Research On Process Parameter Optimization And Precision Prediction Model For Selective Laser Sintering

Selective Laser Sintering is one of the branches of additive manufacturing technology.It has become the focus of current manufacturing companies,due to its wide selection of materials,high manufacturing flexibility,no need for construction support,and high material utilization.At present,there is no specific and scientific system guidance for the formulation of process parameters for SLS molded parts.In most cases,it is determined according to experience or using traditional learning algorithms.However,the results can’t meet the molding precision of requirements of SLS molded parts.In this paper,SLS molded parts are taken as research objects,aiming at reducing the warpage deformation and cylindricity error of molded parts.Using the combination of finite element simulation and experimental research,the processing parameters affecting printing precision are studied and analyzed.A corresponding optimization prediction model was established to guide the design and formulation of SLS molded parts processing parameters.Firstly,according to the SLS sintering process,the basic principles and characteristics of SLS are introduced.The main factors affecting the precision of SLS molded parts are analyzed.Then,using ANSYS software,the three-dimensional transients in the sintering process of SLS molded parts are constructed.The finite element model realizes the moving application of Gaussian heat source using APDL parametric language under the conditions of thermal conduction,thermal convection,thermal radiation and thermal property parameters with temperature changes.The temperature field and strain of SLS sintering process were analyzed.The warpage deformation data of SLS molded parts were obtained.The simulation results were consistent with the experimental results,which proved the feasibility of obtaining data by simulation.Secondly,the effects of different laser power,scanning speed,scanning pitch and layer thickness on the precision of the molded parts were studied by orthogonal test.The orthogonal table design experiment was used to obtain the corresponding experimental data.The variance analysis method was used to analyze the warpage deformation and cylindricity error of the molded parts.Selecting the optimal combination of printing parameters with the goal of minimum warpage and cylindricity error.Finally,based on the above simulation and experimental data,BP neural network optimized by Particle Swarm Optimization established a precision prediction model for SLS sintered parts to solve the problem difficult to select the precision and process parameters of the molded parts in SLS and the defects of BP neural network itself.The optimal solution of PSO is used as the initial weight threshold of BP neural network algorithm.The optimized BP neural network prediction model is established by using MATLAB.Based on the laser power,scanning speed,scanning pitch and layer thickness,the model aims at the warpage deformation and cylindricity error of the molded part,and constructs the nonlinear relationship between the process parameters.Then compared with predicted result and traditional BP neural network,it shows that the PSO-BP neural network model has good global search ability and convergence,and the accuracy prediction is more accurate,which has practical guidance for the rational selection of processing parameters.