Prediction of tensile strength using machine learning algorithms in fused deposition modeling

Additive manufacturing provides unique advantages in creating complex parts, and among its methods, fused deposition modeling uses molten thermoplastic layers for diverse applications. Machine learning, an artificial intelligence technique, enhances additive manufacturing capabilities by enabling predictions and improving process control, designs, material properties, and production efficiency. The paper aims to utilize accurate machine learning-based algorithms for data analysis, and parameter optimization to predict the mechanical properties of the product. The results obtained through the support vector machine model integrating with the Tsai–Wu theory are in good agreement with experimental values observed at a print speed of 40 mm/s and a layer thickness of 0.2 mm. While the Levenberg–Marquardt algorithm in the artificial neural network has better prediction accuracy with mean absolute percentage errors of 4.22% and 3.53% for tensile and flexural strength, which is higher than the support vector machine. Furthermore, the research emphasizes the significant impact of printing speed on product quality by implementing analysis of variance with a percentage contribution ratio of ∼45%. This comparison allows manufacturers to make informed decisions and effectively optimize the additive manufacturing process. Machine learning algorithm utilization in additive manufacturing holds immense potential for elevating quality, efficiency, and reliability in this transformative industry.