基于混合半云模型的风速-功率曲线建模方法

Wind speed-power curve modeling method based on hybrid half-cloud model

风速-功率曲线的准确建模是风电机组出力态势评估和风电功率预测的关键基础之一。计及风电映射关系的不确定性及功率曲线的分布形态,提出一种基于混合半云模型的建模策略来实现对风功率数据固有和随机分布特征的挖掘和建模。引入最优组内云熵算法快速有效地剔除异常数据;采用逆向云发生器求取期望、熵与超熵数字特征来定量刻画风速-功率对应关系的不确定性,构建腰部数据的半云模型;通过X条件云发生器和正向云发生器分别求取腰部和上部数据的功率云滴,实现定性数字特征向定量数据的转换。以中国东北某大型风电场的实测数据为例,从数据质量、频率分布和风功率预测等维度分析混合半云模型,验证了所提方法的可行性。

Accurate wind speed-power curve modeling is one of the key basics for situation assessment of wind turbine output and wind power forecasting. Considering the uncertainty of wind power mapping relationship and the distribution pattern of power curve,a modeling strategy based on HHC(Hybrid Half-Cloud model) is proposed for mining and modeling of the inherent and random distribution characteristics of wind power data. The OSCE(Optimal Segmentation Cloud Entropy) algorithm is introduced to eliminate the abnormal data quickly and effectively. The backward cloud generator is adopted to obtain the digital characteristics of expectation,entropy and hyper-entropy for quantify the uncertainty of wind speed-power relationship,and a semi-cloud model of waist data is constructed. The power cloud droplets of waist and upper data are calculated by X condition cloud generator and forward cloud generator respectively,and the transformation from the qualitative features to quantitative data is realized. Taking the measured data of a large scale wind farm in Northeast China as an example,HHC is analyzed from the perspectives of data quality,frequency distribution and wind power forecasting,which verifies the feasibility of the proposed method.