e2ee0da8-1c56-4f18-9029-c7d37d28333d","order_by":1,"name":"BIBIN JOSE","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"BIBIN","middleName":"","lastName":"JOSE","suffix":""},{"id":411490764,"identity":"8b3600ab-8e6e-4d99-8a0a-fe2123d5bf09","order_by":2,"name":"JINA VARGHESE","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"JINA","middleName":"","lastName":"VARGHESE","suffix":""},{"id":411490765,"identity":"f15f3fe3-13bb-4ac7-af7d-a214e4c35596","order_by":3,"name":"ADITHYA BHARATH SHAH","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"ADITHYA","middleName":"BHARATH","lastName":"SHAH","suffix":""},{"id":411490766,"identity":"43930c98-2f69-4515-a3fd-ccc457e17658","order_by":4,"name":"GEORGE ANTONY","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"GEORGE","middleName":"","lastName":"ANTONY","suffix":""},{"id":411490767,"identity":"2c65d69e-171c-4347-8da2-fa61b0e1f11a","order_by":5,"name":"V S CHRIS","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"V","middleName":"S","lastName":"CHRIS","suffix":""},{"id":411490768,"identity":"11571a64-c1a1-464a-b9d8-b2321def4615","order_by":6,"name":"YASEEN AHAMMAD","email":"","orcid":"","institution":"Muthoot Institute of Technology \u0026 Science","correspondingAuthor":false,"prefix":"","firstName":"YASEEN","middleName":"","lastName":"AHAMMAD","suffix":""}],"badges":[],"createdAt":"2025-01-29 07:38:26","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-5922562/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-5922562/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":75734986,"identity":"f40aa71c-fe14-4e81-9d90-5b6be9c7014c","added_by":"auto","created_at":"2025-02-07 15:27:40","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":116169,"visible":true,"origin":"","legend":"\u003cp\u003eExperimental methodology\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-5922562/v1/642bcb3e6acfed76ab32bdbf.png"},{"id":75734959,"identity":"e5a797a6-c044-4a67-919e-9dad2a0af88f","added_by":"auto","created_at":"2025-02-07 15:27:38","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":348806,"visible":true,"origin":"","legend":"\u003cp\u003e(a-d)\u003cstrong\u003e \u003c/strong\u003eNumerical and experimental analysis of optimized track (a) FEA Simulation, (b) Longitudinal macrostructure, (c) Cross-sectional macrostructure and (d) SEM-EDS analysis of the multi-layered track\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-5922562/v1/c66e957b8583113187f7efe3.png"},{"id":75734954,"identity":"abf22bde-36c5-4c33-a6d2-8121dbe038bd","added_by":"auto","created_at":"2025-02-07 15:27:38","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":18857,"visible":true,"origin":"","legend":"\u003cp\u003ePredicted track width compared to actual width\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-5922562/v1/babda6aba5dd468afa0b168e.png"},{"id":97249495,"identity":"c80b507e-33db-42b6-b9bf-36b5dfd75695","added_by":"auto","created_at":"2025-12-02 13:12:45","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":949628,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-5922562/v1/91eed012-3285-44bb-9d60-2e0df9d599f8.pdf"},{"id":75736795,"identity":"0a32dd9a-abd8-43e7-9f76-9c8eb4527c22","added_by":"auto","created_at":"2025-02-07 15:43:38","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":6141727,"visible":true,"origin":"","legend":"","description":"","filename":"TABLE2.docx","url":"https://assets-eu.researchsquare.com/files/rs-5922562/v1/1029a5f8ebc05e8497dbeace.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"Optimizing and Predicting Track Quality in Multilayer Laser Melting of Inconel 718: A Numerical, Experimental and Machine Learning Approach","fulltext":[{"header":"1. INTRODUCTION","content":"\u003cp\u003eInconel 718 is widely utilized in the aviation industry for turbine blades, combustion chambers, and other parts due to its exceptional resistance to oxidation and hot corrosion. Inconel 718 industry faces challenges in obtaining intricate and highly accurate parts due to quality standards. Therefore, non-conventional technology is necessary for producing complex, high-accuracy parts with minimal defects [\u003cspan additionalcitationids=\"CR2 CR3 CR4\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. Selective Laser Melting is a popular additive manufacturing technique where a thin powder layer is deposited onto a base, and a laser beam scans through the predefined tracks to fuse the powder particles together. This process is repeated until the