基于Adam优化的改进PSO-RBF牵引变压器故障诊断研究*

doi:10.11985/2023.04.023

摘要/Abstract

摘要：

牵引变压器作为铁路供电的重要设备，承担着电气化铁路电力输送和电压转换的重要作用，其运行情况决定了整个牵引供电系统的安全稳定运行。由于牵引变压器运行环境恶劣，运行时同时受“电-磁-力-热”等多种内外应力，而且负载特性复杂，使得牵引变压器的故障诊断难度增大。对牵引变压器的故障诊断研究，不仅可以在牵引变压器发生故障前及时进行预警，而且可以为牵引变压器的检修提供充分的理论支撑，从而提高铁路供电系统的安全性和可靠性。把基于经典动量概念的Adam优化算法与PSO算法相结合，提出了一种基于油中溶解气体的牵引变压器故障诊断的新方法。首先构建PSO-RBF牵引变压器故障诊断模型，通过加速因子非线性搭配的仿真，采用非线性指数递减搭配来提高粒子群的寻优能力。以某铁路局管辖内牵引变压器的故障数据来进行仿真分析，诊断结果验证了IPSO-RBF-Adam牵引变压器故障诊断模型的诊断率和稳定性优于PSO-RBF模型。

关键词: 牵引变压器, 故障诊断, RBF神经网络, PSO算法

Abstract:

Traction transformer, as an important equipment of railway power supply, plays an important role in power transmission and voltage conversion of electrified railway, and its operation determines the safety and stability operation of the whole traction power supply system. Due to the harsh operating environment of the traction transformer, it is also subjected to various internal and external stresses such as “electric-magnetic-force-thermal”, and complex load characteristics, which makes the difficulty of fault diagnosis of the traction transformer increased. The research on the fault diagnosis of traction transformers can not only give timely warning before the failure of traction transformer, but also can provide sufficient theoretical support for the overhaul of traction transformer, so as to improve the safety and efficiency of railway power supply. A new method for traction transformer fault diagnosis based on dissolved gas in oil is proposed by combining Adam optimization algorithm based on classical momentum concept with PSO algorithm. Firstly, a PSO-RBF traction transformer fault diagnosis model is constructed, through the simulation experiment of nonlinear acceleration factor collocation, the nonlinear exponential decreasing collocation is used to improve the optimization ability of particle swarm optimization. The IPSO-RBF-ADAM traction transformer fault diagnosis model is better than PSO-RBF model in diagnosis accuracy and stability by simulation analysis of traction transformer fault data under the jurisdiction of a railway bureau.

Key words: Traction transformer, fault diagnosis, RBF neural network, particle swarm optimization algorithm

中图分类号:

U226

邵宁宁, 王英. 基于Adam优化的改进PSO-RBF牵引变压器故障诊断研究^*[J]. 电气工程学报, 2023, 18(4): 209-216.

SHAO Ningning, WANG Ying. Research on PSO-RBF Traction Transformer Fault Diagnosis Based on Adam Optimization[J]. Journal of Electrical Engineering, 2023, 18(4): 209-216.

图/表 17

表1

表2

四种测试函数"

函数名称	测试函数
Sphere函数	$f 1 = ∑ i = 1 d x i 2$
Rosenbrock函数	$f 2 = ∑ i = 1 d [100 (x i + 1 − x i 2) 2 + (x i − 1) 2]$
Griewank函数	$f 3 = 1 4 000 ∑ i = 1 d x i 2 − ∏ i = 1 d cos x i i + 1$
Rastrigrin函数	$f 4 = ∑ i = 1 d [x i 2 − 10 cos (2 π x i) + 1 0]$

