Electricity price Jump-Diffusion forecasting model integrating dynamical structure and KMCrossformer deep learning

LI Jike; HU Zhijian

doi:10.14188/j.1671-8844.2025.0267

您当前的位置：

首页 >

文章列表页 >

Electricity price Jump-Diffusion forecasting model integrating dynamical structure and KMCrossformer deep learning

更新时间：2026-10-16

- Electricity price Jump-Diffusion forecasting model integrating dynamical structure and KMCrossformer deep learning
- Engineering Journal of Wuhan University Pages: 1-13(2026)
- 作者机构：
  
  武汉大学电气与自动化学院，湖北武汉 430072
- 作者简介：
- 基金信息：
- DOI：10.14188/j.1671-8844.2025.0267
  CLC： F407.61
- Received：27 June 2025，
  
  Online First：16 October 2026，
- 稿件说明：
移动端阅览
LI Jike, HU Zhijian. Electricity price Jump-Diffusion forecasting model integrating dynamical structure and KMCrossformer deep learning[J/OL]. Engineering Journal of Wuhan University, 2026, 1-13.
DOI：

LI Jike, HU Zhijian. Electricity price Jump-Diffusion forecasting model integrating dynamical structure and KMCrossformer deep learning[J/OL]. Engineering Journal of Wuhan University, 2026, 1-13. DOI： 10.14188/j.1671-8844.2025.0267.

摘要

随着电力市场环境的日益复杂，电价序列中呈现出愈加显著的非线性波动特征及尖峰行为。为提升电价预测的精度，提出一种跳跃扩散变体框架下的动力学预测模型。首先，基于局部离群因子（Local Outlier Factor

LOF）识别出局部离群点并通过时间缓冲机制聚合为跳跃事件簇；其次，引入Koopman算子对系统的非线性演化过程进行升维建模，从中解耦出跳跃项在状态空间中的主导动力学模式；最后，在升维层面将解耦模式作为高维特征嵌入至KMCrossformer跨维度注意力机制网络，在深层次融合跳跃项与扩散项的同时，实现对电价序列长期趋势与短时剧烈扰动的协同建模。研究通过澳大利亚昆士兰州数据集的算例分析证明了所提出预测模型在尖峰捕捉能力与整体预测分布拟合度方面的优越性。

Abstract

With the increasing complexity of the electricity market environment

electricity price series exhibit increasingly pronounced nonlinear fluctuations and spike behaviors. To enhance the accuracy of electricity price forecasting

this study proposes a dynamical prediction model under a Jump-Diffusion variant framework. First

the Local Outlier Factor (LOF) method is employed to identify local outliers

which are then aggregated into jump event clusters via a temporal buffering mechanism. Second

the Koopman operator is introduced to perform a lifted modeling of the system’s nonlinear evolution

enabling the decoupling of dominant dynamical modes associated with jump components in the state space. Finally

at the lifted level

the decoupled modes are embedded as high-dimensional features into the KMCrossformer network with a cross-dimensional attention mechanism. This design allows for deep integration of jump and diffusion components

thereby achieving coordinated modeling of both long-term trends and short-term abrupt disturbances in electricity price series. Case studies based on the Queensland

Australia dataset demonstrate the proposed model’s superiority in spike capturing ability and overall distributional fit of predictions.

关键词

Keywords

references

陈宇聪 , 白晓清 . 考虑时序二维变化的日前市场电价预测模型 [J ] . 电力系统及其自动化学报 , 2024 , 36 ( 07 ): 22 - 29 .

CHEN Yucong , BAI Xiaoqing . Forecasting model of day-ahead market electricity price considering 2D variation of time series [J ] . Proceedings of the CSU-EPSA , 2024 , 36 ( 07 ): 22 - 29 .

刘丽燕 , 邹小燕 . GARCH族模型在电力市场电价预测中的比较研究 [J ] . 电力系统保护与控制 , 2016 , 44 ( 04 ): 57 - 63 .

LIU Liyan , ZOU Xiaoyan . Comparison of GARCH models in electricity price forecasting [J ] . Power System Protection and Control , 2016 , 44 ( 04 ): 57 - 63 .

李国庆 , 刘钊 , 金国彬 , 等 . 基于随机分布式嵌入框架及BP神经网络的超短期电力负荷预测 [J ] . 电网技术 , 2020 , 44 ( 2 ): 437 - 445 .

LI Guoqing , LIU Zhao , JIN Guobi , et al . Ultra short-term power load forecasting based on randomly distributive embedded framework and BP neural network [J ] . Power System Technology , 2020 , 44 ( 2 ): 437 - 445 .

