基于VAE-WSVM的二次采样回路测试校核评估

doi:10.11930/j.issn.1004-9649.202405011

摘要/Abstract

摘要：

分析二次系统采样回路中存在的隐患，针对现有方法难以获取大数据量的缺陷，基于历史测量数据，采用变分自编码器实现数据增殖，接着基于增殖后的数据集训练支持向量机模型，利用灰狼优化算法改进模型重要参数，并对二次系统采样回路误差范围进行分类校核。实验结果表明：所提方法训练的分类模型评估效果良好。

关键词: 二次系统采样回路, 误差评估, 小样本

Abstract:

Analyze the hidden dangers existing in the sampling circuit of the secondary system, and in view of the defects that the existing methods are difficult to obtain large amount of data, based on the historical measurement data, the variational autoencoder is used to achieve data proliferation, and then the support vector machine model is trained on the proliferated data set, and the important parameters of the model are improved by using the grey wolf optimization algorithm, and the error range of the secondary system sampling is classified and checked. The experimental results show that the classification model trained by the proposed method has a good evaluation effect.

Key words: secondary system sampling loop, error evaluation, small sample size

尹相国, 刘慧娟, 张冠育, 毛玉荣, 龚世玉. 基于VAE-WSVM的二次采样回路测试校核评估[J]. 中国电力, 2025, 58(3): 162-167.

Xiangguo YIN, Huijuan LIU, Guanyu ZHANG, Yurong MAO, Shiyu GONG. Sampling Loop Test Check Evaluation of Secondary System based on VAE-WSVM Data Mining[J]. Electric Power, 2025, 58(3): 162-167.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202405011

https://www.electricpower.com.cn/CN/Y2025/V58/I3/162

图/表 8

图 1 变分自编码器的基本流程

Fig.1 The basic flow of VAE

图 2 多类核函数对比分析

Fig.2 Comparative analysis of multiple types of kernel functions

图 3 二次系统采样回路测试校核评估流程

Fig.3 Evaluation process for secondary system sampling loop

图 4 基于变分自编码器算法的数据增殖

Fig.4 Data multiplication based on VAE

图 5 模型关键参数寻优

Fig.5 Optimization of model key parameters

表 1 各样本构成情况

Table 1 The composition of each sample

采样回路误差状态	代号	样本数量	测试集	训练集
正常	N	1100	330	770
计量误差	M1	300	90	210
测量误差	M2	300	90	210
保护误差	L1	300	90	210

表 2 递增不平衡度的训练样本构成情况

Table 2 The composition of training samples with increasing unbalance degree

代号	训练样本构成情况
代号	训练集1	训练集2	训练集3	训练集4
N	100	200	400	770
M1	100	200	210	210
M2	100	200	210	210
L1	100	200	210	210

表 3 各评估方法结果

Table 3 Results of each evaluation method

评估方法	准确率/%
评估方法	训练集1	训练集2	训练集3	训练集4
统计量	75.17	75.08	75.40	75.71
简单神经网络	96.90	88.33	79.67	79.39
支持向量机	83.33	86.92	82.21	79.29
VAE-SVM	92.92	94.33	96.96	98.57
VAE-WSVM	95.83	97.08	98.71	99.52

