Online melt pool depth estimation during directed energy deposition using coaxial infrared camera, laser line scanner, and artificial neural network

Abstract

Melt pool monitoring techniques aid in the quality assurance and control of directed energy deposition (DED) additive manufacturing. Typically, the monitoring is based on the characterization of melt pool geometries, such as width, height, and depth. Among these, the melt pool depth cannot be measured directly. However, it indicates the distance from the deposited surface to the deepest point of the melt pool and is a key factor that determines the metallurgical bond between layers. In this study, an online melt pool depth estimation technique was developed for the DED process using a coaxial infrared (IR) camera, a laser line scanner, and an artificial neural network (ANN). Initially, the width and length of the melt pool at a particular position were measured using the coaxial IR camera. Simultaneously, the laser line scanner measured the build height and deposited track profile of the same position online. Features extracted from these measurements were used as inputs to the ANN model, and the melt pool depth was estimated online during multi-layer and multi-track printing. The performance of the proposed technique was verified considering multiple values of laser power, scanning speed, build height, and hatch spacing. The estimation results were compared with those obtained from optical microscopy inspection. The overall accuracy of the melt pool depth estimation was approximately 25.97 mu m. These results demonstrate the effectiveness and potential of the proposed online melt pool depth estimation technique for DED process monitoring.