required part is created [\u003cspan additionalcitationids=\"CR7\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eThe influence of the process parameters on the laser melting process of Inconel 718 which includes the effect of the laser energy on the powder layer and the densification associated with each of the combinations were observed that with high energy density the densification was increased [\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. The investigation of the SLMed track's microstructure revealed a fine structure, enhancing the microhardness of the formed part. Numerous experimental studies have examined the mechanical characteristics and microstructure of Inconel 718 components using SLM, but no research has been explored the impact of processing parameters and their optimization on IN718's multilayer laser melting [\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAnother strategy for parameter optimization in SLM is to investigate the relation between the process parameters and the resulting track using a machine learning algorithm.\u003c/p\u003e \u003cp\u003eThis paper presents a finite element model for studying temperature distribution during powder melting and the formation of multi-layer laser melted tracks. The model is applied to experimentally vary process parameters using a macroscope to capture surface characteristics. The model is used to predict track width and depth based on unknown parameter combinations, reducing time and cost.\u003c/p\u003e"},{"header":"2. METHODOLOGY","content":"\u003cp\u003eThe material considered for this work was Inconel 718 and numerical analysis was conducted initially to determine the temperature distribution by varying the parameters and laser melting of the multi layered tracks were fabricated. The surface characteristics was observed through the macrosocope for identifying the dimensions of the tracks formed. This data was used as training data for machine learning model for prediction.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003e2.1 Numerical Analysis\u003c/h2\u003e \u003cp\u003eThe FEA analysis was conducted on ANSYS to determine the temperature distribution and to reduce the complexity and minimize computation time. Certain assumptions were considered such as the flow of the molten pool due to Marangoni effect were neglected and the powder layer were modeled as homogenous. The laser source is modeled as gaussian heat source which were made to travel through x axis through a length of 10 mm and the subsequent temperature distribution was compared with the experimental results.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec4\" class=\"Section2\"\u003e \u003ch2\u003e2.2 Experimentation\u003c/h2\u003e \u003cp\u003eInconel 718 was used as the substrate. The substrate was 80 mm x 60 mm x 3mm of size which was cut from IN718 base plate. The substrate was subjected to grinding using a surface grinding machine to make the surface even and emery abrasive paper (grit 150) was used to make the surface of the substrate coarser for facilitating the proper deposition of the powder layer. Sieve shaker machine was used to separate the powder particles for accurate particle separation. IN718 powder was initially spread over the substrate to a layer thickness of 0.3 mm (300 \u0026micro;m). Then the specimen was preheated to a temperature of 600\u0026deg;C using a muffle furnace. Process parameters such as laser power, laser speed and laser diameter were varied to study the effect of these parameters during the multilayer laser melting process. In the experiment Nd-Yag laser was used as the source for the melting of the IN718 powder. Due to certain limitations for conducting full experiments, the L9 Taguchi array shown in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e is used to determine the possible combinations of the parameters.\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\u003eL9 Taguchi array\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\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 \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eRun No.