Abstract: Prediction of offshore wind power is a prerequisite for the stable operation of large-scale wind farms.The accuracy of wind power prediction plays an important role in improving the quality and consistency of the power grid.To improve the wind power prediction error,the Auto Regressive Integrated Moving Average(ARIMA)and the improved k-Nearest Neighbor(kNN)method were used to predict the offshore wind power.Results show that wind power could be predicted by different prediction techniques according to different characteristics,the prediction error was less than 20%,and the prediction accuracy could be improved by improving the prediction method.Finally,the improved two prediction methods were verified by a specific example,and the results were compared to verify the effectiveness of the algorithm.

Key words: wind power, auto regressive integrated moving average, improvement k-nearest neighbor, prediction model

摘要： 海上风电功率预测是大型风电场并网稳定运行的先决条件,风电功率预测的准确性对提高电网的质量和一致性具有重要作用。为提高风电功率预测误差,采用自回归移动平均(ARIMA)和改进的k-最近邻(kNN)两种预测方法对海上风电功率进行预测,结果表明,风电功率根据不同的特性可以用不同的预测方法进行预测,预测误差小于20％,通过改进预测方法,能够提高预测精度。最后,通过具体算例对改进的两种预测方法进行了验证,并将结果进行了对比,验证了两种算法在不同预测时间范围的有效性。

关键词: 风电功率, 自回归移动平均, 改进的k-最近邻, 预测模型