The Research Of Silica Dioxide Composite Materials Artificial Neural Network Expert System

With small size, high purity, low density, high specific surface area, gooddispersion properties, nano-silica dioxide has an irreplaceable role in many scientificfields. The polymer nanocomposites not only can play the excellent performance ofsilica, but also show the advantages of polymer materials. Therefore, in recent years,silica dioxide composite materials have been a hot research topic at home and abroad.In this paper, based on the application of Visual Studio2005and SQL Server2005database, a silica dioxide composite artificial neural network expert system isestablished. This is in reference to the large number of foreign literature on nano-silicadioxide composites, collecting, summarizing and analyzing the experimental data in thefield. Through the human-computer interface and use ADO.NET access to the database,System realize the data query, modify, delete, and add the database of silicon dioxidecomposite material properties such as mechanical, thermal and electrical, then takingthese data as training samples, with BP neural network algorithm, system make networktraining. In the training process, adjust the network parameters for optimal training, testrunning results show that the learning rate is0.1~0.9, the number of hidden layer nodesis8or9.Finally, by using the weight matrix saved in the process of training, system predictthe properties of silica composite materials prepared under different process parameters,such as the particle size, density, dielectric properties, tensile strength and elongation, soas to provide reference for the future research on silica composite materials.