3D printing path planning method for complex thin-walled structure object based on reinforcement learning

Abstract

The invention belongs to the field of artificial intelligence and 3D printing and relates to a 3D printing path planning method for a complex thin-walled structure object based on reinforcement learning. The method comprises the following steps: firstly, establishing a simulation environment for path planning; then establishing a return matrix R based on a Q-learning algorithm in reinforcement learning, and generating a status-action matrix Q of the return matrix R; analyzing data of the status-action matrix Q to obtain a printed path planning simulation result; and finally, combining parameter of an actual printer with the simulation result to finish an actual printing process. By means of the Q-learning algorithm, the 3D printing path is learned intelligently. By means of learning and training, a computer can find a proper printed path intelligently, so that the printing efficiency can be improved greatly. The formed effect is also obviously better than that of a conventional path planning algorithm. The method provided by the invention is of universality to a complex thin-walled pattern.

Description

technical field [0001] The invention belongs to the field of artificial intelligence and 3D printing, and relates to a 3D printing path planning method for complex thin-walled structural objects based on reinforcement learning. Background technique [0002] 3D printing technology, also known as additive manufacturing technology, was originally used in the field of industrial design and mold manufacturing to manufacture part mold models, and then used in direct manufacturing of products. After nearly ten years of rapid development, 3D printing technology has been widely used in medical, aerospace, education and other fields. At present, the most widely used method is the layered 3D printing method. By layering the target object model and accumulating printing layer by layer from bottom to top, the target entity is finally obtained. [0003] Among them, after the target object model is layered, the movement path planning of the print head for each layer is an important step i...

AI Technical Summary

Problems solved by technology

However, for objects with thin-walled external and internal structures (often used for lightweight industrial production), the existing traditional path planning methods have certain limitations. If the stress is too concentrated, the heat dissipation effect of the material will be reduced, which will directly lead to deformation, cracks and other phenomena in the product.

The model contour printed by the offset contour algorithm is easy to be disconnected due to the large printing span, forming a disconnected "pocket", and the number of printing interruptions is also large, and the phenomenon of "drawing wire" will appear at the end point, that is, redundant curing will occur glitch

Images