摘要
现有风电功率预测多数为确定性预测,而由于风电功率具有随机性和波动性,确定性预测在不确定性条件下难以为系统调度决策提供有效信息,概率性预测能够提供预测的不确定性信息。提出一种基于小波变换和模糊自适应共振映射的新型确定性预测方法,利用数值天气预报及风电功率历史数据进行确定性预测。同时基于分位数回归分析,并以置信度、锐度、技术评分为指标,对确定性预测结果进行概率性评估。以某风电场为例,给出了确定性预测值及概率性评估,验证了所提方法的有效性。
Most of the existing wind power forecasting is deterministic,but the deterministic forecasting is difficult to provide effective information for the system scheduling decision under uncertainty conditions due to stochastic and volatility of the wind power,and the probability forecasting can provide uncertainty information.A new deterministic forecasting method is proposed based on wavelet transform and fuzzy adaptive resonance mapping by using the historical data of numerical weather forecast and wind power.At the same time,based on quantile regression analysis,deterministic prediction results are probabilistically assessed by regarding confidence,the sharpness and the skill score as the index.Taking a wind electric field as an example,the deterministic forecast value and the probability assessment are given,which verify the validity of the proposed method.
出处
《现代电力》
北大核心
2016年第5期80-86,共7页
Modern Electric Power
关键词
概率性预测
分位数回归
模糊自适应共振映射
小波变换
风电功率预测
probabilistic forecasting
quantile regression
fuzzy adaptive resonance theory map
wavelet transform
wind power forecasting
