基于差分自回归滑动平均模型的风电场短期风速预测

Wind Speed Short-term Forecast for Wind Farms Based on ARIMA Model

目前,风力发电的并网规模越来越大;但是鉴于风力发电特有的间歇性和随机性,难免会对电力系统的稳定运行和电能质量造成巨大影响;也就限制了风电的发展速度与规模。对风力发电场的风速进行中、长、短期的预测可以在一定程度上有效地解决该问题。依据风速序列的自相关性以及时序性,提出了一种基于时间序列分析的风电场短期风速预测ARIMA模型。重点讨论了建模的过程、模型的识别、模型的定阶和模型参数的估计。最后结合风电场实际,对比于持续法预测,给出了相应的预测结果和误差分析,验证了所提出的ARIMA模型用于风电场风速预测的可行性。

The size of the wind powenetwork igrowing apresent. In view of the intermittenand stochasticharacteristicof wind powegeneration, which will inevitably have tremendouimpacon the stable operation of the powesystem and powequality, also limitthe speed and scale of the developmenof wind power. The long, short-term wind speed forecastof the wind farm can be effective to solve the problem to some extent. In view of the wind speed sequence with timing and correlation, wind speed forecasmodel, ARIMA, based on time serieanal- ysis, ipresented. The modeling process, pattern recognition, model ordedetermination and parameteestimation are focused on. Finally, the forecasresultand erroanalysiare given with the contrasto the continuoumethod forecast, the resultverify the feasibility of the proposed ARIMmodel fowind speed forecasting.