基于改进CNN-LSTM的高原地区光伏功率短期预测方法

普璇; 谢敏; 刘洋; 刘明波

doi:10.16516/j.ceec.2025-348

欢迎来到能源电力期刊网！

English Version 期刊登录

您当前的位置：

首页 >

文章列表页 >

基于改进CNN-LSTM的高原地区光伏功率短期预测方法

极端气候事件下的能源与交通韧性 | 更新时间：2026-10-14

- 基于改进CNN-LSTM的高原地区光伏功率短期预测方法
- Short-Term Photovoltaic Power Forecasting in Plateau Regions Based on an Improved CNN-LSTM Model
- 南方能源建设 2026年13卷第3期页码：103-118
- 作者机构：
  
  华南理工大学电力学院,广东,广州,510640
- 作者简介：
- 基金信息：
  
  广东省基础与应用基础研究基金资助项目“适应宽功率波动海上风电特性的制氢系统建模与集成协调控制方法研究”（2022A1515240074）;广东省基础与应用基础研究基金重点项目“海上风电集群主动支撑电力系统电压和频率的控制方法”（2024B1515250007）
- DOI：10.16516/j.ceec.2025-348
  中图分类号： TK519;TM615;TP18
- 纸质出版：2026
- 稿件说明：
移动端阅览
普璇, 谢敏, 刘洋, 等. 基于改进CNN-LSTM的高原地区光伏功率短期预测方法[J]. 南方能源建设, 2026, 13(3): 103-118.

.
普璇, 谢敏, 刘洋, 等. 基于改进CNN-LSTM的高原地区光伏功率短期预测方法[J]. 南方能源建设, 2026, 13(3): 103-118. DOI： 10.16516/j.ceec.2025-348.

. DOI： 10.16516/j.ceec.2025-348.

摘要

目的

为响应国家“双碳”目标，我国高原地区光伏项目建设正快速推进。但由于该地区光伏出力波动性较强，传统预测方法易出现形状拟合度不佳和预测精度不高的问题，给电力系统调控带来诸多挑战。

方法

基于此，文章提出一种改进CNN-LSTM的高原地区光伏短期功率预测方法。该方法在传统CNN-LSTM模型中引入DTW距离作为曲线形状修正指标，通过Kendall不确定性加权法对损失函数中均方误差MSE和DTW距离两损失项的权重进行自适应加权，解决了多目标损失函数中权重系数依赖经验调参的难题。同时，为解决传统Kendall不确定性加权方法中存在的过拟合与局部最优等问题，在模型训练过程中引入演化博弈-突变机制对损失项的权重进行寻优，以实现对各损失项权重的动态调整，提升训练效果和预测精度。

结果

研究表明，改进的CNN-LSTM模型能够适应高原地区光伏强波动性特征，有效提升该地区的预测精度。相较于传统主流预测模型而言，改进的CNN-LSTM模型MAE平均下降37.5%，RMSE平均下降31.27%，

平均提升8%。

结论

经算例验证，所提方法在高原地区光伏短期功率预测场景下能够取得较好的预测效果，具备较强的应用潜力。

Abstract

Objective

In response to the country's "dual carbon" goals

the construction of photovoltaic projects in China's plateau regions is rapidly advancing. However

due to the strong variability of photovoltaic output in these areas

traditional forecasting methods often face issues such as poor shape fitting and low prediction accuracy

posing challenges for power system regulation.

Method

To address this

this paper proposed an improved CNN-LSTM short-term photovoltaic power forecasting method for plateau regions. This method introduced DTW distance into the traditional CNN-LSTM model as a curve shape correction metric

and used the Kendall uncertainty weighting method to adaptively adjust the weights of the two loss terms

mean squared error (MSE) and DTW distance

in the loss function

overcoming difficulties in tuning loss term weights during multi-objective model training. At the same time

to solve problems such as overfitting and local optima present in the traditional Kendall uncertainty weighting method

an evolutionary game-mutation mechanism was introduced during model training to optimize the weights of the loss terms

enabling dynamic adjustment of each loss term's weight to improve training effectiveness and prediction accuracy.

Result

The study shows that the improved CNN-LSTM model can adapt to the strong variability of photovoltaic output in plateau regions and effectively enhance prediction accuracy. A photovoltaic power station in the southwestern plateau of Guizhou is taken as an example

with validation conducted across multiple typical days under different seasonal and meteorological conditions. The results indicate that

compared with traditional mainstream forecasting models

the model's MAE decreases by an average of 37.5%

RMSE decreases by an average of 31.27%

and

increases by an average of 8%.

Conclusion

The case studies demonstrate that the proposed method can achieve good forecasting performance for short-term photovoltaic power in plateau regions

showing strong application potential.

关键词

Keywords

references

LIU W, YANG K N. Short term photovoltaic power prediction in high-altitude and cold regions based on numerrical weather prediction factor expansion and improved ensemble learning [J]. Electrical engineering, 2024, 25(8): 1-10, 17. DOI: 10.3969/j.issn.1673-3800.2024.08.001.

ZHANG J, XIONG G J. Short-term photovoltaic power prediction combination model considering seasonal characteristic and data window [J]. Electric power engineering technology, 2025, 44(1): 183-192. DOI: 10.12158/j.2096-3203.2025.01.019.

BREITKREUZ H, SCHROEDTER-HOMSCHEIDT M, HOLZER-POPP T, et al. Short-range direct and diffuse irradiance forecasts for solar energy applications based on aerosol chemical transport and numerical weather modeling [J]. Journal of applied meteorology and climatology, 2009, 48(9): 1766-1779. DOI: 10.1175/2009jamc2090.1.

