Comprehensive Research On Intellective Deformation Analysis Method

With the development of the science and computer technology, the new methods of the deformation data processing come forth increasingly, the artificial intelligence method has become the focus of the deformation data processing. Compare to the tradition data processing method, the artificial intelligence method can learn itself, adapt itself, adjust error and forecast accuracy and so on. It plays a more importance role in the deformation analysis and forecasting.This paper research on combining some method using in the deformation analysis and forecasting, bases on the merit of the artificial intelligence method. The main contents including:1,It researches the method of the deformation measurement data processing. This paper uses the dyadic wavelet transformation method to remove noise of the raw data and has used mean square error variance,correlation coefficient,signal-noise ratio and flatness as four evaluation indexes and then put them between 0 and 1 to determine the decomposing progression using comprehensive analyzing methods.2,It researches the deformation forcasting precision of the Wavelet Neural Network. Based on the principle that the Neural Network sometimes has local minimum and slow convergence speed, the article has used Genetic Algorithms and Wavelet Neural Network activiation function to form Genetic Wavelet Neural Network and has forecasted deformation, and at last get more accurate forecast values.3,Uses the Genetic Wavelet Neural Network to forcast deformation. Based on the principle that the Neural Network sometimes has local minimum and slow convergence speed,the article has used the Genetic Algorithms to seek all the data in order to calculate the best result and combine the Genetic Algorithms and Wavelet Neural Network to form the Genetic Wavelet Neural Network and a more accurate forecast method using BP Neural Network and Wavelet Neural Network has been realized.4,Uses the method's forecast stability and validity to evaluate it . Analysed the deformation forecast stability and validity of BP Neural Network, Wavelet Neural Network and Genetic Wavelet Neural Network, the forecast ability and validity of Genetic Wavelet Neutral Network has been proved further.5,Analyses the problems of building the deformation survey system. The problems of how can the sensors join in the system, the data storage and management, data fusion and analysis.