Optimization of polylactic acid 3D printing parameters based on support vector regression and cuckoo search

Abstract

Abstract3D printing, also known as additive manufacturing (AM), is evolving from a rapid prototyping tool to a pillar of the Industry 4.0 revolution and widely used in various industries, since it can quickly and efficiently create workpieces with complex structures and integrated functions. After analyzing the 3D printing principle of fused deposition modeling (FDM), this paper proposes a deformation control‐oriented optimization method for process parameters of FDM of the polylactic acid (PLA) materials, based on support vector regression (SVR) and cuckoo search (CS). First, FDM printing principle and its main process parameters were analyzed. Two key parameters, printing temperature and printing speed, were selected for research, with the maximum shrinkage deformation as the workpiece contour accuracy index. Combining Latin Hypercube Sampling (LHS) with Finite Element Analysis (FEA), finite sample sets were generated. Second, learning from the sample data, the SVR surrogate model was built to predict the workpiece contour accuracy, and the nonlinear relationship between FDM process parameters and PLA shrinkage deformation was obtained. Finally, taking the combination of printing temperature and speed as the design variable and minimizing the maximum shrinkage deformation as the optimization objective, the FDM process parameters optimization model was established, and CS was used to search for the optimal parameters combination. Experimental comparison showed that the FEA results and SVR model in this method are correct and effective, and the PLA workpiece shrinkage deformation is smaller under the optimal parameters combination, which effectively improves the shape accuracy of FDM workpieces.Highlights Combining support vector regression (SVR) and cuckoo search (CS) can optimize polymer 3D printing process parameters. Effective finite element model was built for polymer 3D printing. SVR model established by simulation data can predict workpiece deformation. The optimal process parameters were found using CS algorithm.