Study And Design On Data Acquisition And Prediction For Systems Of Brushless Doubly-fed Wind Turbine

With the prompt development of the world economy, the problem of the resources, environment and the power sources is getting worse and worse. Countries are all working at development and making use of the new resources, which is renewable, unpolluted and low cost. The study and development of the wind power system especially wind power generation system is a hot topic in the new energy territory. Wind Power Generation System with Brushless Doubly-Fed Generator can realize variable speed by controlling variable pitch of the double-driver wind generators, and this advanced technology has been the study point.The variable pitch system of the brushless doubly-fed energy is analyzed, and Support Vector Regression, SVR is for shot, algorithm is put forward to predict pitch angel, and System of the high speed DSP controlling Data Acquisition is designed in this paper.SVR algorithm is used, including the regression analysis and the choosing of proper kernel function, and the best pitch angle is designed in every wind speed for the limited samples. The MATLAB is used to simulate and checkout the sample data of training and prediction. The result indicates that the accuracy of this algorithms is better than the past neural network algorithms, for Generalization ability is getting better, improving the controlling accuracy and the efficiency of the whole system.DSP is used to collect data of wind speed, and the core unit is TMS320F28335 of TI. Through wind speed sensor, the wind speed signal is analog, witch is input to ADC, and in witch the digital signal that is given to DSP is translated. Peripheral circuit include design of power supply circuit, reset circuit, ADC part of the DSP, storage part, RS-232 interface part and interconnect portion, and several Software design processes of the main circuit are designed in this paper, initially reach the design purpose.