\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eLaser Power\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eLaser Speed\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSpot Diameter\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\u003e350\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6\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\u003e350\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.2\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\u003e350\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.2\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e400\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e450\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e40\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.4\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e450\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e45\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.6\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e450\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e50\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.2\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=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003e2.3 Machine learning models\u003c/h2\u003e \u003cp\u003eThe proposed models involve multiple regression and RNN to predict track width and depth based on input dimensions of process parameters. The error metric used is maximum error. Linear regression is a statistical method that models the relationship between independent and continuous variables and aims to find the best fitting line that minimizes the gap between predicted and actual values. RNN are generally used to process the sequential information. The inputs are given sequentially and the input in each timestep uses the processed information from its previous timestep.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cdiv id=\"Sec6\" class=\"Section3\"\u003e \u003ch2\u003e2.3.1. Data Augmentation\u003c/h2\u003e \u003cp\u003eThe multivariate normal distribution is a variation of the univariate normal distribution for two or more variables, describing the joint distribution of multiple random variables with different mean and variance and may be correlated with others, ensuring consistency in the distribution of the random vectors. In this distribution, the probability density function is characterized by a mean vector and a covariance matrix, which captures the relationships and variability among the variables. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e. shows the overall experimental methodology.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e"},{"header":"3. RESULT AND DISCUSSION","content":"\u003cdiv id=\"Sec8\"\u003e\n \u003ch2\u003e3.1 Numerical analysis and Experimentation\u003c/h2\u003e\n \u003cp\u003eThe detailed numerical and experimental analysis of all the 9 sets of tracks is presented in Table 2. Figure 2 (a-d) shows the numerical and experimental analysis of optimized track (track 9 in Table 2). Figure 2 (a) shows the FEA Simulation of the optimized track. From the numerical analysis conducted it was observed that when the laser power P\u0026thinsp;=\u0026thinsp;450 W, scan speed v\u0026thinsp;=\u0026thinsp;50 mm/s and spot diameter was 0.2 mm, the temperature developed was over 2000\u0026deg;C which was sufficient for the melting of the powder particles.\u003c/p\u003e\n \u003cp\u003eFigure 2 (b) show longitudinal macrostructure and Fig. 2 (c) Cross-sectional macrostructure of the track. The macroscopic analysis of the track is smooth and continuous due to the optimum set of parameters. The SEM analysis conducted as shown in Fig. 2 (d) shows that the track was formed properly, and the powder particles had melted properly without any distortions making the parameters optimum for fabrication of parts during SLM process. EDAX analysis was also conducted on the track and was observed that no major variation from the original chemical composition and the elements which constituted the majority of the composition was Nickel and Chromium which improves the corrosion and oxidation resistance. Composition of Oxygen was less than 5% indicating minor oxidation.\u003c/p\u003e\n\u003c/div\u003e\n\u003cdiv id=\"Sec9\"\u003e\n \u003ch2\u003e3.2 Prediction of Track width and height\u003c/h2\u003e\n \u003cp\u003eThe trained models are compared using maximum error. The predicted values of the width are shown in Fig. 3. The error is least in the linear regression model in comparison with Recurrent Neural Network (RNN). In the linear regression model the test data of the width given was 0.637 mm and the model after training predicted 0.581mm. while in the RNN model the test data of the width given was 0.685 mm and the model predicted 0.570 mm. While in the linear regression model the test data of the depth given was 0.691mm and the model after training predicted 0.6971 mm, which is closely accurate. In RNN model the test data of the depth given was 0.722 mm and model predicted 0.193 mm. Maximum error obtained by the linear regression model is 4.27%, while the maximum error obtained by the RNN model is 52.8%. In comparison with the models it has been found that linear regression predicts with less error than RNN.