摘要
电力系统是强耦合、多变量系统,对其多元混沌时间序列预测是当前研究难点。提出了一种基于优化的时谱图神经网络,用于电力系统的混沌预测。利用潜在相关层挖掘多元时间序列之间的相关性,通过序列转换单元将时间序列转换为频域信号并学习其特征,结合多种算法优化模型实现更好的预测效果。试验表明经优化后的时谱图神经网络不仅能对电力系统的多状态变量进行混沌预测,而且比其他参考模型具有更高的预测精度和稳定性。
Power system is a strong coupling and multivariable system,and the prediction of its multivariate chaotic time series is difficult at present.In this paper,a time spectrum neural network based on optimization was proposed for chaos prediction of power system.Firstly,the potential correlation layer was used to mine the potential correlation between multivariate time series.Then,the time series were converted into frequency domain signals through the sequence conversion unit to learn their characteristics.Finally,a variety of algorithms were combined to optimize the model to achieve better prediction effect.The results show that the optimized time spectrum neural network can predict the multivariable chaos of power system,and has higher prediction accuracy and stability than other reference models.
作者
卢英东
韦笃取
LU Yingdong;WEI Duqu(College of Electronic Engineering,Guangxi Normal University,Guilin 541004,China)
出处
《振动与冲击》
EI
CSCD
北大核心
2023年第11期156-162,共7页
Journal of Vibration and Shock
基金
国家自然科学基金(62062014)
广西自然科学基金(2021JJA170004)
广西研究生教育创新计划项目(XYCSZ2021001)。
关键词
神经网络
电力系统
混沌
多元时间序列预测
优化算法
neural network
power system
chaos
multivariate time series prediction
optimization algorithm
