Additive manufacturing (also known as 3D printing) is widely used in automotive, aerospace, shipbuilding, medicine, and other industries. Despite the name, most commercial 3D printing systems currently in use operate in 2.5-D mode, in which materials are accumulated layer upon layer in planes with a fixed direction (usually the opposite direction of gravity). Although this approach has lower hardware complexity and software development costs, it is faced with staircase effects, the need for support structures, and even a lack of manufacturability. Multiaxis 3D printing methods remove the problem of support structure required by traditional methods, showing extremely high manufacturing flexibility. It also offers a wide range of applications in support-free 3D printing of complex structures. This article reviews multiaxis support-free 3D printing process planning methods. These methods are classified as overhang structure decomposition, skeletonization, constraint optimization, curved layer decomposition, and inner or outer volume decomposition based on their capacity to deal with complex models. Following that, the issues and challenges of multiaxis support-free 3D printing are examined in terms of surface quality, overhanging area manufacturability, precision control, path planning, and so on. Finally, in view of the current problems and challenges, the prospects for multiaxis support-free 3D printing are discussed for future development.

The method based on overhang structure decomposition was the first proposed process planning method for multiaxis support-free 3D printing. The core of this method is to distinguish between the core (or buildable) and overhang (or unbuildable) volumes, which requires the overhanging feature to have strong concave edges or loops. To improve the ability to deal with complex geometric parts even further, the researchers obtained the centroid axis, or skeleton, from the geometric information of the input model by following the intrinsic characteristics of the part, and then used the skeleton to guide the decomposition into subvolumes without supports. This method can implement the support-free fabrication of multibranched or tree-like structures (as shown in Fig. 11), but it cannot effectively handle models without obvious skeleton features, implying that depending just on the skeleton for guided slicing is insufficient. Combining constrained optimization methods to minimize the area of the support structure yields a better decomposition result. Following that, a variety of methods, including the ant colony algorithm, beam-guided search algorithm, and downward flooding search algorithm, were proposed to find the optimal solution. In Table 2, the characteristics of constraint-based optimization methods are compared. However, to manufacture on a 3+ 2 axis platform, most of the above methods use a plane as both a dividing plane and a base plane. This planner layer-based constraint imposes constraints on the fabrication of more complex parts, and new approaches are attempted to fully utilize the flexibility provided by multiple degrees of freedom. The curved layer decomposition method attempts to divide the volume into a series of roughly equal-thickness surfaces while satisfying support-free and manufacturability constraints. There is a possibility of nozzle collision and local embedding if the resulting surface layer is concave. As a result, obtaining the convexly curved layers becomes the focal point of this method. Recently, an ellipsoid-based curved layer decomposition algorithm was proposed. Due to the ellipsoid’s convexity, the problem of nozzles’ local embedding was successfully avoided. However, this method is overly convoluted. To achieve a balance between algorithm, control complexity, and manufacturing efficiency, an inner or outer volume decomposition method was proposed, in which 5- and 2.5-axis depositions were applied externally and internally, respectively, to obtain a denser internal entity, which was important for metal parts. Table 4 summarizes the characteristics of the proposed multiaxis support-free 3D printing process planning method and their impact on the manufacturing process.

This paper also summarizes some current issues and challenges. First, previous work has focused on path generation with suboptimal surface quality under support-free and collision-free constraints. To improve surface quality, methods such as microcurved slicing, helical slicing, conformal printing, remelting, and hybrid fabrication have recently been proposed. Second, to improve the overhang area’s manufacturability, methods, such as changing Z-direction increments and tilting the nozzle, are used to expand the maximum permitted overhang angle. Third, while most of the current multiaxis 3D printing systems are in an open-loop state, the control quantities in the printing process are usually coupled with each other. Researchers try to smooth the uneven top surface caused by the variable height deposition strategy with the PI controllers. Finally, as illustrated in Fig. 24, the path generation is further optimized to meet the collision requirements while avoiding singularities as much as possible.

To summarize, the process planning methods of multiaxis support-free 3D printing for complex structures has yielded intriguing results, including the ability to decompose and print volumes with nonsharp edges, but it is still in its early stages. The following aspects are expected to be prioritized in the development of multiaxis support-free 3D printing methods. The first is more widespread process applications. At the moment, most methods are concentrated on the FDM process. To realize support-free printing of complex metal components, the process requirements and metal additive manufacturing characteristics must be further combined. The second is improved precision and performance printing. Feedback control is used to improve the robot’s positioning accuracy and to further optimize the motion trajectory to improve manufacturing accuracy. Simultaneously, by introducing requirements, such as mechanical properties into process planning, the mechanical properties of the parts will be improved even further. Additionally, functionally gradient structure printing of complex structures is a potential research topic. Another option is to use multiaxis hybrid manufacturing. The CNC machine has a high level of precision and efficiency in manufacturing. Combining with a CNC machine can improve manufacturing flexibility even further. There will also be a more open community. Hardware and software compatibility, as well as device openness, have an impact on multiaxis support-free process planning. However, some open source motion and process planning methods have emerged, and more opportunities are expected in the future.