基于蚁群优化算法的电力系统负荷序列的聚类分析

依据神经网络原理短期负荷预测模型的性能,负荷样本空间的分布特性对预测精度有大的影响,并且外部气象因素对负荷敏感性的复杂非线性关系也将使预测精度降低.运用负荷序列特征的聚类分析与模式识别相结合原理可解决该问题.该文提出了基于蚁群优化算法（ant colony optimization Algorithm,ACOA）的电力系统负荷序列聚类分析.通过对实际地区负荷系统的聚类分析显示其优越性;并证实基于ACOA的聚类比Kohonen神经网络聚类对气候异常情况、高温区域、节假日都具有更高的敏感性和分辨率;对负荷曲线轮廓的相似性具有更细腻和更均匀的聚类特性.上述的聚类特性对STLF精度的提高是极其重要的.

CLUSTERING ANALYSIS OF POWER SYSTEM LOAD SERIES BASED ON ANT COLONY OPTIMIZATION ALGORITHM

According to the performance of short-term load forecasting(STLF)model based on the principle of artificial neural networks(ANN),the forecasting accuracy is influenced by the distributed feature of load sample space,and the complex nonlinear relation which is formed by the sensibility of external weather factors to power load will also lead to the reduce of forecasting accuracy.To use power load series for characteristic clustering combination with pattern recognition may use as one method of solving the problem.In this paper,the characteristic clustering and its analysis to power load series based on Ant Colony Optimization Algorithm(ACOA) was presented.The load clustering performance of ACOA in actual load system has shown its superiority,which has more sensitivity and resolution to climatic anomaly circumstances,high temperature area,to festival and holiday condition than Kohonen neural network based clustered method,and which has more exquisite and even of the clustering characteristics on the similarity of load curve profile.The above clustering performance has a most important significance to improve the accuracy of STLF.