基于改进VMD-XGBoost-BiLSTM组合模型的光伏发电异常检测

doi:10.19678/j.issn.1000-3428.0067583

摘要/Abstract

摘要：

光伏发电是我国大力发展的重要新能源发电形式，其异常检测是为系统运维决策提供依据的重要环节。由组件老化、故障或不良因素造成的光伏系统运行状态异常将直接影响发电效率和能力，进而会对系统安全性和经济效益造成影响。然而，现有检测方法存在识别异常类型不全面、对标注数据数量依赖性强、更新模型成本高、对噪声和测量误差敏感等局限性以及不适合大规模推广部署的缺点。为解决这一问题，提出一种基于历史发电量以及气象监测数据的光伏发电异常检测方法。利用基于异常值去除和相关性分析的预处理步骤去除原始数据中的噪声并筛选最佳特征。通过变分模态分解（VMD）将数据分解成多个模态分量以提取光伏发电量的周期和非周期特征。构建改进VMD-XGBoost-BiLSTM组合模型，利用自适应赋权、Attention机制和改进鲸鱼优化算法的特点完成光伏发电量常态预测。在此基础上，通过与实际测量值进行对比，利用设定的规则进行异常判断。实验结果表明，该方法相较于单一BiLSTM和XGBoost模型平均误差下降幅度超过20%，其中约15.67%的性能提升得益于所提改进措施。

关键词: 光伏发电异常检测, 神经网络, 变分模态分解, 注意力机制, 改进鲸鱼优化算法

Abstract:

Photovoltaic is a vital and rapidly developing form of renewable energy generated in China, and anomaly detection is a crucial reference for making decisions on its system operation and maintenance. The abnormal operation status of a photovoltaic system caused by component aging, failure, and other adverse factors affect the power generation efficiency and capacity, which impact the system security and benefits. However, existing solutions have limitations, such as identification of limited types of abnormalities, significantly relying on labeled amount of data, additional costs for model updating, sensitivity to noise and measurement errors, and shortcomings unsuitable for large-scale promotion and deployment. A photovoltaic anomaly detection method based on historical power generation and meteorological monitoring data is proposed to solve this problem. Initially, the preprocessing steps based on spike removal and correlation analysis are used for original data denoising and feature refinement. Next, the Variational Mode Decomposition(VMD) is used to decompose the data into multiple Intrinsic Mode Function (IMF) to extract the periodic and nonperiodic characteristics of photovoltaic power generation. Subsequently, an improved VMD-XGBoost-BiLSTM hybrid model is established to precisely and stably predict normal photovoltaic power generation, benefiting from adaptive weighting, Attention mechanism, and improved Whale Optimization Algorithm (WOA). Finally, the prediction results are compared to actual data, and thus, abnormalities can be detected based on the proposed rules. The experimental results indicate that compared to a single BiLSTM and XGBoost model, this method reduces the average error by more than 20%, in which 15.67% is contributed by the series of improvement.

Key words: photovoltaic anomaly detection, neural network, Variational Mode Decomposition(VMD), attention mechanism, improved Whale Optimization Algorithm(WOA)

赵博超, 马嘉骏, 崔磊, 栾文鹏, 朱静. 基于改进VMD-XGBoost-BiLSTM组合模型的光伏发电异常检测[J]. 计算机工程, 2024, 50(3): 306-316.

Bochao ZHAO, Jiajun MA, Lei CUI, Wenpeng LUAN, Jing ZHU. Anomaly Detection for Photovoltaic Based on Improved VMD-XGBoost-BiLSTM Combination Model[J]. Computer Engineering, 2024, 50(3): 306-316.

https://www.ecice06.com/CN/Y2024/V50/I3/306

图/表 16

图1 原始数据示例

Fig.1 Original data example

图2 关联矩阵

Fig.2 Correlation matrix

图3 晴朗无云时光伏发电量示例

Fig.3 Example of photovoltaic generation for sunny days

图4 特征改进后的关联矩阵

Fig.4 Correlation matrix with improved features

图5 LSTM结构示意图

Fig.5 Schematic diagram of LSTM structure

图6 BiLSTM结构示意图

Fig.6 Schematic diagram of BiLSTM structure

图7 Attention机制结构

Fig.7 Attention mechanism structure

图8 IWOA优化流程

Fig.8 IWOA optimization procedure

图9 基于改进VMD-XGBoost-BiLSTM组合模型的光伏发电量异常检测方法流程

Fig.9 Procedure of photovoltaic anomaly detection method based on improved VMD-XGBoost-BiLSTM combination model

