A novel process planning method of 3+2-axis additive manufacturing for aero-engine blade based on machine learning

Additive manufacturing (AM) is an emergingly technology in aerospace such as aero-engine blade fabrication, which has benefits in complex shape creation with little post processing required. In this paper, a machine learning algorithm is proposed for powder-saving and support-free process planning in multi-axis metal AM, improving the printing efficiency and the surface quality of printed blade. Firstly, a self-adaptive spectral clustering algorithm is developed to carry out two functions: one is to decompose the blade into sub-blocks in a global view; the other one is to automatically obtain the optimal clustering number, addressing the contradiction issue between printing efficiency and decomposition performance. Secondly, the global constraint formula and the normalized area weight are introduced to obtain main printing orientations (MPOs). Each sub-block can be built along the corresponding MPO with high-quality surface, free support, and low powder leakage. A sample blade is built on the 3 + 2 axis laser metal deposition (LMD) machine to validate the feasibility of the proposed method. Experimental results indicate that the proposed method has advantages of less powder consumption, higher decomposition performance and printing efficiency compared to the existed method.