Multi-object optimal design of rapid prototyping based on uniform experiment

Abstract Experiment of rapid prototyping (RP) is a multi-level and multi-factor problem with long process cycle. As powder-based RP method, three dimensional printing (3DP), and selective laser sintering (SLS) are the most prominent RP methods for their flexibility in material selecting. The research of 3DP process shows that much experiments work can be greatly decreased by the method of uniform design with restricted mixtures for RP products' performance to its forming material. Then, according to backward regression methods, the binomial and inverse terms nonlinear regression equations were set up by statistic analysis and computation. Together with evaluation function method, multi-object optimization model for performance of products to the powder mixtures was developed, and the optimization problem was solved by software MATLAB. After that, the optimum results were solved and tested by experiment with the same condition as before. Differences between the calculation results and testing results, which including product's density, compression strength, surface evaluation, and deformation in three dimensions, are less than 10%, and have the same error direction. Uniform design method for regression and optimization was proved to be an excellent method to obtain optimum results for multi-level and multi-factor experiments of RP and other process which have a long processing cycle.