[1]侯鹏飞,母宇宸,王月璠,等.基于BKA-BiLSTM-KD 模型的超短期光伏发电功率预测[J].机械与电子,2026,44(08):103-109.
　HOU Pengfei,MU Yuchen,WANG Yuepan,et al.Ultra-short-term Photovoltaic Power Prediction Based on BKA-BiLSTM-KD Model[J].Machinery & Electronics,2026,44(08):103-109.
点击复制

基于BKA-BiLSTM-KD 模型的超短期光伏发电功率预测()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
44
期数:
2026年08期
页码:
103-109
栏目:
电力控制
出版日期:
2026-08-25

文章信息/Info

Title:
Ultra-short-term Photovoltaic Power Prediction Based on BKA-BiLSTM-KD Model
文章编号:
1001-2257(2026)08-0103-07
作者:
侯鹏飞1; 母宇宸1; 王月璠1; 常奇峰2; 李俏菡1
1.华北水利水电大学电气信息工程学院,河南 郑州 450045;
2.许继电气股份有限公司,河南 许昌 461000
Author(s):
HOU Pengfei1; MU Yuchen1; WANG Yuepan1; CHANG Qifeng2; LI Qiaohan1
(1.School of Electrical Engineering,North China University of Water Resources and Electric Power,Zhengzhou 450045,China;
2.XJ Electric Co.,Ltd.,Xuchang 461000,China)
关键词:
功率预测; 随机森林; 双向长短期记忆网络; 黑翅鸢优化算法; 知识蒸馏
Keywords:
power forecasting; random forest; bidirectional long short-term memory; black-winged kite algorithm; knowledge distillation
分类号:
TM614;TP183
文献标志码:
A
摘要:
为提高超短期光伏发电功率的预测精度,缩短预测时间,提出了一种基于BKA-BiLSTM-KD 的超短期光伏发电功率预测方法。首先,采用K-近邻算法(KNN)清洗异常数据以提升样本质量,并基于随机森林(RF)进行特征选择与组合增强;其次,提出黑翅鸢优化算法(BKA)优化双向长短期记忆网络(BiLSTM)预测模型,并对该预测模型进行知识蒸馏(KD),得到计算时间更短的学生模型;最后,以第17届“中国电机工程学会杯”比赛数据集和华北某光伏场站实际功率为算例样本,分别采用BiLSTM、BKA-BiLSTM、WOA-BiLSTM 和BKA-BiLSTM-KD共4种方法5种模型进行预测。预测试结果表明,与BKA-BiLSTM 模型相比,所提方法的决定系数R2 提升了2.7%以上,预测耗时缩短超过了67.8%,能够为光伏场站生产规划和电网调度决策提供高效、精准的数据支撑。
Abstract:
To improve the accuracy and shorten the time consumption of ultra-short-term photovoltaic (PV) power forecasting,a method based on BKA-BiLSTM-KD is proposed.First,the K-Nearest Neighbors (KNN) algorithm is employed to clean abnormal data and enhance sample quality,and the random forest (RF)is utilized for feature selection and combination augmentation.Second,the black-winged kite algorithm (BKA) is used to optimize the prediction model of Bidirectional Long Short-Term Memory (BiLSTM),and the knowledge distillation (KD) is applied to this model to obtain a more computationally efficient student model.Finally,taking the data set from the 17th “The Electrician Mathematical Contest in Modeling” competition and actual power data of a PV station in North China as case samples,four methods with five models,namely BiLSTM,BKA-BiLSTM,WOA-BiLSTM and BKA-BiLSTM-KD,are used for prediction respectively.The prediction results demonstrate that,compared with the BKA-BiLSTM model,the proposed method improves the coefficient of determination (R2) by more than 2.7% and reduces the prediction time consumption by over 67.8%.It can provide efficient and accurate data support for PV station production planning and power grid dispatching decisions.

参考文献/References:

[1] 茅庭瑞,李培强.基于改进VMD 和CSABO-TCN-BiGRU 的中短期光伏功率预测[J].太阳能学报,2025,46(12):168-177.

