相空间重构和支持向量机结合的电力负荷预测模型研究

发布时间：2018-03-22 23:00

本文选题：支持向量回归　切入点：混沌　出处：《电测与仪表》2014年24期 　论文类型：期刊论文

【摘要】：风电机组集中并网会对电网安全稳定运行带来影响,为了合理规划各类供电机组高效运行,对电力负荷预测的精度提出了更高的要求。电网负荷时间序列具有混沌特性,普通预测方法难以描述其特性和内在规律。利用混沌相空间重构理论对负荷时间序列进行研究,用互信息法和CAO方法分别求得时间延迟和嵌入维数,并由此得到系统最大李雅普诺夫指数,证明其具有混沌特性。然后根据时间延迟和嵌入维数对样本数据相空间重构,在此基础上利用支持向量回归算法(PSR-SVR)对电力负荷进行预测,支持向量回归采用网格寻优确定参数。最后将预测的结果同时间序列模型和BP神经网络模型的预测结果进行对比,结果表明,这是一种误差小,精度高的电网负荷预测方法器。
[Abstract]:Centralized grid connection of wind turbines will have an impact on the safe and stable operation of the power grid. In order to reasonably plan the efficient operation of all kinds of power supply units, a higher requirement for the accuracy of power load forecasting is put forward. The time series of power load has chaotic characteristics. The general prediction method is difficult to describe its characteristics and inherent laws. The chaotic phase space reconstruction theory is used to study the load time series, and the time delay and embedding dimension are obtained by mutual information method and CAO method, respectively. The maximum Lyapunov exponent of the system is obtained, which is proved to be chaotic. Then the phase space of the sample data is reconstructed according to the time delay and embedding dimension, and then the power load is predicted by using the support vector regression algorithm (PSR-SVR). Support vector regression uses grid optimization to determine parameters. Finally, the prediction results are compared with those of time series model and BP neural network model. The results show that this method is of low error and high precision.
【作者单位】：中北大学机械与动力工程学院;
【基金】：山西省自然科学基金资助项目(2013011026-2)
【分类号】：TM715

【参考文献】