\u003c/p\u003e\n \u003cp\u003e\u003cstrong\u003eCONCLUSION\u003c/strong\u003e\u003c/p\u003e\n \u003cp\u003eThis study conducted an extensive investigation, utilizing both numerical and experimental methods, to examine the influence of laser parameters\u0026mdash;namely scanning speed, laser power, and beam diameter\u0026mdash;on the multi-layer laser melting of Inconel 718 powder. The key findings from the study are as follows:\u003c/p\u003e\n \u003cul\u003e\n \u003cli\u003e\n \u003cp\u003eA FEA was performed to determine the temperature distribution resulting from various process parameters, and these findings were validated through experimental methods.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eOptimal parameters were identified: a laser power of 450W, scan speed of 50 mm/s, and beam diameter of 0.2 mm produced a smooth and continuous track without distortions. SEM and EDAX were used to analyze this track.\u003c/p\u003e\n \u003c/li\u003e\n \u003cli\u003e\n \u003cp\u003eML models were also explored, with two models trained to predict track width and depth based on process parameters. The Linear regression model outperformed the RNN model, yielding more accurate predictions.\u003c/p\u003e\n \u003c/li\u003e\n \u003c/ul\u003e\n\u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eFunding Declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that no funding for conducting this research work.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthics and Consent to Publish declarations : Not applicable\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData Availability Declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that the data supporting the findings of this work are available within the manuscript\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting Interest Declaration\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare that no competing Interest for this research work\u003c/p\u003e\u003ch2\u003eAuthor Contribution\u003c/h2\u003e\u003cp\u003eThe authors Dr Antony Kurian and Dr Bibin Jose wrote the main content of the manuscript. The Machine Learning approach and programming was done by Dr Jina Varghese. The experimental runs were performed by Mr Adithya bharath shah, George Antony,V S Chris and Yaseen Ahammad\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003ePatalas-Maliszewska, J., Feldshtein, E., Devojno, O., Śliwa, M., Kardapolava, M., \u0026amp; Lutsko, N. (2020). Single tracks as a key factor in additive manufacturing technology\u0026mdash;analysis of research trends and metal deposition behavior. \u003cem\u003eMaterials\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(5), 1115.\u003c/li\u003e\n\u003cli\u003eRazavykia, A., Brusa, E., Delprete, C., \u0026amp; Yavari, R. (2020). An overview of additive manufacturing technologies\u0026mdash;a review to technical synthesis in numerical study of selective laser melting. \u003cem\u003eMaterials\u003c/em\u003e, \u003cem\u003e13\u003c/em\u003e(17), 3895.\u003c/li\u003e\n\u003cli\u003eSoujon, M., Kallien, Z., Roos, A., Zeller-Plumhoff, B., \u0026amp; Klusemann, B. (2022). Fundamental study of multi-track friction surfacing deposits for dissimilar aluminum alloys with application to additive manufacturing. \u003cem\u003eMaterials \u0026amp; Design\u003c/em\u003e, \u003cem\u003e219\u003c/em\u003e, 110786.\u003c/li\u003e\n\u003cli\u003eJose, B., Manoharan, M., Natarajan, A., Muktinutalapati, N. R., Madhusudhan Reddy, G., \u0026amp; Meshram, S. D. (2023). Current research and developments in welding of 18% nickel maraging steel. \u003cem\u003eProceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and Applications\u003c/em\u003e, \u003cem\u003e237\u003c/em\u003e(6), 1295-1318.\u003c/li\u003e\n\u003cli\u003eJia, Qingbo \u0026amp; Gu, Dongdong. (2014). Selective laser melting additive manufacturing of Inconel 718 superalloy parts: Densification, microstructure and properties. Journal of Alloys and Compounds. 585. 713-721.