Deposition Modeling And Path Planning For Cold Spray Additive Manufacturing

Additive manufacturing technology is often used for custom design and manufacturing of complex-shaped parts,and has gradually become one of the key technologies in current molding and manufacturing.Traditional additive manufacturing technologies slice 3D part models in parallel layers on software.Layer-by-layer accumulation is formed along the set direction to achieve overall molding printing.However,there are still bottlenecks in this technology,such as the existence of solidification tissue defects inside the finished parts,which cannot fully meet the preparation requirements of aluminum,magnesium and other light metals.Cold Spraying Additive Manufacturing(CSAM)technology has the advantages of low processing temperature,simple operation,and low thermal impact on materials.It is increasingly promising in the field of additive manufacturing.However,compared with the traditional thermal spray additive manufacturing technology,the cold spraying additive manufacturing technology is still in the development stage.The problems of profile shape prediction and geometry control in the deposition process and the analysis and design of the gun path in the printing process still need to be studied in depth.In this paper,a neural network model improved by a meta-heuristic algorithm is used to predict and control the geometry of the deposition profile in the cold spray additive manufacturing process.A hierarchical slicing method for printed part models is explored.This is also used to analyze and design the print path planning method in detail.The main work of this paper includes:(1)Starting from a theoretical study,the forming mechanism of the cold spray manufacturing process and the influencing factors of the deposition profile are analyzed.The model design and experimental verification of the cold spray deposition layer profile are carried out.A cold spray additive manufacturing deposition distribution model based on the theory related to the profile distribution is constructed,and the equation for the measurement of the deposition layer thickness is obtained.An improved Marine Predator Algorithm(MPA)combined with an Artificial Neural Network(ANN)was used.The input layer parameters of the neural network were screened.A modified Marine Predator Algorithm-Artificial Neural Network(MPA-ANN)model was designed to predict the deposition profile of cold spray additive manufacturing.A comparison is made with other prediction models.(2)The detailed process of slicing STL(Standard Template Library)files for additive manufacturing parts is analyzed.The characteristics of existing layered slicing methods are discussed.According to the neighboring triangle surface slice search method,the contour line orientation in the slice is analyzed and judged to ensure the complete search of the outer surface contour curve.The layering thickness is determined by minimizing the outer contour line error,and an adaptive layering method based on the ratio of contour line deviation is derived.(3)Combined with the characteristics of the additive manufacturing process,the printing area of each layer is determined based on the contour lines obtained after slicing.A cold spray additive manufacturing print path planning model was constructed to obtain a mathematical description of the shortest spray path for the contours inside and outside the print area.The picture information of the print path is learned and predicted by combining deep learning methods.It is compared with a linear model to verify the effectiveness of using convolutional neural networks to predict the shortest print path for additive manufacturing.Through the above work,the prediction capability of the proposed neural network algorithm for cold spray additive manufacturing deposition model is verified.The planning goal of the shortest print trajectory for cold spray additive manufacturing is proposed.The overall printing time is reduced and the printing efficiency is improved.The feasibility of using convolutional neural networks to solve the additive manufacturing print path problem is demonstrated.