Offshore Wind Speed Load Predicting Based on GA-SVM

The accurate measurement of the effective wind speed is the basis of the wind generation system operation. The effective wind speed cannot be directly measured accurately. Because offshore wind farm is a three-dimensional time-varying environment, the wind speed distribution is different on the plane of blade rotation in the wind generation system. This paper uses support vector machine with genetic algorithm (GA-SVM) to estimate the effective wind speed. The simulation results show that this model has high prediction accuracy, which the mean error of the training data can reach 0.5689 m/s. This model can improve the control capability of the wind generation system, when the wind speed changes in a wide range.

[1]  Yong Liu,et al.  Research on construction system of large scale non-grid-connected wind power offshore wind farms , 2010, 2010 World Non-Grid-Connected Wind Power and Energy Conference.

[2]  Hua Li,et al.  An algorithm of soft fault diagnosis for analog circuit based on the optimized SVM by GA , 2009, 2009 9th International Conference on Electronic Measurement & Instruments.

[3]  He Chuan.  Simulation of wind speed estimation for MPPT control , 2010 .

[4]  Wei Shen,et al.  A soft sensor modeling approach using support vector machines , 2003, Proceedings of the 2003 American Control Conference, 2003..

[5]  Bimal K. Bose,et al.  Fuzzy logic based intelligent control of a variable speed cage machine wind generation system , 1995 .

[6]  Ming-Syan Chen,et al.  On the Design and Analysis of the Privacy-Preserving SVM Classifier , 2011, IEEE Transactions on Knowledge and Data Engineering.

[7]  Liu Yingming,et al.  Research and simulation of direct drive wind turbine VSCF characteristic , 2008, 2008 IEEE International Conference on Automation and Logistics.