参考文献 19

1	钟鸣, 陶军, 刘洵宇, 等. 智能变电站光纤虚实回路映射及故障诊断技术[J]. 中国电力, 2023, 56 (10): 171- 178.
	ZHONG Ming, TAO Jun, LIU Xunyu, et al. Smart substation optical fiber virtual and real loop mapping and fault diagnosis technology[J]. Electric Power, 2023, 56 (10): 171- 178.
2	易亚文, 赵静朴, 柳灿, 等. 基于改进ED及CM-BPNN算法的保护测量回路误差状态评估方法[J]. 中国电力, 2023, 56 (11): 143- 150.
	YI Yawen, ZHAO JingPu, LIU Can, et al. Error status evaluation method for protection measurement circuit based on improved ED and CM-BPNN algorithms[J]. Electric Power, 2023, 56 (11): 143- 150.
3	王亚飞, 张朋丰, 王讯, 等. 采用集群测控的智能变电站新型监控方案技术经济性分析[J]. 山东电力技术, 2023, 50 (4): 32- 40. DOI
	WANG Yafei, ZHANG Pengfeng, WANG Xun, et al. Technical and economic analysis of a novel supervision system for smart substation using cluster measurement and control devices[J]. Shandong Electric Power, 2023, 50 (4): 32- 40. DOI
4	张洁, 卢晓雄, 陈佳, 等. 基于阻抗测量的电压互感器二次回路压降测试技术研究[J]. 广东电力, 2023, 36 (9): 51- 59. DOI
	ZHANG Jie, LU Xiaoxiong, CHEN Jia, et al. Research on voltage transformer secondary circuit voltage drop measurement technology based on impedance measurement[J]. Guangdong Electric Power, 2023, 36 (9): 51- 59. DOI
5	王达, 高林, 张爱平, 等. 基于调控云数据分析的二次故障智能诊断系统设计与应用[J]. 山东电力技术, 2023, 50 (9): 73- 80.
	WANG Da, GAO Lin, ZHANG Aiping, et al. Design and application of secondary failure intelligent diagnosis system based on dispatching and control cloud data analysis[J]. Shandong Electric Power, 2023, 50 (9): 73- 80.
6	栗磊, 梁亚波, 赫嘉楠, 等. 基于差动电流相位差的和应涌流识别及其与内部故障的区分方法[J]. 电网与清洁能源, 2023, 39 (8): 64- 72. DOI
	LI Lei, LIANG Yabo, HE Jianan, et al. A method of identifying sympathetic inrush current based on phase difference of differential current and distinguishing it from internal faults[J]. Power System and Clean Energy, 2023, 39 (8): 64- 72. DOI
7	赵东生, 周原, 范亚洲, 等. 输电线路在线监测设备供能技术研究综述[J]. 广东电力, 2024, 37 (7): 107- 117.
	ZHAO Dongsheng, ZHOU Yuan, FAN Yazhou, et al. Overview of energy supply technology for online monitoring equipment of transmission lines[J]. Guangdong Electric Power, 2024, 37 (7): 107- 117.
8	郑翔, 杜奇伟, 阮黎翔, 等. 基于WOA-SVM的智能变电站二次系统故障参数映射模型[J]. 浙江电力, 2024, 43 (1): 36- 44.
	ZHENG Xiang, DU Qiwei, RUAN Lixiang, et al. A WOA-based fault parameter mapping model for the secondary systems of intelligent substations[J]. Zhejiang Electric Power, 2024, 43 (1): 36- 44.
9	陈辉, 黄林滨, 李朝兵, 等. 二次再热锅炉30%负荷下燃烧优化调整研究[J]. 电力科技与环保, 2023, 39 (2): 129- 137.
	CHEN Hui, HUANG Linbin, LI Chaobing, et al. Study on combustion optimization and adjustment of secondary reheat boiler at 30%load[J]. Electric Power Technology and Environmental Protection, 2023, 39 (2): 129- 137.
10	钟鸣, 翟寅, 付晓艺. 基于线性二次型积分调节器的风电机组功率载荷协同变桨控制[J]. 内蒙古电力技术, 2023, 41 (5): 81- 87.
	ZHONG Ming, ZHAI Yin, FU Xiaoyi. LQRI-based cooperative pitch control of wind turbine power and load[J]. Inner Mongolia Electric Power, 2023, 41 (5): 81- 87.
