基于物理信息神经网络的套管电容芯体等效热导率反演方法

doi:10.19595/j.cnki.1000-6753.tces.251735

Abstract
Figure/Table
References (29)
Related Citation (10)

Download: PDF (8838 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract

Dry-type resin-impregnated paper (RIP) bushings have been widely applied in converter transformers and high-voltage direct-current (HVDC) systems owing to their oil-free insulation, strong moisture resistance, and low maintenance requirements. However, the absence of internal oil circulation causes heat to accumulate within the bushing core, which increases the thermal stress on the composite insulation structure and accelerates insulation aging. Temperature-field modeling of RIP bushing are performed by finite element methods (FEM) and the results of temperature distribution simulations are highly dependent on the selected model parameters. Because the thermal conductivity of the multilayer capacitive screen cannot be directly characterized using conventional measurement techniques, determining its equivalent thermal properties remains a major challenge for accurate thermal simulation and temperature-rise analysis. This study addresses this problem by proposing a data-efficient inversion method, identifying the thermal conductivity of the capacitive screen using sparse temperature information.
A physics-informed neural network (PINN) based inversion framework was developed to estimate the equivalent radial and axial thermal conductivities of the capacitive screen. A two-dimensional axisymmetric heat-conduction model was constructed using domain partitioning to represent the conductor, capacitive core, bottom flange and outer sheath. The multilayer screen region was homogenized into an anisotropic medium, enabling the model to retain directional thermal-transfer characteristics while reducing geometric and computational complexity. The steady-state heat-conduction partial differential equation (PDE) and interface conduction conditions across different materials were embedded directly into the total loss function of PINN method. This formulation restricts the solution space to physically consistent states and enables the network to infer internal parameters directly from sparse temperature measurements. To enhance PINN’s training stability, an adaptive gradient-balancing strategy was employed, which dynamically regulated the weights of the PDE, boundary, and observation losses and strengthen convergence robustness.
To evaluate the effectiveness of the proposed PINN method, a 110 kV RIP bushing temperature-rise platform was established. Under the consistency of other variables such as ambient temperature and humidity, steady-state temperature data of four different operating current levels (60%I_n, 80%I_n, I_n, and 110%I_n) at several typical positions were collected, including conductor rod, the connecting seat, bottom flange and the outer sheath. After reversion, the identified thermal conductivity parameters were imported into a FEM model for independent verification. The simulations replicated the measured temperature-rise characteristics with deviations held within ±2 K across all monitored points, indicating that the inferred parameters accurately reflect the thermal behavior of the composite capacitive screen. These results demonstrate that the proposed PINN-based inversion framework provides a reliable and data-efficient means for identifying effective thermal properties in multilayer insulation systems, offering a practical tool for improving thermal-field modeling and enhancing the design and evaluation of power-equipment insulation structures.
The results show that the proposed PINN inversion framework can effectively recover the anisotropic equivalent thermal conductivity of multilayer composite insulation from sparse obtainable temperature data. The method relies on only a small set of training data and retains high physical consistency and interpretability, making it possible to infer otherwise unmeasurable thermal conductivity from discrete temperature measurements. The framework can be extended to other power equipment, including transformer windings, reactor insulation cylinders, and dielectric components in gas-insulated switchgear (GIS). The proposed approach provides an effective pathway for improving thermal-field modeling accuracy, refining temperature-rise evaluation, and supporting future digital-twin applications for the thermal management of power equipment.

Key words： Physics-informed neural network thermal conductivity inversion resin impregnated paper bushing parameter identification temperature rise experiment

Received: 13 October 2025

PACS:

TM835.4

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors

Cite this article:

URL:

https://dgjsxb.ces-transaction.com/EN/10.19595/j.cnki.1000-6753.tces.251735 OR https://dgjsxb.ces-transaction.com/EN/Y2026/V41/I19/6753

