基于自适应神经模糊推理与反馈数据驱动储能变流器电容老化指征参数监测方法

doi:10.19595/j.cnki.1000-6753.tces.251366

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (3417 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract Capacitors are critical components in power conversion systems (PCS) for energy storage, and their lifespan and reliability are key factors that limit the overall performance and safe operation of the converter. Electrolyte drying and dielectric aging, influenced by factors such as temperature and ripple current, lead to a decrease in capacitance and an increase in equivalent series resistance (ESR). Accurate online monitoring of these key state parameters is a prerequisite for aging assessment and life prediction. Traditional offline methods require disassembly and cannot reflect real operating conditions. Existing online methods, such as recursive least squares (RLS) or model-based observers, suffer from model dependence and noise sensitivity. Pure neural network approaches lack interpretability.
This paper proposes a dual-stage method that integrates an adaptive neuro-fuzzy inference system (ANFIS) for parameter identification and a data-driven voltage-error feedback mechanism for validation and refinement. Firstly, based on the physical structure and application scenario, a simplified capacitor equivalent model, consisting of a series connection of capacitance (C) and ESR (R), is established. The dynamic voltage-current relationship forms the basis for parameter identification. Secondly, an ANFIS model is constructed for online identification. A hybrid learning algorithm combining least squares estimation (LSE) for consequent parameters and gradient descent for premise parameters is employed. The physical parameters (C and R) are extracted by analyzing the network output's gradient with respect to the input current. Thirdly, a novel data-driven evaluation mechanism based on voltage-differential-error propagation is introduced. The parameters identified by ANFIS are used to reconstruct the capacitor voltage. A differential-error model is derived to establish a direct mapping between parameter errors and voltage differential errors. By substituting multiple data points into this model, a set of lines is generated in the parameter error plane. The density distribution of their intersections indicates the most probable parameter errors, enabling quantitative assessment and refinement. Finally, simulation models in Matlab/Simulink and a 2 kW experimental platform are built. Gaussian white noise is added to simulate measurement inaccuracies, and a Kalman filter is used for signal pre-processing.
Simulation results under ideal conditions demonstrate high accuracy, with ANFIS identification errors of 0.87% for capacitance and 0.83% for ESR. Under simulated noisy conditions (1% voltage noise, 10% current noise), the identification errors increase but remain within acceptable limits. The intersection density plot becomes more dispersed, reflecting the increased uncertainties, yet still indicates the most probable error region. The evaluation offsets for capacitance and ESR are within 5% and 25%, respectively, demonstrating robustness. Experimental results show that the ANFIS method identifies parameters that are close to those obtained from offline LCR bridge measurements. The subsequent error evaluation provides predicted ranges for the actual online parameters and the corresponding loss tangent (tanδ). The evaluated tanδ range includes the offline- measured value, confirming the predictive capability. The proposed ANFIS-based approach achieved higher identification accuracy than the RLS method.
The following conclusions can be drawn. (1) The ANFIS-based identification model successfully integrates the interpretability of fuzzy systems with the learning ability of neural networks for high-precision online identification of capacitor parameters. (2) The voltage differential error propagation mechanism addresses the challenge of lacking a direct online validation benchmark for quantitative assessment of parameter errors. (3) By retaining physical meaning, the method overcomes the “black-box” limitation of pure neural networks, enhancing engineering applicability.

Key words： Filter capacitors power converters parameter estimation adaptive neuro-fuzzy inference system (ANFIS)

Received: 06 August 2025

PACS:	TM46
	TM53

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	Tian Yanjun
	Zhai Haizhi
	Qu Zhangshun
	Liu Qing

Cite this article:

Tian Yanjun,Zhai Haizhi,Qu Zhangshun等. Monitoring Method for Capacitor Aging Indication Parameters in Energy Storage Converters Based on Adaptive Neuro-Fuzzy Inference and Feedback Data-Driven Approaches[J]. Transactions of China Electrotechnical Society, 2026, 41(14): 4960-4973.

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.251366 OR https://dgjsxb.ces-transaction.com/EN/Y2026/V41/I14/4960