陈杰尧 , 陶春华 , 马光文 , 等 . 基于数据挖掘与支持向量机的现货市场出清价预测方法 [J ] . 电网与清洁能源 , 2020 , 36 ( 10 ): 14 - 19+27 .

CHEN Jieyao , TAO Chunhua , MA Guangwen , et al . Forecasting method of spot market clearing price based on data mining and support vector machine [J ] . Power System and Clean Energy , 2020 , 36 ( 10 ): 14 - 19+27 .

徐任超 , 阎威武 , 王国良 , 等 . 基于周期性建模的时间序列预测方法及电价预测研究 [J ] . 自动化学报 , 2020 , 46 ( 06 ): 1136 - 1144 .

XU Ren-Chao , YAN Wei-Wu , WANG Guo-Liang , et al . Time series forecasting based on seasonality modeling and its application to electricity price forecasting [J ] . Acta Automatica Sinica , 2020 , 46 ( 06 ): 1 136 - 1144 .

吉兴全 , 曾若梅 , 张玉敏 , 等 . 基于注意力机制的CNN-LSTM短期电价预测 [J ] . 电力系统保护与控制 , 2022 , 50 ( 17 ): 125 - 132 .

JI Xingquan , ZENG Ruomei , ZHANG Yumin , et al . CNN-LSTM short-term electricity price prediction based on an attention mechanism [J ] . Power System Protection and Control , 2022 , 50 ( 17 ): 125 - 132 .

赵雅雪 , 王旭 , 蒋传文 , 等 . 基于最大信息系数相关性分析和改进多层级门控LSTM的短期电价预测方法 [J ] . 中国电机工程学报 , 2021 , 41 ( 01 ): 135 - 146+404 .

ZHAO Yaxue , WANG Xu , JIANG Chuanwen , et al . A novel short-term electricity price forecasting method based on correlation analysis with the maximal information coefficient and modified multi-hierachy gated LSTM [J ] . Proceedings of the CSEE , 2021 , 41 ( 01 ): 135 - 146+404 .

Vaswani A , Shazeer N , Parmar N , et al . Attention is all you need [J ] . Advances in neural information processing systems , 2017 , 30 .

Cheng Qilin , Li Zihao , Wang Xin , et al . Autoformer based electricity market price forecasting for risk mitigation [C ] // 2023 IEEE 7th Conference on Energy Internet and Energy System Integration (EI2) . IEEE , 2023 : 3426 - 3431 .

Nie Yuqi , Nguyen N H , Sinthong P , et al . A time series is worth 64 words: Long-term forecasting with transformers [J ] . arXiv preprint arXiv: 2211.14730 , 2022 .

Afanasyev D O , Fedorova E A . On the impact of outlier filtering on the electricity price forecasting accuracy [J ] . Applied Energy , 2019 , 236 : 196 - 210 .

Christensen T M , Hurn A S , Lindsay K A . Forecasting spikes in electricity prices [J ] . International Journal of Forecasting , 2012 , 28 ( 2 ): 400 - 411 .

Cuaresma J C , Hlouskova J , Kossmeier S , et al . Forecasting electricity spot-prices using linear univariate time-series models [J ] . Applied Energy , 2004 , 77 ( 1 ): 87 - 106 .

Huisman R , Mahieu R . Regime jumps in electricity prices [J ] . Energy economics , 2003 , 25 ( 5 ): 425 - 434 .

Chan K F , Gray P , van Campen B . A new approach to characterizing and forecasting electricity price volatility [J ] . International Journal of Forecasting , 2008 , 24 ( 4 ): 728 - 743 .

Zhang Chenxu , Fu Yong , Gong Lin . Short-term electricity price forecast using frequency analysis and price spikes oversampling [J ] . IEEE Transactions on Power Systems , 2022 , 38 ( 5 ): 4739 - 4751 .

Vu D H , Muttaqi K M , Agalgaonkar A P , et al . Short-term forecasting of electricity spot prices containing random spikes using a time-varying autoregressive model combined with kernel regression [J ] . IEEE Transactions on Industrial Informatics , 2019 , 15 ( 9 ): 5378 - 5388 .

Shi Wei , Wang Yufeng , Chen Yiyuan , et al . An effective two-stage electricity price forecasting scheme [J ] . Electric Power Systems Research , 2021 , 199 : 107416 .

Liu Shengyuan , Jiang Yicheng , Lin Zhenzhi , et al . Data-driven two-step day-ahead electricity price forecasting considering price spikes [J ] . Journal of Modern Power Systems and Clean Energy , 2022 , 11 ( 2 ): 523 - 533 .

Mount T D , Ning Yumei , Cai Xiaobin . Predicting price spikes in electricity markets using a regime-switching model with time-varying parameters [J ] . Energy Economics , 2006 , 28 ( 1 ): 62 - 80 .

