基于SVM模型的风电功率预测

为解决因风电机组功率波动产生的调度问题,运用支持向量机法对风电机组功率的输出进行实时预测,结果表明单台机组预测的均方根误差为2.16%,相关系数为77.605 4%,58台机组预测的均方误差为0.7%,相关系数为90.321 4%。说明了风机机组汇聚得越多,机组系统越稳定。最后还探索了进一步提高SVM预测精度的方法。

The Forecasting of Wind Power Based on SVM

To solve the problem of electricity scheduling which is caused by power fluctuation of wind turbines,this article focuses on the real-time forecast on power output of wind turbines with the use of support vector machine.The result shows that the MSE of single turbine forecast is 0.02,the related coefficient is 77.605 4%,the MSE of fifty-eight turbines forecast is 0.007,the related coefficient is 90.321 4%.It shows the wind turbine convergence is beneficial to the stable of turbine system.It also makes an exploration on further improvement of forecasting accuracy.