[1] 冯艺萱, 边晓燕, 陈雯, 等. 新型电力系统灵活性资源成本回收机制分析及挑战[J]. 电工技术学报, 2025, 40(21): 7013-7028.
Feng Yixuan, Bian Xiaoyan, Chen Wen, et al.Analysis and challenges of new power system flexibility resource cost recovery mechanisms[J]. Transactions of China Electrotechnical Society, 2025, 40(21): 7013-7028.
[2] 宋少鹏, 田艳军, 李欣, 等. 电压源型并网变流器状态监测与故障诊断方法综述[J]. 电气工程学报, 2024, 19(3): 1-13.
Song Shaopeng, Tian Yanjun, Li Xin, et al.Review of condition monitoring and fault diagnosis methods for grid-connected VSC[J]. Journal of Electrical Engin-eering, 2024, 19(3): 1-13.
[3] 张云霄, 张佳瑞, 蔡英健, 等. 多物理场下薄膜电容器用聚丙烯材料劣化及诊断方法研究进展[J]. 电工技术学报, 2025, 40(15): 4905-4926.
Zhang Yunxiao, Zhang Jiarui, Cai Yingjian, et al.Research progress on degradation and diagnostic methods of polypropylene materials for film capa-citors under multiple physical fields[J]. Transactions of China Electrotechnical Society, 2025, 40(15): 4905-4926.
[4] Zhao Zhaoyang, Davari P, Lu Weiguo, et al.Online DC-link capacitance monitoring for digital-controlled Boost PFC converters without additional sampling devices[J]. IEEE Transactions on Industrial Elec-tronics, 2023, 70(1): 907-920.
[5] 王要强, 李锐, 陈天锦, 等. 基于长周期暂态信号分析的电机驱动变换器直流母线电容在线状态监测[J]. 电力系统保护与控制, 2024, 52(16): 120-131.
Wang Yaoqiang, Li Rui, Chen Tianjin, et al.Online monitoring of a DC-link capacitor for a motor drive converter based on long period transient signal analysis[J]. Power System Protection and Control, 2024, 52(16): 120-131.
[6] Zhao Zhaoyang, Davari P, Wang Yaoqiang, et al.Online capacitance monitoring for DC/DC Boost converters based on low-sampling-rate approach[J]. IEEE Journal of Emerging and Selected Topics in Power Electronics, 2022, 10(5): 5192-5204.
[7] 李林卓, 周国伟, 孙启贺, 等. 特高压换流站交流滤波电容组件电热特性分析[J]. 河北电力技术, 2024, 43(6): 30-35.
Li Linzhuo, Zhou Guowei, Sun Qihe, et al.Analysis of electrothermal characteristics of AC filter capacitor components in ultra voltage converter station[J]. Hebei Electric Power, 2024, 43(6): 30-35.
[8] Amaral A M R, Marques Cardoso A J. An economic offline technique for estimating the equivalent circuit of aluminum electrolytic capacitors[J]. IEEE Transac-tions on Instrumentation and Measurement, 2008, 57(12): 2697-2710.
[9] Amaral A M R, Marques Cardoso A J. A simple offline technique for evaluating the condition of aluminum-electrolytic-capacitors[J]. IEEE Transac-tions on Industrial Electronics, 2009, 56(8): 3230-3237.
[10] 罗兵, 李君洛, 王少杰, 等. 接枝聚丙烯电容薄膜的高温储能特性[J]. 中国电机工程学报, 2024, 44(9): 3371-3377.
Luo Bing, Li Junluo, Wang Shaojie, et al.High temperature energy storage characteristics of grafted polypropylene capacitor film[J]. Proceedings of the CSEE, 2024, 44(9): 3371-3377.
[11] 卢伟国, 曹琪, 张婷婷, 等. 基于改进谐波注入的单相Boost功率因数校正变换器直流母线电容参数在线检测[J]. 电工技术学报, 2025, 40(8): 2601-2614.
Lu Weiguo, Cao Qi, Zhang Tingting, et al.Online capacitance detection of DC-link capacitor in single-phase Boost PFC converter based on improved harmonic injection[J]. Transactions of China Elec-trotechnical Society, 2025, 40(8): 2601-2614.
[12] 谷苗, 夏超英, 田聪颖. 基于综合型卡尔曼滤波的锂离子电池荷电状态估算[J]. 电工技术学报, 2019, 34(2): 419-426.
Gu Miao, Xia Chaoying, Tian Congying.Li-ion battery state of charge estimation based on com-prehensive Kalman filter[J]. Transactions of China Electrotechnical Society, 2019, 34(2): 419-426.
[13] 赖纪东, 崔玉妹, 董玮, 等. 基于遗传因子递推最小二乘法微网逆变器线路阻抗辨识[J]. 太阳能学报, 2025, 46(2): 293-298.
Lai Jidong, Cui Yumei, Dong Wei, et al.Line impedance identification of microgrid inverters based on genetic factor-based recursive least squares method[J]. Acta Energiae Solaris Sinica, 2025, 46(2): 293-298.
[14] 范兴明, 封浩, 张鑫. 最小二乘算法优化及其在锂离子电池参数辨识中的应用[J]. 电工技术学报, 2024, 39(5): 1577-1588.
Fan Xingming, Feng Hao, Zhang Xin.Optimization of least squares method and its application in parameter identification of lithium-ion battery model[J]. Transac-tions of China Electrotechnical Society, 2024, 39(5): 1577-1588.
