改进麻雀搜索算法优化VMD-CNN-BiLSTM的光伏阵列故障诊断研究

于洋; 郭栋; 徐艺; 王鹏伟; 高超; 隋永佳

doi:10.19912/j.0254-0096.tynxb.2024-2338

English Version Journal Login

您当前的位置：

首页 >

文章列表页 >

改进麻雀搜索算法优化VMD-CNN-BiLSTM的光伏阵列故障诊断研究

更新时间：2026-07-21

- 改进麻雀搜索算法优化VMD-CNN-BiLSTM的光伏阵列故障诊断研究
- Vol. 47, Issue 5, Pages: 627-638(2026)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.19912/j.0254-0096.tynxb.2024-2338
  CLC： TM615
- Online First：04 June 2026，
  
  Published：2026
- 稿件说明：
移动端阅览
于洋, 郭栋, 徐艺, et al. 改进麻雀搜索算法优化VMD-CNN-BiLSTM的光伏阵列故障诊断研究[J]. 2026, 47(5): 627-638.
DOI：

于洋, 郭栋, 徐艺, et al. 改进麻雀搜索算法优化VMD-CNN-BiLSTM的光伏阵列故障诊断研究[J]. 2026, 47(5): 627-638. DOI： 10.19912/j.0254-0096.tynxb.2024-2338.

摘要

为提高光伏阵列故障诊断的精度与可靠性

并实现故障类型的准确识别

提出一种基于改进麻雀搜索算法（ASFSSA）优化变分模态分解（VMD）与卷积神经网络-双向长短期记忆网络（CNN-BiLSTM）的故障诊断方法。首先

通过ASFSSA算法优化VMD的模态分量K和惩罚系数α

提升信号分解的性能和时间效率;然后

利用优化后的VMD对光伏阵列故障信号进行多层次分解

提取高辨识度的特征分量;最后

将这些特征输入CNN-BiLSTM模型

进行深度特征学习和故障类型分类。实验结果表明

所提方法能有效提取光伏阵列的故障特征

故障诊断的平均精度达到98.97%

在实际光伏电站实测数据集上的诊断准确率达到99.30%

优于其他对比模型

证实了所提方法的可行性和有效性。

Abstract

To address the challenges in precise fault identification and diagnostic reliability enhancement for photovoltaic （PV） arrays

this study develops an integrated fault diagnosis framework incorporating Adaptive Sparrow Search Algorithm （ASFSSA）-optimized Variational Mode Decomposition （VMD） with a hybrid Convolutional Neural Network-Bidirectional Long Short-Term Memory （CNN-BiLSTM） architecture. The framework comprises three principal phases： Initially

ASFSSA-driven parameter optimization is implemented for VMD's critical parameters （modal components and penalty coefficient ）

enhancing computational efficiency and decomposition precision. Subsequently

the optimized VMD algorithm performs multiscale decomposition of PV array fault signals to selectively extract discriminative feature components. Ultimately

these processed features are fed into the CNN-BiLSTM network for hierarchical feature abstraction and multi-class fault pattern recognition. Comprehensive experimental evaluations demonstrate the framework's superior performance

achieving 98.97% mean diagnostic accuracy on benchmark datasets and 99.30% accuracy on operational PV plant fault data

surpassing benchmark models in both controlled and real-world scenarios. These quantitative outcomes substantiate the technical viability and diagnostic effectiveness of the proposed intelligent fault diagnosis system.

关键词

Keywords

references

肖先勇, 郑子萱. “双碳”目标下新能源为主体的新型电力系统:贡献、关键技术与挑战[J]. 工程科学与技术, 2022, 54(1): 47-59.

任惠, 夏静, 卢锦玲, 等. 基于红外图像和改进MobileNet-V3的光伏组件故障诊断方法[J]. 太阳能学报, 2023, 44(8): 238-245.

WAN L, CHEN Y, LI H,et al.Rolling-element bearing fault diagnosis using improved LeNet-5 network[J]. Sensors, 2020, 20(6): 1693.

ZHANG Q, DENG L.An intelligent fault diagnosis method of rolling bearings based on short-time Fourier transform and convolutional neural network[J]. Journal of failure analysis and prevention, 2023, 23(2): 795-811.

ETANA B B, DAWUD A A, MALENGIER B,et al.Discrete wavelet transform based processing of embroidered textile-electrode electromyography signal acquired with load and pressure effect[J]. Journal of industrial textiles, 2024, 54: 15280837241232449.

RIOS R A, MELLO R F.Applying empirical mode decomposition and mutual information to separate stochastic and deterministic influences embedded in signals[J]. Signal processing, 2016, 118: 159-176.