STRAUB N, KARALUS S, HERZBERG W, et al. Satellite‐based solar irradiance forecasting: replacing cloud motion vectors by deep learning [J]. Solar RRL, 2024, 8(24): 2400475. DOI: 10.1002/solr.202400475.

BHAVSAR S, PITCHUMANI R, ORTEGA-VAZQUEZ M A. Machine learning enabled reduced-order scenario generation for stochastic analysis of solar power forecasts [J]. Applied energy, 2021, 293: 116964. DOI: 10.1016/j.apenergy.2021.116964.

LI Y Z, RU L, NIU J C. Power structure optimization for grid-connected photovoltaic system based on Markov decision processes [C]//IEEE. Proceedings of 2011 6th IEEE Conference on Industrial Electronics and Applications, Beijing, China, June 21-23, 2011. New York: IEEE, 2011: 2688-2692. DOI: 10.1109/ICIEA.2011.5976051.

WEN B, ZHANG X Q, FU W L, et al. Short-term PV power generation forecasting based on quadratic mode decomposition reconstruction and BiTCN-BiGRU model [J]. Power system protection and control, 2025, 53(18): 74-87. DOI: 10.19783/j.cnki.pspc.241500.

CHEN X X, WU K T, HE J, et al. Short term photovoltaic output interval prediction based on integrated machine learning model [J]. Smart power, 2024, 52(2): 87-93, 107. DOI: 10.3969/j.issn.1673-7598.2024.02.012.

DONG J, LIU R, SHU H C, et al. Short-term distributed photovoltaic power generation prediction based on BIRCH clustering and L-Transformer [J]. High voltage engineering, 2024, 50(9): 3883-3893. DOI: 10.13336/j.1003-6520.hve.20230809.

TAN H W, YANG Q L, XING J C, et al. Photovoltaic power prediction based on combined XGBoost-LSTM model [J]. Acta energiae solaris sinica, 2022, 43(8): 75-81. DOI: 10.19912/j.0254-0096.tynxb.2021-0005.

WANG C Y, DUAN Q Q, ZHOU K, et al. A hybrid model for photovoltaic power prediction of both convolutional and long short-term memory neural networks optimized by genetic algorithm [J]. Acta physica sinica, 2020, 69(10): 100701. DOI: 10.7498/aps.69.20191935.

ZOU L L, QUAN H Y, TANG P, et al. A photovoltaic power prediction framework based on multi-stage ensemble learning [J]. Energies, 2025, 18(17): 4644. DOI: 10.3390/en18174644.

YUAN J H, XIE B B, HE B L, et al. Short term forecasting method of photovoltaic output based on DTW-VMD-PSO-BP [J]. Acta energiae solaris sinica, 2022, 43(8): 58-66. DOI: 10.19912/j.0254-0096.tynxb.2020-1405.

WANG H, YAN J, ZHANG J W, et al. Short-term integrated forecasting method for wind power, solar power, and system load based on variable attention mechanism and multi-task learning [J]. Energy, 2024, 304: 132188. DOI: 10.1016/j.energy.2024.132188.

LI Z, CUI L Y, TAO Y J, et al. The combined model for PV power forecasting based on combination weighting approach of game theory [J]. Electrical engineering, 2017(5): 24-29. DOI: 10.3969/j.issn.1673-3800.2017.05.005.

YANG L C, BOTTOU L, BENGIO Y, et al. Gradient-based learning applied to document recognition [J]. Proceedings of the IEEE, 1998, 86(11): 2278-2324. DOI: 10.1109/5.726791.

HOCHREITER S, SCHMIDHUBER J. Long short-term memory [J]. Neural computation, 1997, 9(8): 1735-1780. DOI: 10.1162/neco.1997.9.8.1735.

BERNDT D J, CLIFFORD J. Using dynamic time warping to find patterns in time series [C]//AAAI Press. Proceedings of the 3rd International Conference on Knowledge Discovery and Data Mining, Seattle, USA, July 31, 1994-August 1, 1994. AAAI Press, 1994: 359-370.

TAYLOR P D, JONKER L B. Evolutionary stable strategies and game dynamics [J]. Mathematical biosciences, 1978, 40(1/2): 145-156. DOI: 10.1016/0025-5564(78)90077-9.

HUANG K N. Evolutionary games and evolutionary economics [J]. Economic research journal, 2009, 44(2): 132-145.

CIPOLLA R, GAL Y, KENDALL A. Multi-task learning using uncertainty to weigh losses for scene geometry and semantics [C]//IEEE. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, June 18-23, 2018. New York: IEEE, 2018: 7482-7491. DOI: 10.1109/CVPR.2018.00781.

BALDUZZI D, RACANIÈRE S, MARTENS J, et al. The mechanics of n -player differentiable games [C ] //PMLR. Proceedings of the 35th International Conference on Machine Learning, Stockholm, Sweden, July 10-15, 2018. Proceedings of Machine Learning Research, 2018: 354-363.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

吸附式液态二氧化碳储能系统热经济性分析

高比例风光集群并网多时间尺度惯量协同与暂态稳定优化控制

锂离子电池组TEC-液冷复合热管理系统仿真与性能分析

Cu/Al共掺杂P2-Na0.72Fe0.3Mn0.7O2正极材料的制备及其电化学性能

储能技术在发电侧的综合应用与进展