Layered and subregional control strategy based on model-free adaptive iterative learning for laser additive manufacturing process

Improving stability in process and the quality of manufactured parts is critical to laser additive manufacturing (LAM). Developing appropriate control methods is essential for this purpose. Layered control approaches are commonly utilized in current research on LAM process control, with a primary focus on the pattern of layer-by-layer deposition. However, these strategies might neglect the dynamic and evolving characteristics of the process along the scanning direction. Due to the complex nature of the LAM process, such layered control approaches could be inadequate in addressing intra-layer instability, potentially resulting in metallurgical defects. To overcome these limitations, we propose a layered-and-subregional model-free adaptive iterative learning control (LS-MFAILC) method. Considering the cyclic and the layer-by-layer processing principle of LAM, we develop a subregional control strategy and integrate it with a layered control framework. In addition, the model-free adaptive iterative learning control (MFAILC) algorithm is utilized as the controller due to its ability to control complex and time-varying systems with repetitive running characteristics. In the proposed method, each target layer to be printed in process is divided into multiple non-overlapping regions along the scanning direction, and the process parameters for each region are adaptively controlled based on historical information acquired from previous layers during the process. The temperature information of the melt pool is captured and extracted using a thermal imager integrated with the printing system. Experimental results and comparison with the layered control approach demonstrate the effectiveness and superiority of our method.