考虑负荷特性的母线负荷异常数据修正方法

介绍了异常数据的种类、特点和识别母线负荷异常数据的方法根据母线负荷特性的影响因素生成特征向量,通过灰色关联度来选取相似参考样本,以相似样本为基础采用模糊C均值聚类获得聚类中心曲线,并根据识别出的异常数据种类,结合聚类中心曲线与修正算法进行异常数据修正实例证明该修正方法对不同的异常数据类型都能有较好的修正效果,确保了母线数据的完整性,提高了母线负荷预测的工作效率和预测精度

Correction Method for Abnormal Data of Bus Load Considering Load Characteristics

This article introduces the types and characteristics of abnormal data as well as the method to recognize the bus with abnormal data.Firstly,a feature vector is generated according to the factors affecting bus load characteristic,and the similar reference sample is selected through gray correlation degree.Then the clustering center curve is obtained by using the fuzzy C-mean clustering based on the similar samples.Finally,according to the identified abnormal data types,the bad data is concctcd by means of clustering center curve and correction algorithm.An example shows that the method has a better correction effect for different types of abnormal data,ensures the integrity of the bus data,and further improves the predictive efficiency and accuracy of bus load.