图10 VMD结果

Fig.10 VMD results

图11 不同天气下的对比结果

Fig.11 Comparison results under different weather conditions

图12 异常检测结果

Fig.12 Anomaly detection results

参考文献 30

1	王青, 孙頔, 张海霞, 等. 中国光伏行业2021年回顾与2022年展望. 电气时代, 2022,(5): 20- 28.
	WANG Q, SUN D, ZHANG H X, et al. Review of photovoltaic industry in China in 2021 and prospect for 2022. Electric Age, 2022,(5): 20- 28.
2	HONG Y Y, PULA R A. Methods of photovoltaic fault detection and classification: a review. Energy Reports, 2022, 8, 5898- 5929. doi: 10.1016/j.egyr.2022.04.043
3	LIVERA A, THERISTIS M, MAKRIDES G, et al. Recent advances in failure diagnosis techniques based on performance data analysis for grid-connected photovoltaic systems. Renewable Energy, 2019, 133, 126- 143. doi: 10.1016/j.renene.2018.09.101
4	TSANAKAS J A, HA L, BUERHOP C. Faults and infrared thermographic diagnosis in operating c-Si photovoltaic modules: a review of research and future challenges. Renewable and Sustainable Energy Reviews, 2016, 62, 695- 709. doi: 10.1016/j.rser.2016.04.079
5	ALIPPI C, NTALAMPIRAS S, ROVERI M. Model-free fault detection and isolation in large-scale cyber-physical systems. IEEE Transactions on Emerging Topics in Computational Intelligence, 2017, 1(1): 61- 71. doi: 10.1109/TETCI.2016.2641452
6	QU J, QIAN Z, PEI Y, et al. An unsupervised hourly weather status pattern recognition and blending fitting model for PV system fault detection. Applied Energy, 2022, 319, 119271. doi: 10.1016/j.apenergy.2022.119271
7	LI B, DELPHA C, DIALLO D, et al. Application of artificial neural networks to photovoltaic fault detection and diagnosis: a review. Renewable and Sustainable Energy Reviews, 2021, 138, 110512. doi: 10.1016/j.rser.2020.110512
8	DE BENEDETTI M, LEONARDI F, MESSINA F, et al. Anomaly detection and predictive maintenance for photovoltaic systems. Neurocomputing, 2018, 310, 59- 68. doi: 10.1016/j.neucom.2018.05.017
9	WANG K, QI X, LIU H. Photovoltaic power forecasting based LSTM-convolutional network. Energy, 2019, 189, 116225. doi: 10.1016/j.energy.2019.116225
10	CHOUDER A, SILVESTRE S. Automatic supervision and fault detection of PV systems based on power losses analysis. Energy Conversion and Management, 2010, 51(10): 1929- 1937. doi: 10.1016/j.enconman.2010.02.025
11	MIAO S, NING G, GU Y, et al. Markov chain model for solar farm generation and its application to generation performance evaluation. Journal of Cleaner Production, 2018, 186, 905- 917. doi: 10.1016/j.jclepro.2018.03.173
12	AGOUA X G, GIRARD R, KARINIOTAKIS G. Short-term spatio-temporal forecasting of photovoltaic power production. IEEE Transactions on Sustainable Energy, 2018, 9(2): 538- 546. doi: 10.1109/TSTE.2017.2747765
13	GAROUDJA E, HARROU F, SUN Y, et al. Statistical fault detection in photovoltaic systems. Solar Energy, 2017, 150, 485- 499. doi: 10.1016/j.solener.2017.04.043
14	ZHU X, DANG Y, DING S. Using a self-adaptive grey fractional weighted model to forecast Jiangsu's electricity consumption in China. Energy, 2020, 190, 116417. doi: 10.1016/j.energy.2019.116417
15	WANG J Y, QIAN Z, ZAREIPOUR H, et al. Performance assessment of photovoltaic modules based on daily energy generation estimation. Energy, 2018, 165, 1160- 1172. doi: 10.1016/j.energy.2018.10.047
16	POLO F A O, BERMEJO J F, FERNÁNDEZ J F G, et al. Failure mode prediction and energy forecasting of PV plants to assist dynamic maintenance tasks by ANN based models. Renewable Energy, 2015, 81, 227- 238. doi: 10.1016/j.renene.2015.03.023
17	HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proceedings of the Royal Society of London Series A: Mathematical, Physical and Engineering Sciences, 1998, 454(1971): 903- 995. doi: 10.1098/rspa.1998.0193
18	SMITH J S. The local mean decomposition and its application to EEG perception data. Journal of the Royal Society Interface, 2005, 2(5): 443- 454. doi: 10.1098/rsif.2005.0058
19	LIU D, NIU D, WANG H, et al. Short-term wind speed forecasting using wavelet transform and support vector machines optimized by genetic algorithm. Renewable Energy, 2014, 62, 592- 597. doi: 10.1016/j.renene.2013.08.011
20	NAZARI M, SAKHAEI S M. Successive variational mode decomposition. Signal Processing, 2020, 174, 107610. doi: 10.1016/j.sigpro.2020.107610
21	LIU H, MI X, LI Y. Smart multi-step deep learning model for wind speed forecasting based on variational mode decomposition, singular spectrum analysis, LSTM network and ELM. Energy Conversion and Management, 2018, 159, 54- 64. doi: 10.1016/j.enconman.2018.01.010
22	DRAGOMIRETSKIY K, ZOSSO D. Variational mode decomposition. IEEE Transactions on Signal Processing, 2014, 62(3): 531- 544. doi: 10.1109/TSP.2013.2288675
23	ZHAO Y, LI C, FU W, et al. A modified variational mode decomposition method based on envelope nesting and multi-criteria evaluation. Journal of Sound and Vibration, 2020, 468, 115099. doi: 10.1016/j.jsv.2019.115099
24	HOCHREITER S, SCHMIDHUBER J. Long short-term memory. Neural Computation, 1997, 9(8): 1735- 1780. doi: 10.1162/neco.1997.9.8.1735
25	SIAMI-NAMINI S, TAVAKOLI N, NAMIN A S. The performance of LSTM and BiLSTM in forecasting time series[C]//Proceedings of 2019 IEEE International Conference on Big Data. Washington D. C., USA: IEEE Press, 2019: 3285-3292.
26	CHEN T Q, GUESTRIN C. XGBoost: a scalable tree boosting system[C]//Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York, USA: ACM Press, 2016: 785-794.
27	NIU D, YU M, SUN L, et al. Short-term multi-energy load forecasting for integrated energy systems based on CNN-BiGRU optimized by attention mechanism. Applied Energy, 2022, 313, 118801. doi: 10.1016/j.apenergy.2022.118801
28	MIRJALILI S, LEWIS A. The whale optimization algorithm. Advances in Engineering Software, 2016, 95, 51- 67. doi: 10.1016/j.advengsoft.2016.01.008
29	ZHANG H, TANG L, YANG C, et al. Locating electric vehicle charging stations with service capacity using the improved whale optimization algorithm. Advanced Engineering Informatics, 2019, 41, 100901. doi: 10.1016/j.aei.2019.02.006
30	LI J H, GUO H. A hybrid whale optimization algorithm for plane block parallel blocking flowline scheduling optimization with deterioration effect in lean shipbuilding. IEEE Access, 2021, 9, 131893- 131905. doi: 10.1109/ACCESS.2021.3112742