表2

图1

图2

图3

图4

表3

图5

图6

表4

表5

图7

图8

图9

图10

图11

表6

参考文献 22

[1]	李波. 智能牵引变压器故障诊断技术研究[J]. 铁道工程学报, 2020, 37(8):71-76.
	LI Bo. Research on fault diagnosis technology of intelligent traction transformer[J]. Journal of Railway Engineering Society, 2020, 37(8):71-76.
[2]	范贤浩, 刘捷丰, 张镱议, 等. 融合频域介电谱及支持向量机的变压器油浸纸绝缘老化状态评估[J]. 电工技术学报, 2021, 36(10):2161-2168.
	FAN Xianhao, LIU Jiefeng, ZHANG Yiyi, et al. Evaluation of aging state of oil-immersed paper insulation of transformer by integrating frequency domain dielectric spectrum and support vector machine[J]. Transactions of China Electrotechnical Society, 2021, 36(10):2161-2168.
[3]	梁永亮, 郭汉琮, 薛永端. 基于特征气体关联特征的变压器故障诊断方法[J]. 高电压技术, 2019, 45(2):386-392.
	LIANG Yongliang, GUO Hancong, XUE Yongduan. Transformer fault diagnosis method based on characteristic gas correlation characteristics[J]. High Voltage Engineering, 2019, 45(2):386-392.
[4]	郭林, 唐晶, 唐黎哲, 等. 一种基于改进BP神经网络的变压器故障诊断方法[J]. 控制与信息技术, 2021(5):71-77.
	GUO Lin, TANG Jing, TANG Lizhe, et al. A transformer fault diagnosis method based on improved BP neural network[J]. Control and Information Technology, 2021(5):71-77.
[5]	王保义, 杨韵洁, 张少敏. 改进BP神经网络的SVM变压器故障诊断[J]. 电测与仪表, 2019, 56(19):53-58.
	WANG Baoyi, YANG Yunjie, ZHANG Shaomin. SVM transformer fault diagnosis based on improved BP neural network[J]. Electrical Measurement and Instrument, 2019, 56(19):53-58.
[6]	夏飞, 罗志疆, 张浩, 等. 混合神经网络在变压器故障诊断中的应用[J]. 电子测量与仪器学报, 2017, 31(1):118-124.
	XIA Fei, LUO Zhijiang, ZHANG Hao, et al. Application of hybrid neural network in transformer fault diagnosis[J]. Journal of Electronic Measurement and Instrument, 2017, 31(1):118-124.
[7]	张奎, 王建南, 王肖峰. 基于神经网络的变压器故障诊断[J]. 电子测量技术, 2017, 40(12):98-101.
	ZHANG Kui, WANG Jiannan, WANG Xiaofeng. Transformer fault diagnosis based on neural network[J]. Electronic Measurement Technology, 2017, 40(12):98-101.
[8]	KENNEDY J, EBERHART R C. Particle swarm optimization[C]//Proceeding of IEEE International Conference on Neural Networks. Piscataway:IEEE, 1885,1492-1498.
[9]	孟荣, 赵冀宁, 周通. 基于改进惯性权重粒子群算法的抢修小组快速调配策略[J]. 电网与清洁能源, 2021, 37(7):17-24.
	MENG Rong, ZHAO Jining, ZHOU Tong. Rapid deployment strategy based on improved inertial weight particle swarm algorithm[J]. Power System and Clean Energy, 2021, 37(7):17-24.
[10]	朱明星, 张德龙. RBF网络基函数中心选取算法的研究[J]. 安徽大学学报, 2000(1):72-78.
	ZHU Mingxing, ZHANG Delong. Study of RBF network[J]. Journal of Anhui University, 2000(1):72-78.
[11]	郭小艳. 改进粒子群优化RBF网络在变压器故障诊断中的应用[D]. 西安: 西安建筑科技大学, 2019.
	GUO Xiaoyan. Application of improved particle swarm optimization RBF network in transformer fault diagnosis[D]. Xi’an: Xi’an University of Architecture and Technology, 2019.
[12]	KINDMA D, BA J. A method for stochastic optimization[C]// Proceeding of the 3rd International Conference on Learning Representations, April 14-16,2014, San Diego. USA: Workshop Track, 2015:1-13.
[13]	施恂山, 马宏忠, 张琳, 等. PSO改进RBPNN在变压器故障诊断中的应用[J]. 电力系统保护与控制, 2016, 44(17):39-44.
	SHI Xunshan, MA Hongzhong, ZHANG Lin, et al. PSO improved application of RBPNN in transformer fault diagnosis[J]. Power System Protection and Control, 2016, 44(17):39-44.
[14]	JIANG Yan, HAN Fei. A hybrid algorithm of adaptive particle swarm optimization based on adaptive moment estimation method[C]// International Conference on Intelligent Computing, March 10-12,2017, Bangkok. Huang DS: Bevilacqua V, 2017:7-20.
[15]	TRAPPEY A J C, TRAPPEY C V, MA L, et al. Intelligent engineering asset management system for power transformer maintenance decision supports under various operating conditions[J]. Computers & Industrial Engineering, 2015, 84:3-11. doi: 10.1016/j.cie.2014.12.033
[16]	李笑竹, 陈志军, 樊小朝, 等. 基于ACS-SA文化基因算法的BP神经网络变压器故障诊断[J]. 高压电器, 2018, 54(2):134-146.
	LI Xiaozhu, CHEN Zhijun, FAN Xiaochao, et al. Fault diagnosis of BP neural network transformer based on ACS-SA culture gene algorithm[J]. High Voltage Apparatus, 2018, 54(2):134-146.
[17]	肖云波, 范菁, 张宜, 等. 基于改进粒子群算法与油中溶解气体的变压器故障诊断的研究[J]. 电子测量技术, 2021, 44(18):122-128.
	XIAO Yunbo, FAN Jing, ZHANG Yi, et al. Based on improved particle swarm algorithm and transformer fault diagnosis of dissolved gas in oil[J]. Electronic Measurement Technology, 2021, 44(18):122-128.
[18]	倪雪松. 基于动态阈值的牵引变压器故障诊断技术研究[D]. 成都: 西南交通大学, 2016.
	NI Xuesong. Research on traction transformer failure diagnosis technology based on dynamic thresholding[D]. Chengdu: Southwest Jiaotong University, 2016.
[19]	谢国民, 倪乐水, 曹媛. 基于VSRP与β-GWO-SVM的变压器故障辨识方法[J]. 高电压技术, 2021, 47(10):3635-3641.
	XIE Guomin, NI Leshui, CAO Yuan. Transformer fault identification method based on VSRP and β-GWO-SVM[J]. High Voltage Engineering, 2021, 47(10):3635-3641.
[20]	刘仲民, 翟玉晓, 张鑫, 等. 基于DBN-IFCM的变压器故障诊断方法[J]. 高电压技术, 2020, 46(12):4258-4265.
	LIU Zhongmin, ZHAI Yuxiao, ZHANG Xin, et al. Transformer fault diagnosis method based on DBN-IFCM[J]. High Voltage Engineering, 2020, 46(12):4258-4265.
[21]	王雨虹, 王志中, 付华, 等. 多策略改进麻雀算法与BiLSTM的变压器故障诊断研究[J]. 仪器仪表学报, 2022, 43(3):39-46.
	WANG Yuhong, WANG Zhizhong, FU Hua, et al. Study of improving sparrow algorithm and BiLSTM[J]. Chinese Journal of Scientific Instrument, 2022, 43(3):39-46.
[22]	谢乐, 衡熙丹, 刘洋, 等. 基于线性判别分析和分步机器学习的变压器故障诊断[J]. 浙江大学学报, 2020, 54(11):2266-2272.
	XIE Le, HENG Xidan, LIU Yang, et al. Transformer fault diagnosis based on linear discriminant analysis and step-wise machine learning[J]. Journal of Zhejiang University, 2020, 54(11):2266-2272.

