基于多特征优选机制与OOB优化的随机森林光伏阵列故障诊断方法

骆守康1, 温克欢1, 张昱波1, 孙晓佳1, 李 扬1, 孙文静1, 李天风2

分布式能源 ›› 0

PDF(1616 KB)
PDF(1616 KB)
分布式能源 ›› 0 DOI: 10.16513/j.2096-2185.DE.26100426

基于多特征优选机制与OOB优化的随机森林光伏阵列故障诊断方法

  • 骆守康1,温克欢1,张昱波1,孙晓佳1,李 扬1,孙文静1,李天风2*
作者信息 +

A Random Forest Fault Diagnosis Method for Photovoltaic Arrays Based on Multi-feature Optimization Mechanism and OOB Optimization

  • LUO Shoukang1, WEN Kehuan1, ZHANG Yubo1, SUN Xiaojia1, LI Yang1, SUN Wenjing1, LI Tianfeng2*
Author information +
文章历史 +

摘要

针对分布式光伏阵列故障频发、传统随机森林 (random forest, RF) 诊断方法准确率与泛化能力不足的问题,提出一种融合多特征优选与袋外 (out-of-bag, OOB) 误差动态优化的 RF 故障诊断方法。首先,基于光伏电池单二极管等效电路,在 MATLAB/Simulink 中构建光伏阵列仿真模型,模拟短路、开路、阴影遮挡、老化及旁路二极管失效等 5 种典型故障,并提取电压、电流、功率等多维故障特征。进而,引入 OOB 误差作为模型训练过程的动态评估指标,实时监控并优化基学习器集成策略,以提升收敛性能与泛化能力;同时,采用“置换重要性初筛+主成分分析二次降维”的两阶段多特征优选方法,从高维特征中筛选出关键故障特征子集,增强模型鲁棒性。实验结果表明,所提 OOB-RF 模型在测试集上的故障诊断准确率达 95.44%,较传统 RF 法提升 8.12 个百分点,过拟合指数显著降低 51.26%,在分类精度与稳定性方面均表现更优,为分布式光伏系统的智能运维提供了有效解决方案。

Abstract

To address the frequent faults in distributed photovoltaic (PV) arrays and the limitations of traditional random forest (RF) methods in diagnostic accuracy and generalization capability, this paper proposes a fault diagnosis method based
on  RF  that  integrates  multi-feature  optimization  and  dynamic  out-of-bag  (OOB)  error  optimization.  First,  a  PV  array simulation model is constructed in MATLAB/Simulink based on the single-diode equivalent circuit of a PV cell. The model simulates five typical fault types, including short circuit, open circuit, partial shading, aging degradation, and bypass diode failure, and extracts multi-dimensional fault features covering voltage, current, and power etc. Furthermore, the OOB error is introduced as a dynamic evaluation metric during model training to monitor and optimize the ensemble strategy of base learners  in  real  time,  thereby  improving  convergence  performance  and  generalization  capability.  Simultaneously,  a  twostage multi-feature optimization strategy—“permutation importance preliminary screening + principal component analysis secondary  dimensionality  reduction”—is  adopted  to  screen  the  most  critical  fault  feature  subset  from  the  original  highdimensional  features,  enhancing  model  robustness.  Experimental  results  demonstrate  that  the  proposed  OOB-RF  model achieves  a  fault  diagnosis  accuracy  of  95.44%  on  the  test  set,  which  is  8.12  percentage  points  higher  than  that  of  the traditional RF model, while the overfitting index is significantly reduced by 51.26%. The proposed method exhibits superior classification  accuracy  and  stability,  providing  an  effective  solution  for  the  intelligent  operation  and maintenance  of distributed PV systems.

关键词

光伏阵列 / 故障诊断 / 随机森林 (RF) 算法 / 袋外 (OOB) 误差 / 多特征优选

Key words

photovoltaic  / array / fault  / diagnosis / random  / forest  / (RF)  / algorithm / out-of-bag  / (OOB)  / error / multi-feature optimization

引用本文

导出引用
骆守康1, 温克欢1, 张昱波1, 孙晓佳1, 李 扬1, 孙文静1, 李天风2. 基于多特征优选机制与OOB优化的随机森林光伏阵列故障诊断方法[J]. 分布式能源, 0 https://doi.org/10.16513/j.2096-2185.DE.26100426.
LUO Shoukang1, WEN Kehuan1, ZHANG Yubo1, SUN Xiaojia1, LI Yang1, SUN Wenjing1, LI Tianfeng2. A Random Forest Fault Diagnosis Method for Photovoltaic Arrays Based on Multi-feature Optimization Mechanism and OOB Optimization[J]. Distributed Energy, 0 https://doi.org/10.16513/j.2096-2185.DE.26100426.
中图分类号: TK     51&emsp      

基金

中国南方电网有限责任公司创新项目 (SZKJXM20222146)

PDF(1616 KB)

Accesses

Citation

Detail

段落导航
相关文章

/

〈 〉