[1]	王志浩, 钱沄涛. 基于Swin Transformer的双流遥感图像时空融合超分辨率重建[J]. 计算机工程, 2024, 50(9): 33-45.
[2]	李俊俊, 董建刚, 李坤. 基于Kubernetes的集群节能策略研究[J]. 计算机工程, 2024, 50(9): 82-91.
[3]	林畅, 郭伟, 任哲聪, 金海波. 基于Transformer的目标跟踪与分割统一算法[J]. 计算机工程, 2024, 50(9): 130-141.
[4]	李泽霖, 吕兆峰, 陈富强, 李克. 基于多跳信息融合的实体对齐模型[J]. 计算机工程, 2024, 50(9): 142-152.
[5]	王汝英, 马嘉骏, 董建强, 刘万龙, 张海涛, 尹凯, 赵博超. 基于MTS-BiGRU-DMHSA的工业负荷预测方法[J]. 计算机工程, 2024, 50(9): 169-178.
[6]	张鲁, 田春伟, 宋焕生, 刘侍刚. 用于低剂量CT图像去噪的多级双树复小波网络[J]. 计算机工程, 2024, 50(9): 266-275.
[7]	朱凯, 李理, 张彤, 江晟, 别一鸣. 基于Transformer的多阶段运动模糊图像修复网络[J]. 计算机工程, 2024, 50(9): 276-285.
[8]	张天鹏, 韩晶, 吕学强. 基于多任务学习的超分辨率辅助小目标检测[J]. 计算机工程, 2024, 50(9): 304-312.
[9]	郭敏, 张熙涵, 李阳. 融合注意力的教师互一致性半监督医学图像分割[J]. 计算机工程, 2024, 50(9): 313-323.
[10]	高煜宝, 文志诚. 基于注意力机制的双路解码器图像去噪方法[J]. 计算机工程, 2024, 50(9): 324-332.
[11]	曾钰琦, 刘博, 钟柏昌, 钟瑾. 智慧教育下基于改进YOLOv8的学生课堂行为检测算法[J]. 计算机工程, 2024, 50(9): 344-355.
[12]	饶日昕, 王怡文, 曾砺志, 童心恬, 赵海涛. 面向废旧电缆检测的轻量化网络模型[J]. 计算机工程, 2024, 50(8): 22-30.
[13]	李华昱, 张智康, 闫阳, 岳阳. 基于知识图谱增强的领域多模态实体识别[J]. 计算机工程, 2024, 50(8): 31-39.
[14]	王蕾, 党时鹏, 潘丰. 基于卷积神经网络的隐匿性旁路预测模型[J]. 计算机工程, 2024, 50(8): 40-49.
[15]	陈瀚, 赵春蕾, 蒋昊达, 王春东. 基于融合模型与语义网络的App用户意图识别研究[J]. 计算机工程, 2024, 50(8): 50-63.

选择文件类型/文献管理软件名称

选择包含的内容