[1] Alassi A, Bañales S, Ellabban O, et al.HVDC transmission: technology review, market trends and future outlook[J]. Renewable and Sustainable Energy Reviews, 2019, 112: 530-554.
[2] 刘杉, 高冲, 侯俊义, 等. 基于去离子水冷却±800 kV换流变压器阀侧套管的冷却效果和参数的影响[J]. 电工技术学报, 2024, 39(12): 3884-3894.
Liu Shan, Gao Chong, Hou Junyi, et al.Cooling effect and effect of parameters of the deionized water cooling based ERIP type bushing of ±800 kV converter transformer[J]. Transactions of China Electrotechnical Society, 2024, 39(12): 3884-3894.
[3] 陈铭, 刘轩东, 李原, 等. 采用频域介电谱法的高压干式套管屏间局部绝缘劣化诊断[J]. 电网技术, 2018, 42(7): 2359-2365.
Chen Ming, Liu Xuandong, Li Yuan, et al.Local inter-foil insulation deterioration diagnosis of RIP bushing based on frequency-domain dielectric spectroscopy method[J]. Power System Technology, 2018, 42(7): 2359-2365.
[4] 严嘉璇. 电热联合作用下胶浸纸材料绝缘劣化过程研究[D]. 北京: 华北电力大学, 2024.
Yan Jiaxuan.Research on insulation degradation process of resin-impregnated paper materials under the joint influence of electrcal and thermal factors[D]. Beijing: North China Electric Power University, 2024.
[5] 盛戈皞, 钱勇, 罗林根, 等. 面向新型电力系统的数字化电力设备关键技术及其发展趋势[J]. 高电压技术, 2023, 49(5): 1765-1778.
Sheng Gehao, Qian Yong, Luo Lingen, et al.Key technologies and development trends of digital power equipment for new type power system[J]. High Voltage Engineering, 2023, 49(5): 1765-1778.
[6] Jyothi N S, Ramu T S, Mandlik M.Temperature distribution in resin impregnated paper insulation for transformer bushings[J]. IEEE Transactions on Dielectrics and Electrical Insulation, 2010, 17(3): 931-938.
[7] 赵建利, 姚顺, 岳永刚, 等. 500kV SF6瓷质套管多工况仿真与故障分析[J]. 电工技术学报, 2021, 36(增刊2): 736-745.
Zhao Jianli, Yao Shun, Yue Yonggang, et al.Simulation and failure analysis of 500kV SF6 porcelain bushing under complicated working conditions[J]. Transactions of China Electrotechnical Society, 2021, 36(S2): 736-745.
[8] Akbari M, Mostafaei M, Rezaei-Zare A.Estimation of hot-spot heating in OIP transformer bushings due to geomagnetically induced current[J]. IEEE Transactions on Power Delivery, 2023, 38(2): 1277-1285.
[9] 刘亚东, 严英杰, 严波, 等. 电力装备内部状态反演重构研究框架与应用展望[J]. 高电压技术, 2022, 48(8): 2883-2896.
Liu Yadong, Yan Yingjie, Yan Bo, et al.Framework and application prospect of internal state inversion and reconstruction of power equipment[J]. High Voltage Engineering, 2022, 48(8): 2883-2896.
[10] 杨帆, 吴涛, 郝翰学, 等. 基于矩阵低秩分解的变压器绕组温度场降阶模型快速更新方法[J]. 电工技术学报, 2025, 40(24): 7846-7862.
Yang Fan, Wu Tao, Hao Hanxue, et al.Fast updating method for reduced-order model of transformer winding temperature field based on matrix low-rank decomposition[J]. Transactions of China Electrotechnical Society, 2025, 40(24): 7846-7862.
[11] 杨帆, 胡星宇, 王鹏博. 油浸式电力变压器温升计算场路耦合和降阶模型研究[J]. 电工技术学报, 2025, 40(13): 4071-4084.
Yang Fan, Hu Xingyu, Wang Pengbo.Study on temperature rise calculation and reduced-order model of oil-immersed transformer with field-circuit coupling[J]. Transactions of China Electrotechnical Society, 2025, 40(13): 4071-4084.
[12] 陈思, 盛戈皞, 张立静, 等. 基于绕组等效热参数的干式变压器温度场分布快速仿真方法[J]. 高电压技术, 2025, 51(1): 97-109.
Chen Si, Sheng Gehao, Zhang Lijing, et al.Fast simulation method of temperature field distribution for dry-type transformer based on equivalent thermal parameters of windings[J]. High Voltage Engineering, 2025, 51(1): 97-109.
[13] 张志宇, 寇家俊, 张重远, 等. 基于本征正交分解和克里金代理模型的油浸式变压器绕组温度场快速计算方法[J]. 电工技术学报, 2025, 40(7): 2215-2225.