[15] 谢文超, 赵延明, 方紫微, 等. 带可变遗忘因子递推最小二乘法的超级电容模组等效模型参数辨识方法[J]. 电工技术学报, 2021, 36(5): 996-1005.
Xie Wenchao, Zhao Yanming, Fang Ziwei, et al.Variable forgetting factor recursive least squales based parameter identification method for the equivalent circuit model of the supercapacitor cell module[J]. Transactions of China Electrotechnical Society, 2021, 36(5): 996-1005.
[16] Pang Hui, Chen Kaiqiang, Geng Yuanfei, et al.Accurate capacity and remaining useful life predi-ction of lithium-ion batteries based on improved particle swarm optimization and particle filter[J]. Energy, 2024, 293: 130555.
[17] 向超群, 尹雪瑶, 伍珣, 等. 基于物理信息神经网络的牵引变流器直流支撑电容参数辨识方法[J]. 电工技术学报, 2024, 39(15): 4654-4667.
Xiang Chaoqun, Yin Xueyao, Wu Xun, et al.Parameter identification of DC-link capacitor in traction converter based on physical information neural network[J]. Transactions of China Electro-technical Society, 2024, 39(15): 4654-4667.
[18] 吴祎, 王友仁, 陈妤, 等. 基于BPNN的电力电子电路参数性故障诊断方法[J]. 电子测量与仪器学报, 2012, 26(5): 458-462.
Wu Yi, Wang Youren, Chen Yu, et al.Parametric fault diagnosis method of power electronic circuit based on BPNN[J]. Journal of Electronic Measurement and Instrument, 2012, 26(5): 458-462.
[19] 张菲菲, 谷正气, 张沙, 等. 基于EKF-BP网络的矿用自卸车轮胎材料参数辨识[J]. 振动与冲击, 2016, 35(17): 71-76.
Zhang Feifei, Gu Zhengqi, Zhang Sha, et al.Tire material parametric identification of mining dump trucks based on EKF-BP[J]. Journal of Vibration and Shock, 2016, 35(17): 71-76.
[20] Soualhi A, Makdessi M, German R, et al.Heath monitoring of capacitors and supercapacitors using the neo-fuzzy neural approach[J]. IEEE Transactions on Industrial Informatics, 2018, 14(1): 24-34.
[21] Kamel T, Biletskiy Y, Chang Liuchen.Capacitor aging detection for the DC filters in the power electronic converters using ANFIS algorithm[C]// 2015 IEEE 28th Canadian Conference on Electrical and Computer Engineering (CCECE), Halifax, NS, Canada, 2015: 663-668.
[22] 陈龙, 易琼洋, 贲彤, 等. 全局优化算法在Preisach磁滞模型参数辨识问题中的应用与性能对比[J]. 电工技术学报, 2021, 36(12): 2585-2593, 2606.
Chen Long, Yi Qiongyang, Ben Tong, et al.Application and performance comparison of global optimization algorithms in the parameter identi-fication problems of the preisach hysteresis model[J]. Transactions of China Electrotechnical Society, 2021, 36(12): 2585-2593, 2606.
[23] 徐乐, 邓忠伟, 谢翌, 等. 锂离子电池电化学-热耦合模型对比研究[J]. 机械工程学报, 2022, 58(22): 304-320.
Xu Le, Deng Zhongwei, Xie Yi, et al.Comparative study of electrochemical-thermal models for Li-ion batteries[J]. Journal of Mechanical Engineering, 2022, 58(22): 304-320.
[24] Wei Xing, Jin Long, Yao Bo, et al.Application-driven data acquisition for condition monitoring of power semiconductors in traction inverter applications[J]. IEEE Transactions on Transportation Electrification, 2024, 10(3): 6812-6824.
[25] 张超, 杜博超, 崔淑梅, 等. 一种为实现电动汽车薄膜母线电容准在线故障诊断的容值辨识方法[J]. 电工技术学报, 2020, 35(增刊1): 220-227.
Zhang Chao, Du Bochao, Cui Shumei, et al.Capacitance identification method for realizing quasi-online fault diagnosis of thin film bus capacitance of electric vehicle[J]. Transactions of China Electro-technical Society, 2020, 35(S1): 220-227.
[26] Yao Bo, Wei Xing, Zhang Yichi, et al.Accelerated degradation testing and failure mechanism analysis of metallized film capacitors for AC filtering[J]. IEEE Transactions on Power Electronics, 2024, 39(5): 6256-6270.
[27] 于希宁, 程锋章, 朱丽玲, 等. 基于T-S模型的自适应神经模糊推理系统及其在热工过程建模中的应用[J]. 中国电机工程学报, 2006, 26(15): 78-82.
Yu Xining, Cheng Fengzhang, Zhu Liling, et al.ANFIS modeling based on T-S model and its appli-cation for thermal process[J]. Proceedings of the CSEE, 2006, 26(15): 78-82.
[28] 刘彤宇, 陆起涌, 李旦, 等. 随机工况下基于改进ANFIS的锂电池容量衰减实时估计[J]. 仪器仪表学报, 2024, 45(5): 218-226.
Liu Tongyu, Lu Qiyong, Li Dan, et al.Real-time estimation of lithium battery capacity degradation based on an improved neural fuzzy inference system under random operating conditions[J]. Chinese Journal of Scientific Instrument, 2024, 45(5): 218-226.