基于ARMA模型的短期风功率预测研究

ARMA Model Based Study on Short-time Wind Power Predication

针对风功率预报中出现的资料获取困难、预报精度差等问题,提出采用基于时间序列数据的ARMA模型,并重点对ARMA模型进行识别和参数进行估计。在一定范围内,枚举输入AIC值,并采用大量数据进行模拟,同时采用MAE、NMAE和NRMSE三种指标对模拟结果进行评价,得到了适合于风功率预报的ARMA模型。同时将模型用于预报,发现预报结果精度比较高,表明ARMA模型有较好的实用性。

ARMA model based on time series is proposed, its identification and parameters are predicated especially aiming at difficulty in data acquisition in wind power predication. In certain scope, A/C values are exemplified and input and data in large quantity are utilized o simulate. Simultaneously, three indexes of MAE,NMAE and NRMSE are applied to assess the simulating results. Finally, ARMA mod-el suitable for the wind power prediction is obtained. Furthermore, the model is applied for prediction. It is found that the prediction pre-cision is higher. This present that ARMA model is quite practical.