The Research Of Barium Strontium Titanate Artificial Neural Network Expert System

Along with the development of computer technology and the improvement of data processing algorithm, the advanced research of artificial neural network in materials science has been made. Barium strontium titanate with the high insulation resistance, the low dielectric loss (high frequency and low frequency) and high capacitance temperature stability, has been researched by many people.Based on the grasping of the artificial neural network, the performance of barium strontium titanate ceramics and the database development technology, the artificial neural network expert system for the barium strontium titanate was designed by using the MS SQL SER.VER2005to store data and by using the adaptive step-size BP neural network algorithm to train data. This system contains a friendly man-machine interface, which is used for saving, editing and deleting experiment data into database. And then with the basic knowledge of materials research, the BP neural network was built after processing these data. Finally the training results are being used to forecast the dielectric properties and the curie temperature and the lattice constant of barium strontium titanate, where these barium strontium titanate samples were prepared under different process conditions (sintering temperature, sintering time, pre-sintering temperature, pre-sintering time, molding pressure, substrate, annealing temperature, the amount of Ba2+, the amount of Sr2+, the amount of Ti4+) and tested at different frequency, temperatures and electric field strength, respectively.