| In this paper,Visual Basic(VB)program is used to establish the optimizing software system of experimental design method,thereby realizing the selection of the experimental design methods based on a certain special purpose.The validity of the software is confirmed according to the comparison between the results of full factorial design method and that of the experimental method selected by the developed software.Moreover,the data mining and analysis were performed on the typical cases related to material research by different data processing methods.The aim of this project is to develop a convenient and practical tool for the engineering technicians,by which,the optimization of the experimental designs and data analysis methods could be achieved.The optimizing software of experimental design methods developed in the present paper includes the current three common methods:the full factorial design,the orthogonal experimental design and the uniform design.The users can select the appropriate experimental design method using this software according to the specific requirements and the evaluation range of the experiments.Compressive tests of SiC/AI composite are designed by using full factorial design method and the method selected by the developed optimizing software,respectively.Compressive tests are then simulated by finite element analysis based on the models established by SolidWorks.The results indicate the validity of the experimental design optimization software.For data processing method,the following conclusions can be derived by instance analysis from the present work:regression analysis is excellent in solving linear problems,which can accurately describe the correlation between experimental factors and resulting data;The gray correlation method can be used to sort the correlation between the experimental factors and the target factors,showing the effect of the different experimental factors on the target factors;The artificial neural network has a strong ability to deal with non-linear analysis and is suitable for high-throughput experiments with a large number of experimental data,which will be one of the most important calculating prediction tools in material genetic engineering. |