GEETHA K, HOTA M K.Microseismic signal denoising based on variational mode decomposition with adaptive non-local means filtering[J]. Pure and applied geophysics, 2023, 180(11): 3709-3731.

李军, 江水, 徐启胜, 等. 基于VMD分解与卷积神经网络的液压泵故障诊断[J]. 锻压装备与制造技术, 2022, 57(3): 82-87.

曹沂风. 基于支持向量机的风轮不平衡故障诊断方法研究[J]. 太阳能学报, 2024, 45(8): 613-620.

柯希成, 刘永豪, 赵南洋, 等. 基于变分时频分解与随机森林的柴油机气门故障诊断方法[J]. 北京化工大学学报(自然科学版), 2024, 51(4): 99-106.

杨帅, 曾文伟, 杨凌云, 等. 基于GOA-SVM的光伏阵列故障诊断方法研究[J]. 电力科学与技术学报, 2024, 39(5): 172-180.

XU G, LIU M, JIANG Z, et al.Bearing fault diagnosis method based on deep convolutional neural network and random forest ensemble learning[J]. Sensors, 2019, 19(5): 1088.

李斌, 高鹏, 郭自强. 改进蜣螂算法优化LSTM的光伏阵列故障诊断[J]. 电力系统及其自动化学报, 2024, 36(8): 70-78.

BAKIR H, KUZHIPPALLIL F A, MERABET A.Automatic detection of deteriorated photovoltaic modules using IRT images and deep learning (CNN,LSTM) strategies[J]. Engineering failure analysis, 2023, 146: 107132.

刘行行, 帕孜来·马合木提, 程志江, 等. 基于XGBoost的光伏阵列故障诊断方法研究[J]. 电子测量技术, 2023, 46(12): 8-14.

贾亚飞, 朱永利, 王刘旺. 基于VMD和Wigner-Ville分布的局放信号时频分析[J]. 系统仿真学报, 2018, 30(2): 569-578.

刘沛津, 付雪峰, 孙浩峰, 等. 一种融合CNN与Transformer的高鲁棒性目标跟踪算[J]. 系统仿真学报, 2024, 36(8): 1854-1868.

GREFF K, SRIVASTAVA R K, KOUTNÍK J,et al. LSTM: a search space odyssey[J]. IEEE transactions on neural networks and learning systems, 2016, 28(10): 2222-2232.

王英杰, 朱景建, 龚智强, 等. 基于DAE-BiLSTM-CNN的滚动轴承故障诊断方法[J]. 机械设计, 2024, 41(11): 123-129.

武文栋, 施保华, 郑传良, 等. 基于改进麻雀搜索算法优化RBF神经网络的光伏阵列故障诊[J]. 智慧电力, 2023, 51(2): 77-83.

OUYANG C, QIU Y, ZHU D.Adaptive spiral flying sparrow search algorithm[J]. Scientific programming, 2021, 2021(1): 6505253.

逯帆. 基于SABO优化VMD-WTD-SVM的滚动轴承故障诊断模型[J]. 制造技术与机床, 2024(7): 32-39.

周娟利, 王城宇. 基于IGWO-MCKD-ROMP的轴承微弱故障信号诊断方法研究[J]. 机械传动, 2024, 48(9): 160-166.

LAZZARETTI A E, COSTA C H, RODRIGUES M P, et al.A monitoring system for online fault detection and classification in photovoltaic plants[J]. Sensors, 2020, 20(17): 4688.

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

DUAL-MODE DECOMPOSITION CNN-LSTM INTEGRATED SHORT-TERM WIND SPEED FORECASTING MODEL

基于OVMD-CNN-BiLSTM的短期光伏发电预测

基于相似日聚类和SSA-LSTM的光伏组件积灰损失预测

ULTRA-SHORT-TERM WIND SPEED PREDICTION BASED ON SAMPLE ENTROPY-BASED DUAL DECOMPOSITION AND SSA-LSTM

ULTRA-SHORT-TERM WIND POWER PREDICTION BASED ON ADAPTIVE QUADRATIC MODE DECOMPOSITION AND CNN-BiLSTM

Related Author

Bi Guihong

Zhao Xin

Li Lu

Chen Shilong

Chen Chenpeng

梁富源

万艳妮

杨国华

Related Institution

School of Electric Power Engineering, Kunming University of Science and Technology

Faulty of Automation, Huaiyin Institute of Technology, Huaian

Key Laboratory of Thermo-Fluid Science & Engineering of MOE, Xi’an Jiaotong University

National Computer Integrated Manufacturing System （CIMS） Engineering Research Center, Tsinghua University

School of Electrical Engineering, Xinjiang University

AI问答

⁰