基于VMD-MSFOA-LSTM短期电力负荷预测

Short-term Power Load Forecasting Based on VMD-MSFOA-LSTM

摘要

为提高电力系统中电力负荷预测的精度，本文提出一种基于多策略改进的海星优化算法（multi strategy improved starfish optimization algorithm，MSFOA），优化变分模态分解（VMD）与长短期记忆（LSTM）网络相结合的电力负荷预测模型（VMD-MSFOA-LSTM）。首先，MSFOA 算法结合了佳点集初始化策略、动态对立学习机制、线性递减系数和精英增强解质量策略；其次，基于 MSFOA 算法优化 LSTM 网络的超参数，搭建 VMD-MSFOA-LSTM 负荷预测模型；最后，选取澳大利亚电力负荷数据集对模型进行评估。实验结果表明，VMD-MSFOA-LSTM 模型较原始 LSTM 模型均方根误差、平均绝对误差和平均绝对百分比误差分别降低了86.53%、86.71%和87.59%，决定系数提高了16.94% ，验证了本文所提模型在电力负荷预测中的优越性能。

Abstract

To enhance the accuracy of power load forecasting in power systems，a power load forecasting model（VMD-MSFOA-LSTM）was proposed，that combined variational mode decomposition （VMD）with a long short-term memory（LSTM）network，optimized by a multi-strategy improved starfish optimization algorithm（MSFOA）.Firstly，MSFOA integrated the Halton sequence initialization strategy，dynamic opposition-based learning mechanism，linearly decreasing coefficient，and elite enhancement solution quality strategy.Secondly，the hyperparameters of the LSTM network were optimized based on MSFOA to construct the VMD-MSFOA-LSTM load forecasting model.Finally，the Australian power load dataset was selected to evaluate the model.The experimental results show that compared with the original LSTM model，the VMD-MSFOA-LSTM model reduces the root mean square error，mean absolute error，and mean absolute percentage error by 86.53%，86.71%，and 87.59%，respectively，and increases the coefficient of determination by 16.94%，verifying the superior performance of the proposed model in power load forecasting.

关键词

电力负荷预测 / 海星优化算法 / 变分模态分解 / 长短期记忆网络

Key words

power load forecasting / starfish optimization algorithm / variational mode decomposition / long short-term memory network

参考文献

[1] Shao L, Guo Q J, Li C, et al. Short-term load forecasting based on EEMD-WOA-LSTM combination model[J]. Applied Bionics and Biomechanics, 2022, 2022:2166082.

[2] VASUDEVAN S, JOTHINATHAN K. Short-term electrical load forecasting through optimally configured long short-term memory[J]. Majlesi Journal of Electrical Engineering, 2024, 18(4):1-10.

[3] 蔡春雷, 刘微, 任腾腾. 基于 HDBO-LSTM 的短期电力负荷预测方法[J]. 沈阳理工大学学报, 2025, 44(5):21-28.

[4] Cai C L, Liu W, Ren T T. Short-term power load forecasting method based on HDBO-LSTM[J]. Journal of Shenyang Ligong University, 2025, 44(5):21-28. (in Chinese)

[5] Fang M K, Zhang F F, Zhu D, et al. The influence of variational mode decomposition on LSTM prediction accuracy-a case study with wind turbine power signals[J]. Renewable Energy, 2025, 245:122863.

[6] Dragomiretskiy K, Zosso D. Variational mode decomposition[J]. IEEE Transactions on Signal Processing, 2014, 62(3): 531-544.

[7] 盛四清, 金航, 刘长荣. 基于 VMD-WSGRU 的风电场发电功率中短期及短期预测[J]. 电网技术, 2022, 46(3):897-904.

[8] Sheng S Q, Jin H, Liu C R. Short-term and mid-short-term wind power forecasting based on VMD-WSGRU[J]. Power System Technology, 2022, 46(3):897-904. (in Chinese)

[9] Awad N H, Ali M Z, Suganthan P N. Ensemble sinusoidal differential covariance matrix adaptation with Euclidean neighborhood for solving CEC2017 benchmark problems[C]//2017 IEEE Congress on Evolutionary Computation(CEC). Donostia,Spain:IEEE,2017:372-379.

[10] Zhang G, Liu H C, Zhang J B, et al. Wind power prediction based on variational mode decomposition multi-frequency combinations[J]. Journal of Modern Power Systems and Clean Energy, 2019, 7(2):281-288.

[11] Van Houdt G, Mosquera C, Nápoles G. A review on the long short-term memory model[J]. Artificial Intelligence Review, 2020, 53(8):5929-5955.

[12] Zhong C T, Li G, Meng Z, et al. Starfish optimization algorithm(SFOA):a bio-inspired metaheuristic algorithm for global optimization compared with 100 optimizers[J]. Neural Computing and Applications, 2025, 37(5):3641-3683.

[13] Oueslati R, Ouertani M W, Amdouni A, et al. MS-FSOA-LightGBM:multi-strategy starfish optimization algorithm with LightGBM for software defect prediction[J]. Procedia Computer Science, 2025, 270:4054-4063.

[14] Liu S L, Jin Z K, Lin H T, et al. An improve crested porcu- pine algorithm for UAV delivery path planning in challenging environments[J]. Scientific Reports, 2024, 14:20445.

[15] Abdollahpour A, Rouhi A, Pira E. An improved gazelle optimization algorithm using dynamic opposition-based learning and chaotic mapping combination for solving optimization problems[J]. The Journal of Supercomputing, 2024, 80 (9): 12813-12843.

[16] 尚凯凯. 融合扰动策略的自适应哈里斯鹰优化算法[J]. 计算机与数字工程, 2025, 53(2):338-346.

[17] Shang K K. Adaptive Harris Hawks optimization algorithm based on disturbance strategies[J]. Computer and Digital Engineering, 2025, 53(2):338-346. (in Chinese)

[18] Mirjalili S, Lewis A. The whale optimization algorithm[J]. Advances in Engineering Software, 2016, 95:51-67.

[19] Hashim F A, Hussain K, Houssein E H, et al. Archimedes optimization algorithm:a new metaheuristic algorithm for solving optimization problems[J]. Applied Intelligence, 2021, 51 (3):1531-1551.

[20] Amiri M H, Mehrabi Hashjin N, Montazeri M, et al. Hippopot- amus optimization algorithm:a novel nature-inspired optimi- zation algorithm[J]. Scientific Reports, 2024, 14:5032.

[21] Guan Y M, Shi Y L, Wang G, et al. Electricity behaviors anomaly detection based on multi-feature fusion and contrastive learning[J]. Information Systems, 2025, 127:102457.

基金资助

辽宁省教育厅高等学校基本科研项目(JYTMS20230189)

沈阳理工大学引进高层次人才科研支持计划项目(1010147001131)