论文部分内容阅读
提出了一种最优FCM聚类分析和最小二乘支持向量机回归算法(LSSVR)相结合的电力系统短期负荷预测方法。在考虑电力系统负荷日周期性的基础上,运用基于改进划分系数最大原则的最优FCM聚类分析获取历史负荷样本的最优数据模式划分,并根据输入样本相似度选取LSS-VR训练样本。既强化了训练样本的数据规律,又保证了数据特征的一致性,从而提高了LSSVR训练速度,改善了预测效果。仿真实验表明:LSSVR点模型的平均预测精度约98%,而本文模型的平均预测精度达到了98.7%,证明了该方法的有效性和实用性。
A method of power system short-term load forecasting based on optimal FCM clustering analysis and least squares support vector machine regression (LSSVR) is proposed. Based on the periodicity of power system load day, the optimal data model partition of historical load samples is obtained by using the optimal FCM clustering analysis based on the maximum principle of improved partition coefficient, and LSS-VR training samples are selected according to the similarity of input samples. This not only enhances the data regularity of the training samples, but also ensures the consistency of the data features, so as to improve the LSSVR training speed and improve the prediction effect. Simulation results show that the average prediction accuracy of LSSVR point model is about 98%, and the average prediction accuracy of this model reaches 98.7%, which proves the effectiveness and practicability of the proposed method.