李正欣 , 赵林度 . 电力市场中的价格钉研究 [J ] . 继电器 , 2008 ,( 01 ): 52 - 56+62 .

LI Zheng-xin , ZHAO Lin-du . Study of price spikes in power market [J ] . RELAY , 2008 ,( 01 ): 52 - 56+62 .

庄池杰 , 张斌 , 胡军 , 等 . 基于无监督学习的电力用户异常用电模式检测 [J ] . 中国电机工程学报 , 2016 , 36 ( 2 ): 379 - 387 .

ZHUANG Chijie , ZHANG Bin , HU Jun , et al . Anomaly detection for power consumption patterns based on unsupervised learning [J ] . Proceedings of the CSEE , 2016 , 36 ( 2 ): 379 - 387 .

崔晓杰 , 王家耀 , 巩现勇 , 等 . 利用模糊密度聚类和双向缓冲区自动识别热点区 [J ] . 武汉大学学报(信息科学版) , 2019 , 44 ( 01 ): 84 - 91 .

CUI Xiaojie , WANG Jiayao , GONG Xianyong , et al . Hotspot area recognition by using fuzzy density clustering and bidirectional buffer [J ] . Geomatics and Information Science of Wuhan University , 2019 , 44 ( 01 ): 84 - 91 .

Brunton S L , Brunton B W , Proctor J L , et al . Koopman invariant subspaces and finite linear representations of nonlinear dynamical systems for control [J ] . PloS one , 2016 , 11 ( 2 ): e0150171 .

贾继莹 . Koopman算符在一些动力系统中的算法和应用研究 [D ] . 北京 : 清华大学 , 2015 .

Jia Jiying . Research on the algorithms and applications of Koopman operator on some dynamical systems [D ] . Beijing : Tsinghua University , 2015 .

杨雨昕 , 姚伟 , 邓贤哲 , 等 . 基于Koopman时序延拓和CNN-Transformer模型的频率稳定指标预测 [J ] . 电网技术 , 2025 , 49 ( 08 ): 3474 - 3485 .

YANG Yuxin , YAO Wei , DENG Xianzhe , et al . Frequency stability index prediction based on Koopman time series extension and CNN-Transformer model [J ] . Power System Technology , 2025 , 49 ( 08 ): 3474 - 3485 .

Susuki Y , Mezić I . Nonlinear Koopman modes and power system stability assessment without models [J ] . IEEE Transactions on Power Systems , 2013 , 29 ( 2 ): 899 - 907 .

周一辰 , 李金泽 , 李永刚 , 等 . Koopman原理内嵌MLP神经网络模型驱动的电力系统非线性振荡特征分析方法 [J ] . 电力自动化设备 , 2024 , 44 ( 10 ): 132 - 139 .

ZHOU Yichen , LI Jinze , LI Yonggang , et al . Analysis method of nonlinear oscillation characteristics for power system driven by MLP neural network model embedded in Koopman principle [J ] . Electric Power Automation Equipment , 2024 , 44 ( 10 ): 132 - 139 .

Zhang Yunhao , Yan Junchi . Crossformer: Transformer utilizing cross-dimension dependency for multivariate time series forecasting [C ] // The eleventh international conference on learning representations . 2023 .

蔡屹 , 张薇 . 基于STL-Crossformer的综合能源系统多元负荷预测 [J ] . 东北电力大学学报 , 2024 , 44 ( 01 ): 34 - 41 .

CAI Yi , ZHANG Wei . Multi-Load Forecasting of Integrated Energy Systems Based on STL-Crossformer [J ] . Journal of Northeast Electric Power University , 2024 , 44 ( 01 ): 34 - 41 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Water level prediction and impact analysis for Gezhouba Hydropower Station based on deep learning

A multi-level edge detection algorithm combining the scale enhancement module and edge attention

UAV LiDAR point cloud semantic segmentation for distribution network live working scenarios based on CSA⁃PointCNN

Research on ecological flow optimal scheduling of the Liuxi River based on deep learning

A Chinese short text classification algorithm based on auxiliary tasks and knowledge distillation

Related Author

ZHANG Wei

LIU Yuan

DENG Wenliang

LIU Pan

FEI Xiaobo

ZHAO Hui

CHENG Lei

LIU Qi

Related Institution

State Key Laboratory of Water Resources Engineering and Management， Wuhan University

China Yangtze Power Co.， Ltd.

Science and Technology Research Institute， China Three Gorges Corporation

School of Electronic Information， Wuhan University

Deqing Power Supply Company，State Grid Zhejiang Electric Power Co．，Ltd．，Deqing

AI问答

⁰