International Journal on Interactive Design and Manufacturing (IJIDeM) https://doi.org/10.1007/s12008-024-01848-5 ORIGINAL ARTICLE Optimization of process parameters and predicting surface finish of PLA in additive manufacturing—a neural network approach S. Panneer Selvan 1 · D. Elil Raja 2 · V. Muthukumar 3 · Tushar Sonar 4 Received: 11 December 2023 / Accepted: 10 April 2024 © The Author(s), under exclusive licence to Springer-Verlag France SAS, part of Springer Nature 2024 Abstract Additive manufacturing (AM), also known as 3D printing, has revolutionized the industrial sector by enabling the production of intricate geometries and specialized parts. However, achieving optimal surface finish and mechanical properties in AM poses challenges due to factors like material properties and machine characteristics. Accurately predicting surface finish is essential for process optimization and minimizing post-processing efforts. This abstract presents an innovative approach to predict surface finish and tensile strength simultaneously in AM. Leveraging advanced machine learning techniques, predictive models are developed using a comprehensive dataset of process parameters and corresponding measurements. The dataset is generated through systematic experimentation in the fused deposition modelling method, focusing on printing speed, layer thickness, and infill density. These models offer significant benefits to the industry, allowing manufacturers to optimize process parameters for desired surface finish and mechanical properties concurrently. By reducing reliance on trial-and-error approaches, they enhance efficiency, productivity, and part quality while lowering costs and accelerating product development cycles. Keywords Additive manufacturing · 3D printing · Surface finish · Tensile strength · Machine learning · Process parameters · Optimization 1 Introduction Additive manufacturing (AM) process develops items layer by layer from digital models utilising materials ranging from plastic to metal, in contrast to traditional manufacturing procedures that utilise subtractive processes like cutting or drilling. This approach offers numerous advantages, includ- ing the ability to produce highly complex and customized B D. Elil Raja elilraja76@gmail.com Tushar Sonar tushar.sonar77@gmail.com; sonart@susu.ru 1 Department of Mechanical Engineering, Rajalakshmi Engineering College, Thandalam, Chennai 602105, India 2 Department of Mechanical Engineering, St.Joseph’s Institute of Technology, OMR, Chennai 600119, India 3 Department of Mechanical Engineering, Saveetha Engineering College, Thandalam, Chennai 602105, India 4 Department of Welding Engineering, Institution of Engineering and Technology, South Ural State University (National Research University), Chelyabinsk 454080, Russia parts that would be difficult or impossible to make with con- ventional techniques [1]. The potential applications of AM are vast, including aerospace, healthcare, automotive, and consumer goods. Technologies for additive manufacturing come in a variety of forms, each with unique advantages and limitations. Some of the most common types are selective laser sintering, stereolithography fused deposition mod- elling, digital light processing and binder jetting. Among these, fused deposition modelling method is used in this work. The quality of the 3D printed part’s surface finish affects its aesthetics, functionality, and post-processing requirements, while mechanical properties, such as tensile strength, dictate its structural integrity and performance. The ability to predict surface finish and tensile strength accurately is of paramount importance for optimizing the manufacturing process and ensuring the reliability of printed components. Achieving the desired surface finish and tensile strength in AM involves a complex interplay of various factors [2]. These factors include material properties, such as compo- sition and rheological behaviour, process parameters like printing speed, layer thickness, and temperature, as well as 123