\u003c/li\u003e\n\u003cli\u003eZhong, C., Gasser, A., Backes, G., Fu, J., \u0026amp; Schleifenbaum, J. H. (2022). Laser additive manufacturing of Inconel 718 at increased deposition rates. \u003cem\u003eMaterials Science and Engineering: A\u003c/em\u003e, \u003cem\u003e844\u003c/em\u003e, 143196.\u003c/li\u003e\n\u003cli\u003eK. Moussaoui, W. Rubio, M. Mousseigne, T. Sultan, F. Rezai. (2018). Effects of Selective Laser Melting additive manufacturing parameters of Inconel 718 on porosity, microstructure and mechanical properties. Materials Science and Engineering: A. 735.\u003c/li\u003e\n\u003cli\u003eAntony, K., Arivazhagan, N., \u0026amp; Senthilkumaran, K. (2014). Numerical and experimental investigations on laser melting of stainless steel 316L metal powders. \u003cem\u003eJournal of Manufacturing Processes\u003c/em\u003e, \u003cem\u003e16\u003c/em\u003e(3), 345-355.\u003c/li\u003e\n\u003cli\u003eKumar, P., Chakravadhhanula, V.S.K., Manwatkar, S.K. et al. (2021) Establishing the Qualitative Relationship Between Process Parameters: Microstructure, Phases and Defects in SLM-PBF Manufactured and Heat-Treated Inconel 718 Alloy. Trans Indian Natl. Acad. Eng. 6, 1083\u0026ndash;1097 \u003c/li\u003e\n\u003cli\u003eJose, B., Manikandan, M., Arivazhagan, N., Muktinutalapati, N. R., \u0026amp; Madhusudhan Reddy, G. (2023). Development of a novel welding technique with reduced heat-input by employing double-pulsed gas metal arc welding for aerospace grade 18% Ni maraging steel. \u003cem\u003eJournal of Manufacturing Science and Engineering\u003c/em\u003e, \u003cem\u003e145\u003c/em\u003e(2), 021002.\u003c/li\u003e\n\u003cli\u003eXu J, Wu Z, Niu J, Song Y, Liang C, Yang K, Chen Y, Liu Y. Effect of Laser Energy Density on the Microstructure and Microhardness of Inconel 718 Alloy Fabricated by Selective Laser Melting. Crystals. 2022; 12(9):1243.\u003c/li\u003e\n\u003cli\u003eNayak, S. K., Mishra, S. K., Jinoop, A. N., Paul, C. P., \u0026amp; Bindra, K. S. (2020). Experimental studies on laser additive manufacturing of Inconel-625 structures using powder bed fusion at 100 \u0026micro;m layer thickness. \u003cem\u003eJournal of Materials Engineering and Performance\u003c/em\u003e, \u003cem\u003e29\u003c/em\u003e, 7636-7647.\u003c/li\u003e\n\u003c/ol\u003e"},{"header":"Tables","content":"\u003cp\u003eTables 2 is available in the Supplementary Files section.\u003c/p\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"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":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"SLM, Inconel 718 powder, Machine learning, SEM, FEA, RNN","lastPublishedDoi":"10.21203/rs.3.rs-5922562/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-5922562/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eIn this paper multilayer fabrication of Inconel 718 using SLM, a widely used additive manufacturing technique was conducted. The quality of the fabricated tracks is dependent on laser power, scanning speed and beam diameter. The effects of these parameters during SLM process were investigated by numerical analysis as well as by experimentation to determine the optimized set of parameters. However, process optimization using experimentation and simulation is expensive as well as time consuming. Therefore, in addition, two machine learning models have been trained to predict the width and height of the fabricated tracks based on the parameters. By numerical analysis, the variation in temperature distribution was determined by varying these parameters and the effect of these parameters was validated experimentally. The effect of these parameters on the geometry, morphology of the track was analyzed by utilizing tools such as macroscope, SEM and EDAX. Among the two ML models, linear regression models demonstrated superior predictive accuracy with less error compared to RNN models.\u003c/p\u003e","manuscriptTitle":"Optimizing and Predicting Track Quality in Multilayer Laser Melting of Inconel 718: A Numerical, Experimental and Machine Learning Approach","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-02-07 15:27:32","doi":"10.21203/rs.3.rs-5922562/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"a2779917-2eb8-4ff5-8814-a88e869ef302","owner":[],"postedDate":"February 7th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-01T17:23:41+00:00","versionOfRecord":[],"versionCreatedAt":"2025-02-07 15:27:32","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-5922562","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-5922562","identity":"rs-5922562","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}