基于集成学习的时序自适应MMC子模块开路故障诊断方法  

Timing adaptive MMC submodule open-circuit fault diagnosis method based on integrated learning

作　　者：魏银图 张旸 温步瀛[1] 王怀远 WEI Yintu;ZHANG Yang;WEN Buying;WANG Huaiyuan(Key Laboratory of New Energy Generation and Power Conversion,College of Electrical Engineering and Automation,Fuzhou University,Fuzhou University,Fuzhou,Fujian 350108,China)

机构地区：[1]福州大学电气工程与自动化学院,新能源发电与电能变换重点实验室,福建福州350108

出　　处：《福州大学学报(自然科学版)》2025年第1期35-41,共7页Journal of Fuzhou University(Natural Science Edition)

基　　金：福建省自然科学基金资助项目(2022J01113)。

摘　　要：针对模块化多电平换流器(modular multilevel converter,MMC)子模块开路故障诊断中固定时刻诊断方法及时性和准确性难以平衡的问题,提出基于集成学习的时序自适应MMC子模块开路故障诊断方法.首先,分析MMC子模块的开路故障特性,选择子模块的电容电压作为故障监测和定位的故障特征参量.然后,基于代价敏感法构建两个具有相反诊断倾向性的集成模型,分别得到确定样本和不确定样本.接着,将当前阶段不确定样本交由下一时刻的模型继续诊断,直至诊断出所有样本的结果.最后,通过实验验证所提出诊断方法的有效性.结果表明,该方法能在更短的诊断周波内显著提高模型的诊断性能.Aiming at the problem that it is difficult to balance the timeliness and accuracy of the fixed-moment diagnosis method in the open-circuit fault diagnosis of modular multilevel converter(MMC)submodules,this paper proposes a timing-adaptive MMC submodule open-circuit fault diagnosis method based on integrated learning.First,the open-circuit fault characteristics of the MMC submodule are analyzed,and the capacitive voltage of the submodule is selected as the fault characteristic parameter for fault monitoring and localization.Then,two integrated models with opposite diagnostic tendencies are constructed based on the cost-sensitive method to obtain deterministic samples and uncertainty samples,respectively.Then,the uncertain samples at the current stage are handed over to the model at the next moment to continue the diagnosis until the results of all the samples are diagnosed.Finally,the experiment verifies the effectiveness of the proposed diagnostic method.The results show that the method can significantly improve the diagnostic performance of the model in a shorter diagnostic cycle.

关 键 词：模块化多电平换流器 子模块开路故障 故障诊断 集成学习 时序自适应 

分 类 号：TM464[电气工程—电器] TP181[自动化与计算机技术—控制理论与控制工程]