基于变分模态分解和混合深度神经网络的短期电价预测  

Short term electricity price prediction based on variational mode decomposition and hybrid deep neural network

作　　者：刘羿萱 杨昭 LIU Yixuan;YANG Zhao(Ultra High Voltage Company of State Grid Shaanxi Electric Power Co.,Ltd,Xi’an 710025)

机构地区：[1]国网陕西省电力有限公司超高压公司,西安710025

出　　处：《电气技术》2025年第3期30-35,共6页Electrical Engineering

摘　　要：针对电力市场中电价数据的非线性、波动性及时序性等特征,提出一种基于变分模态分解(VMD)和混合深度神经网络的短期电价预测方法。首先利用变分模态分解法将原始电价序列分解为多个平稳的子序列,其次采用混合深度神经网络预测模型对各子序列分别进行预测并叠加,得到最终的电价预测结果。该模型将卷积神经网络(CNN)和双向长短期记忆(BiLSTM)网络组合,提取原始电价数据的空间特征和时间特征,并结合Attention机制,对原始电价序列中不同时刻电价数据的重要性进行区分。最后,以美国PJM电力市场实际电价数据进行仿真分析,并与多种电价预测模型进行对比,结果验证了本文所提方法的有效性。A short term electricity price prediction method based on variational mode decomposition and hybrid deep neural network is proposed to address the characteristics of nonlinearity,volatility,and timeliness in electricity price data in the electricity market.Firstly,the original electricity price sequence is decomposed into multiple stationary subsequences using variational mode decomposition(VMD).Secondly,a hybrid deep neural network prediction model is used to predict and superimpose each subsequence separately,obtaining the final electricity price prediction result.This model combines convolutional neural network(CNN)and bidirectional long short term memory(BiLSTM)network to effectively extract spatial and temporal features of the original electricity price data,and combines attention mechanism to effectively distinguish the importance of electricity price data at different times in the original electricity price sequence.Finally,simulation analysis is conducted using actual electricity price data from the PJM electricity market in the United States,and the effectiveness of the proposed method is verified by comparing multiple electricity price prediction models.

关 键 词：短期电价预测 变分模态分解(VMD) 卷积神经网络(CNN) 双向长短期记忆(BiLSTM)网络 注意力机制 

分 类 号：F416.61[经济管理—产业经济] TP183[自动化与计算机技术—控制理论与控制工程]