[2] 刘林君,李琼,封力,等.局部遮挡下光伏组件功率损失预测模型[J].太阳能学报,2025,46(12):243-253.
[3] 李新福,李斌,王广溢,等.基于相似日聚类和SSA-LSTM 的光伏组件积灰损失预测[J].太阳能学报,2025,46(12):195-203.
[4] 顾锦,刘宇,顾雷涛,等.基于气象分类的超短时光伏发电预测模型[J].太阳能学报,2025,46(12):263-270.
[5] 龚莺飞,鲁宗相,乔颖,等.光伏功率预测技术[J].电力系统自动化,2016,40(4):140-151.
[6] 王育飞,付玉超,孙路,等.基于混沌RBF神经网络的光伏发电功率超短期预测模型[J].电网技术,2018,42(4):1110-1116.
[7] Yan Xinyue,Zhang Haipeng,Li Ran,et al.A cost-oriented score evaluation index to align probabilistic forecasting and uncertain decision-making in power systems[ J].IEEE Transactions on Power Systems,2025,40(2):1981-1984.
[8] 王育飞,汪弦哲,薛花,等.基于BO-SVM 和ISO 改进的ENN光伏功率超短期预测模型[J].太阳能学报,2025,46(6):280-288.
[9] 李练兵,高国强,陶鹏,等.基于SFLA 和MSISSAANFIS的超短期光伏功率动态预测方法[J].太阳能学报,2024,45(10):326-335.
[10] 赵洪山,孙承妍,温开云,等.基于有向图卷积循环网络的分布式光伏出力超短期预测方法[J].太阳能学报,2024,45(8):281-288.
[11] 李芬,于淏,孙改平,等.基于优化变分模态分解的光伏功率超短期区间预测方法[J].太阳能学报,2024,45(8):367-376.
[12] 赵洪山,孙承妍,温开云.基于多气象类型加权和改进高斯混合模型的光伏出力超短期概率预测[J].太阳能学报,2024,45(7):567-576.
[13] 汪鸿,朱正甲,陈建华,等.基于人工智能技术与物理方法结合的新能源功率预测研究[J].高电压技术,2023,49(增刊1):111-117.
[14] 史洪岩,李尚达,潘多涛,等.基于VMD-COMTBO-BiLSTM 光伏阵列故障诊断研究[J].太阳能学报,2026,47(2):18-29.
[15] 焦丕华,蔡旭,王乐乐,等.考虑数据分解和进化捕食策略的BiLSTM 短期光伏发电功率预测[J].太阳能学报,2024,45(2):435-442.
[16] 王玉芳,程培浩,闫明.混合增强黑翅鸢优化算法及其应用[J].计算机科学与探索,2026,20(1):99-121.
[17] 王秋杰,姬晨旭,谭洪,等.基于改进开关函数的弹性配电网复杂多重故障区段定位方法[J].电力自动化设备,2025,45(10):84-91.
[18] 周金涛,何山,王维庆,等.基于黑翅鸢算法和VMD的短期风电功率预测[J].太阳能学报,2025,46(12):762-773.
[19] 王昱,朱芝孺.基于改进K 近邻规则的数据库营销分析[J].统计与决策,2018,34(19):175-178.
[20] 逯静,张燕茹,王瑞.基于BWO 优化VMD 和TCN-BiGRU 的短期风电功率预测[J].工程科学与技术,2025,57(3):31-41.
[21] 王瑞,李哲,逯静.基于融合聚类和BKA-VMD-TCN-BiLSTM 的短期光伏功率预测[J].电子科技大学学报,2025,54(4):592-603.
[22] Wang Jun,Wang Wenchuan,Hu Xiaoxue,et al.Blackwinged kite algorithm:a nature-inspired metaheuristic for solving benchmark functions and engineering problems[ J].Artificial Intelligence Review,2024,57:98.

相似文献/References:

[1]李森文,张 伟,李纯宇,等.基于 SSA-LSTM 的海上风电功率预测[J].机械与电子,2022,(06):22.
　LI Senwen,ZHANG Wei,LI Chunyu,et al.Power Prediction of Offshore Wind Farm Based on SSA-LSTM[J].Machinery & Electronics,2022,(08):22.

备注/Memo

备注/Memo:
收稿日期:2026-04-29
基金项目:河南省自然科学基金青年基金项目(252300420471)
作者简介:侯鹏飞 (1991-),男,甘肃崇信人,博士,讲师,研究方向为新能源发电与并网、电力设备状态监测与故障诊断;母宇宸(2006-),男,河南南阳人,研究方向为新能源发电与并网。
更新日期/Last Update: 2026-08-28