故障类型	主要成分	次要成分
中低温过热	CH₄、C₂H₄	H₂、C₂H₆
高温过热	CH₄、C₂H₄、CO、CO₂	H₂、C₂H₆
高能放电	H₂、C₂H₂	CH₄、C₂H₄、C₂H₆
低能放电	H₂、C₂H₂	无
局部放电	H₂、CH₄、CO	C₂H₂、C₂H₆、CO₂

算法类型	相同迭代次数(500次)算法寻优结果
算法类型	f₁函数	f₂函数	f₃函数	f₄函数
PSO	1.75×10^-4	5.99×10^-2	3.09×10^-5	8.01×10^-2
IPSO1	4.51×10^-4	9.54×10^-2	1.24×10^-4	7.29×10^-2
IPSO2	1.42×10^-4	9.37×10^-2	8.93×10^-5	7.58×10^-2
IPSO3	1.31×10^-4	8.82×10^-2	2.37×10^-5	9.96×10^-2
IPSO4	4.73×10^-4	9.58×10^-3	7.65×10^-6	9.92×10^-3

参数	具体设置
学习率η	0.001
一阶矩阵衰减率µ	0.9
二阶矩阵衰减率v	0.999
稳定常数εε	1×10^-8
初始一阶矩估计m₀	0
初始二阶矩估计v₀	0

诊断方法	迭代次数	误差均方值	诊断错误	准确率(%)
PSO-RBF	231	0.050 6	8	91.111
WOA-RBF	165	0.034 9	5	94.444
IPSO-RBF	142	0.021 4	4	95.556
IPSO-RBF-Adam	101	0.004 7	2	97.778