A 3D printing process optimization method based on reinforcement learning

Abstract

The invention discloses a 3D printing process optimization method based on enhanced learning, which is carried out according to the following steps: a. Collect the printing parameters of several printed models to train the Q-learning model, and obtain the optimal printing of the printed model after training Parameters, the optimal printing parameters are stored in the training model obtained after training; b. Establish a 3D model of the model to be printed, cut the 3D model into k slices, and establish a 2D model of each slice; c. 2D model construction and 3D printing path planning of the printed model; d. Match the model to be printed with the printed model, and use the optimal printing parameters of the printed model with the highest matching degree with the model to be printed as the actual printing parameters of the model to be printed; e . Input the 3D printing path planning in step c and the actual printing parameters in step d into the 3D printing device for 3D printing of the model to be printed. The invention has the characteristics of improving the quality of 3D printing, shortening the 3D printing cycle and reducing the cost of 3D printing.

Description

technical field [0001] The invention relates to the technical field of wire-arc additive manufacturing, in particular to a 3D printing process optimization method based on reinforcement learning. Background technique [0002] In recent years, rapid manufacturing and prototyping technologies have developed rapidly, and new prototyping technologies represented by 3D printing have received widespread attention worldwide. The core process of this technology is to melt the metal material in the form of spherical powder or wire through a high-energy beam (including laser or electron beam, etc.) layer by layer with the assistance of numerical control equipment, and then deposit it to form large structural parts. Different from the traditional "removal" cutting method, this technology uses the "growth" concept to deposit layer by layer, which greatly improves the utilization rate of raw materials and the artistry of new products. At the same time, due to avoiding the design and pr...

AI Technical Summary

Problems solved by technology

Such as the selection of process routes and process parameters: On the one hand, since 3D printing is mostly used in small batches, complex structures, and high value-added products, using laser or electron beams as heat sources has the following disadvantages in terms of cost control: 1. For the laser heat source, its forming rate is slow, and the aluminum alloy has a low absorption rate of the laser; 2. For the electron beam heat source, the size of the vacuum furnace body limits the volume of the component

On the other hand, in the traditional 3D printing process, most of the process parameters depend on a large number of process tests to determine, resulting in long cycle, high cost, and few application scenarios.

Images