Development of An Algorithm for The Inspection of Manufacturing Faults in Laser Powder Bed Fusion Processes

<span style="color: rgb(119, 119, 119); font-family: &quot;Open Sans&quot;, Arial, sans-serif; font-size: 14px;">Additive manufacturing (AM), particularly laser powder bed fusion processes, has revolutionized the production of complex geometries and high-performance components. However, defects such as porosity and incomplete fusion significantly affect the mechanical properties and reliability of manufactured parts. This study presents a novel algorithm to predict porosity in L-PBFP processes by analyzing critical production parameters, including laser power, scan speed, and layer thickness. Unlike traditional machine learning approaches requiring extensive data, the proposed algorithm models porosity mechanisms directly, allowing layer-by-layer prediction without the need for large datasets. Experimental validation was conducted using Ti-6Al-4V samples fabricated under varying process conditions, ensuring comprehensive evaluation. Results demonstrate the algorithm's efficiency in identifying porosity trends and its potential to optimize production parameters. The study emphasizes the importance of pre-production optimization to enhance product quality. This research advances L-PBFP technology by offering a fast, adaptable, and cost-effective method for defect prediction, paving the way for improved quality control in additive manufacturing processes.</span>