The Experiment And Optimization Approaches Research On Selective Laser Sintering Process

Selective Laser Sintering (SLS) is one of the Rapid Prototyping (RP) process developed in the recent two decades. Compared with the other RP process, SLS has attracted extensive attention due to its fast building speed, wide range of materials, simple post-treatment et al. However, this developing tendency is limited due to the poor strength and accuracy of parts fabricated by the SLS process. Hence, carrying on the study of the SLS process possesses not only important theoretical values, but also realistic instructional significance. Using the neural network and genetic algorithms approaches, the SLS sintering process and its post-treatment process have been simulated and optimized on the basis of a series of experiments. According to the characteristics of SLS process, some profitable, creative methods and conclusions, and by which some practical problems have been solved, are presented in this dissertation. As to the study of the SLS sintering process, the main work and results in this dissertation are as following: 1.A neural network model was developed to predict the quality attributes for SLS. During this study, SLS process parameters were utilized as the input to the neural network, which include layer thickness(lt), hatch spacing(dh), laser power(W), scanning speed(v), surroundings temperature of work(Te), interval time(Ts) and scanning mode(F). The quality attributes of the SLS process fabricating part were used as the model's target, which include shrinkage, warpage, density and tensile strength. The principle and methods to determine the network parameters such as number of neuron in hidden layer, excitation function and the convergence accuracy have been analyzed in detail. The experimental investigation results show that the neural network model may be used to analyze the relationship between the process parameters and the quality attributes quantitatively. So it is suitable to apply the neural networks approach to study the SLS process. The model will allow us to produce the parts with desired quality attributes by selecting the appropriate parameter values prior to processing. 2.The Genetic algorithms (GAs) approach is used for optimizing the SLS processes on the basis of neural network modeling. According to the feature of SLS, we present an improved genetic algorithm of the initial population and population size determination based on space division. The optimization strategy which suits to the SLS process is determined by analyzing the performance of GAs with the various search strategy. Result show that the genetic algorithm can obtain the optimum solution of the expression described by neural network which used as a global optimization. The results thus obtained by Gas have been validated by actually building parts. The optimization method based on the hybrid of neural network and algorithms presented herein appears to be very effective for the processes that are difficult to be studied or experimentally due to the complex nonlinear characteristics. 3.Based on the simulation and optimization, a quality index of the sintering part is presented, through which the contribution of each process parameter is evaluated. The results demonstrate that the main process parameter effecting both on shrinkage and density is hatch spacing(dh); the process parameters effect on warpage is at the same level, and effect of the hatch spacing(dh) is large compared with the effect of other process parameters; the effect of the process parameters effect on tensile strength is difference, and the hatch spacing(dh) is most effective compared with other process parameters. 4.Appling the neural network model, the effect of each single and multi-parameter on sintering part neural network prediction performance have been studied. The prediction performance is discussed in detail for various process parameter ranges. The potential benefit from understanding the SLS process is twofold. First, the understanding of the process parameter effects on quality attributes of SLS process fabricating part will be beneficial to develop an intelligent process control for SLS equipment manufacturing. Second, the understanding of the SLS process will allow us to strengthen the process management and to improve the method of operation. In the aspect of SLS post-treatment process study, the main research work and results in this dissertation are as following: Based on a series of experiment, the influence of the sintering and post-treatment process parameters on the manufacture part performance such as shrinkage, surface roughness, density and tensile strength have been study by orthogonal experimental method. The results obtained can helps us to produce the parts with desired performance by selecting appropriate parameter values prior to processing.