11	褚雪汝, 陈中, 吴聪颖, 等. 基于深度学习的电气二次图纸语义识别方法[J]. 浙江电力, 2023, 42 (8): 1- 11.
	CHU Xueru, CHEN Zhong, WU Congying, et al. Small target area extraction and semantic recognition method of electrical secondary drawings based on deep learning[J]. Zhejiang Electric Power, 2023, 42 (8): 1- 11.
12	高伟, 饶俊民, 全圣鑫, 等. 不均衡小样本下多特征优化选择的生命体触电故障识别方法[J]. 电工技术学报, 2024, 39 (7): 2060- 2071.
	GAO Wei, RAO Junmin, QUAN Shengxin, et al. Biological electric-shock fault identification method based on multi-feature optimization selection under unbalanced small sample[J]. Transactions of China Electrotechnical Society, 2024, 39 (7): 2060- 2071.
13	祁虹. 基于压缩感知理论的高压断路器故障诊断研究[D]. 兰州: 兰州交通大学, 2022.
	QI Hong. Research on fault diagnosis of high voltage circuit breaker based on compressed sensing theory[D]. Lanzhou: Lanzhou Jiatong University, 2022.
14	李煜, 毕卫红, 孙建成, 等. 紫外-可见吸收光谱结合化学计量学算法的水体总有机碳浓度快速检测[J]. 光谱学与光谱分析, 2024, 44 (3): 722- 730.
	LI Yu, BI Weihong, SUN Jiancheng, et al. Rapid detection of total organic carbon concentration in water using UV-vis absorption spectra combined with chemometric algorithms[J]. Spectroscopy and Spectral Analysis, 2024, 44 (3): 722- 730.
15	李丹, 梁云嫣, 缪书唯, 等. 基于高斯混合聚类和改进条件变分自编码的多风电场功率日场景生成方法[J]. 中国电力, 2024, 57 (12): 17- 29.
	LI Dan, LIANG Yunyan, MIAO Shuwei, et al. Daily power scenario generation method for multiple wind farms based on gaussian mixture clustering and improved conditional variational autoencoder[J]. Electric Power, 2024, 57 (12): 17- 29.
16	胥备, 刘桐. 基于改进高斯混合变分自编码器的半监督情感音乐生成[J]. 计算机科学, 2024, 51 (8): 281- 296.
	XU Bei, LIU Tong. Semi-supervised emotional music generation method based on improved Gaussian mixture variational autoencoders[J]. Computer Science, 2024, 51 (8): 281- 296.
17	王德鑫. 基于对比学习的图像生成系统的设计与实现[D]. 北京: 北京邮电大学, 2023.
	WANG Dexin. Design and realization of graphic generation system based on contrastive learning[D]. Beijing: Beijing University of Posts and Telecommunications, 2023.
18	徐恒辉, 姚杰, 周萍, 等. 基于运行数据的光伏电站状态评估方法研究[J]. 电力科技与环保, 2023, 39 (5): 450- 456.
	XU Henghui, YAO Jie, ZHOU Ping, et al. Research on photovoltaic power plant state evaluation based on operating data[J]. Electric Power Technology and Environmental Protection, 2023, 39 (5): 450- 456.
19	赵冬梅, 孙明伟, 宿梦月, 等. 基于改进SKNet-SVM的网络安全态势评估[J]. 应用科学学报, 2024, 42 (2): 334- 349. DOI
	ZHAO Dongmei, SUN Mingwei, SU Mengyue, et al. Network security situation assessment based on improved SKNet-SVM[J]. Journal of Applied Sciences, 2024, 42 (2): 334- 349. DOI

[1]	贾巍, 黄裕春. 基于小样本数据差分扩容的微电网负荷预测方法[J]. 中国电力, 2023, 56(8): 151-156,165.
[2]	易亚文, 江传宾, 龚世玉, 秦小元, 赵静朴, 李振兴. 基于因子分析与统计学技术的保护测量回路误差评估[J]. 中国电力, 2023, 56(12): 183-190.
[3]	王达梦, 马志勇, 柳亦兵, 滕伟. 小故障样本条件下的风电机组分层贝叶斯可靠性模型[J]. 中国电力, 2019, 52(12): 97-104.