风力发电中风电场风速优化预测仿真  

Simulation of Optimized Wind Speed Forecast in Wind Power Generation

作　　者：荆波 黄宏光[1] 钟俊[1] JING Bo HUANG Hong - guang ZHONG Jun(School of Electrical Engineering and Information,Sichuan University,Chengdu Sichuan 610065, China)

机构地区：[1]四川大学电气信息学院,四川成都610065

出　　处：《计算机仿真》2017年第10期418-423,共6页Computer Simulation

基　　金：四川省科技厅支撑项目(2016GZ0145)

摘　　要：风电场风速预测是风力发电中的重要组成部分,提高预测精度可以保障电力系统的可靠性。风速受地形和天气等因素的影响,表现出随机性、波动性和非平稳性,导致传统的风速预测模型不易定阶,且只能反映风速序列的一部分信息,预测精度较低。为有效提高预测精度,提出一种新的优化预测方法。首先采用求和模型消除原始风速序列中的非平稳性,其次使用扩展自相关函数准确定阶,最后利用自回归条件异方差模型进一步提取残差序列中的有用信息。运用上述优化模型对某地的风速和风电功率进行预测,并对比传统模型的预测结果。结果表明,优化预测模型充分提取了风速序列中的信息,能够更好跟踪风速骤变,提高了预测精度。Wind speed forecast plays a very important role in wind power generation. The reliability of power system is guaranteed with high forecasting accuracy. Wind speed series is random,fluctuant and non -stationary because of terrain and weather. Therefore, there are several problems in the traditional wind speed forecasting model, such as the complexity of ordering, low data utilization rate and low accuracy. An optimized forecasting model is proposed in the paper, to improve the forecasting accuracy. Firstly, an integrated model was introduced to eliminate non - stationa- ry of the original wind speed series. Secondly, an extended autocorrelation function was used for ordering. Finally, an autoregression conditional heteroskedasticity model was used to extract useful information of residual. Wind speed and power were predicted by the above optimized model. Compared with the traditional model, the optimized forecasting model can extract information adequately, track the drastic change of wind speed and improve the forecasting accuracy.

关 键 词：风速预测 自回归滑动平均 扩展自相关 异方差 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]