Zhang Zhiyu, Kou Jiajun, Zhang Zhongyuan, et al.A fast calculation method for temperature field of oil immersed transformer windings based on proper orthogonal decomposition and Kriging surrogate model[J]. Transactions of China Electrotechnical Society, 2025, 40(7): 2215-2225.
[14] Karniadakis G E, Kevrekidis I G, Lu Lu, et al.Physics-informed machine learning[J]. Nature Reviews Physics, 2021, 3(6): 422-440.
[15] Cai Shengze, Wang Zhicheng, Wang Sifan, et al.Physics-informed neural networks for heat transfer problems[J]. Journal of Heat Transfer, 2021, 143(6): 060801.
[16] Tang Pengfei, Zhang Zhonghao, Tong Jie, et al.Predicting transformer temperature field based on physics-informed neural networks[J]. High Voltage, 2024, 9(4): 839-852.
[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] 赵陈琛, 张国钢, 刘洁, 等. 基于物理信息和迁移网络的高压断路器弹簧操动机构机械参数辨识及故障诊断[J]. 电工技术学报, 2026, 41(6): 2073-2085.
Zhao Chenchen, Zhang Guogang, Liu Jie, et al.Mechanical parameter identification and fault diagnosis of spring operating mechanism of high-voltage circuit breaker based on physical information and transfer network[J]. Transactions of China Electrotechnical Society, 2026, 41(6): 2073-2085.
[19] 余波, 甘子玉, 张森林, 等. 基于物理信息神经网络预测2D/3D非稳态温度场及热源[J]. 工程力学, 2025, 42(10): 12-24.
Yu Bo, Gan Ziyu, Zhang Senlin, et al.Prediction of 2D/3D unsteady-state temperature fields and heat sources upon the physics-informed neural networks[J]. Engineering Mechanics, 2025, 42(10): 12-24.
[20] Wang Tongsheng, Wang Zhiheng, Huang Zhu, et al.Multi-domain physics-informed neural network for solving heat conduction and conjugate natural convection with discontinuity of temperature gradient on interface[J]. Science China Technological Sciences, 2022, 65(10): 2442-2461.
[21] Bottou L.Large-scale machine learning with stochastic gradient descent[C]//Proceedings of 19th International Conference on Computational Statistics, Paris, France, 2010: 177-186.
[22] Loshchilov I, Hutter F. Fixing weight decay regularization in Adam[PP/OL]. arXiv(2019-01-04) [2025-09-01]. https://doi.org/10.48550/arXiv.1711.05101.
[23] Cheng S C, Vachon R I.The prediction of the thermal conductivity of two and three phase solid heterogeneous mixtures[J]. International Journal of Heat and Mass Transfer, 1969, 12(3): 249-264.
[24] International Electrotechnical Commission.Insulated bushings for alternating voltages above 1kV: IEC 60137:2017[S]. Geneva: IEC, 2017.
[25] Hodson T O.Root-mean-square error (RMSE) or mean absolute error (MAE): when to use them or not[J]. Geoscientific Model Development, 2022, 15(14): 5481-5487.
[26] DL/T 664—2016带电设备红外诊断应用规范[S].
[27] 黎帅, 张二龙. 变压器套管接线端部典型发热缺陷原因分析[J]. 电工技术, 2024(5): 171-173, 179.
Li Shuai, Zhang Erlong.Causal analysis of typical overheating faults of transformer bushing terminals[J]. Electric Engineering, 2024(5): 171-173, 179.
[28] 罗舜. 电力变压器套管将军帽发热故障的红外诊断分析[J]. 变压器, 2018, 55(1): 50-53.
Luo Shun.Infrared diagnosis analysis of power transformer bushing coupler heating[J]. Transformer, 2018, 55(1): 50-53.
[29] 刘玉安, 毛宝俊, 陈云龙, 等. 地震作用下换流变压器阀侧套管的温度场仿真与响应分析[J]. 高压电器, 2025, 61(4): 30-38.
Liu Yu’an Mao Baojun, Chen Yunlong, et al. Simulation and response analysis on temperature field of valve side bushing of converter transformer under seismic action[J]. High Voltage Apparatus, 2025, 61(4): 30-38.