Adaptive process control for selective laser melting to compensate for pore formation

点击翻译按钮获取中文摘要

With the increasing customization of products, additive manufacturing plays a crucial role while time-consuming, cost-intensive toolmaking becomes obsolete and the direct production of complex geometries is made possible. Powder bed-based processes, like selective laser melting (SLM/PBF-LB/M), are an industrially relevant and widespread form of additive manufacturing. In this process, the energy input in the form of a laser beam melts a layer of powder that turns into a solid compound during the subsequent solidification. This process is repeated layer-by-layer until a three-dimensional component has been formed. For this, digital process preparations extract data of the layers through a virtual 3D model, and a path planning for the laser path is made to provide the additive manufacturing machine with the relevant geometric information of a component. However, the complex interaction between manufacturing parameters, powder and machine characteristics, ambient conditions, and operator-dependent factors during the build phase leads in part to unforeseeable process results. These become evident in the form of faulty component properties, which immensely limit the profitability of the process. Current trial-and-error approaches for the definition of suitable manufacturing parameters for stable process windows reduce the time advantage of direct additive manufacturing compared to conventional methods. Reaching a maximum material density of the component of up to 100 % to achieve optimal material properties is a significant quality feature in additive manufacturing. However, the formation of process-related pores and lack of fusion, each of which worsens the component density as well as the static and dynamic strength in a different way, is a fundamental problem in reaching this goal. At the same time, these different process errors partly depend on the energy input in the process in a complementary way. The consequence of the complex interaction in the process and the fact that control systems used today do not provide real-time manual intervention and interfaces is that a dynamic parameter adjustment of current machine controls is not used in additive manufacturing. Therefore, process parameters, like laser power and scanning speed, are usually kept constant. In order to counteract this limitation, the research of highly dynamic control and regulation concepts in consideration of transient process states is essential. Hence, it is the goal of this research project to develop a self-adapting control system, which determines suitable manufacturing parameters subject to relevant process and measurement factors. Thereby, optimal process states can be ensured and the resulting component quality can be increased.