作者：李琼林, 刘书铭, 郑晨, 王毅, 张博, 代双寅, 唐钰政

作者单位：国网河南省电力公司电力科学研究院，河南 郑州 450052

关键词：电压暂降;发生时间预测;隐马尔可夫模型;监测数据特性

摘要：

通过电能质量监测系统(power quality monitoring system, PQMS)中蕴含的电网历史故障变化、趋势等重要信息，对未来电压暂降进行预测，可为用户和电网公司合理规划生产，避免经济损失提供有力帮助。该文提出一种基于隐马尔可夫模型的电压暂降发生时间(occurrence time of voltage sag, OTVS)预测方法。首先对电压暂降发生时间的变量可预测性、数据冗余性、事件混沌性进行分析，揭示电压暂降监测数据特性；然后针对这三种特性，提出基于模糊C-均值聚类算法(fuzzy C-means algorithm, FCMA)和赤池信息准则(Akaike information criterion, AIC)的电压暂降历史状态识别与划分方法，以区间型变量刻画监测数据中的历史变化信息；建立考虑暂降历史变化信息和电网扰动变化信息的隐马尔可夫模型，实现对未来电压暂降的预测。最后，利用中部某省10个监测点的历史数据进行验证，所提方法的预测准确率最高可达92.85%，所提方法的预测性能较其他典型预测方法约高5%~30%。

Prediction of occurrence time of voltage sag based on hidden Markov model
LI Qionglin, LIU Shuming, ZHENG Chen, WANG Yi, ZHANG Bo, DAI Shuangyin, TANG Yuzheng
State Grid Henan Electric Power Research Institute, Zhengzhou 450052, China
Abstract: The prediction of future voltage sag based on the important information such as the changes and trends of power grid faults in power quality monitoring system (PQMS) can help users and utilities to plan production reasonably and avoid economic losses. In this paper, a prediction method of occurrence time of voltage sag (OTVS) based on hidden Markov model is proposed. Firstly, the predictability of variable, redundancy of data and chaos of voltage sag are analyzed, and the characteristics of monitoring data of voltage sag are revealed. Then, a new method based on fuzzy C-means algorithm (FCMA) and Akaike information criterion (AIC) is proposed to identify and divide voltage sag historical states, and the historical information in monitoring data is characterized by interval variables. A hidden Markov model considering the historical variation information of sag and the change information of power grid disturbance is established, and realize the prediction of future voltage sag. Finally, the historical data from 10 monitors in a central province are used for verification. The prediction accuracy of the proposed method is up to 92.85%, and the prediction performance of the proposed method is about 5%-30% higher than that of other typical prediction methods.
Keywords: voltage sag;prediction of occurrence time;hidden Markov model;characteristics of monitoring data
2023, 49(4):106-113  收稿日期: 2021-08-16;收到修改稿日期: 2021-11-25
基金项目: 国家电网公司总部科技项目资助 (5400-202124153A- 0-0-00)
作者简介: 李琼林(1980-),男,湖北英山县人,教授级高级工程师,博士,主要从事电能质量分析与控制技术工作
参考文献
[1] 肖先勇, 杨洪耕, 陈武, 等. 敏感设备电压暂降敏感度的模糊随机评估[J]. 中国电机工程学报, 2009, 29(34): 90-95
[2] XU Y, WU Y, ZHANG M, et al. Sensitivity of programmable logic controllers to voltage sags[J]. IEEE Transaction on Power Delivery, 2019, 34(1): 2-10
[3] 戴志辉, 李川, 焦彦军. IIDG低压穿越模型及其在配网故障分析中的应用[J]. 电力系统及其自动化学报, 2018, 30(7): 16-23
[4] ATALIK T, ÇADIRCI I, DEMIRCI T, et al. Multipurpose platform for power system monitoring and analysis with sample grid applications[J]. IEEE Transaction on Instrumentation and Measurement, 2014, 63(3): 566-582
[5] 符金伟, 关石磊, 左思然, 等. 基于GM(1, 1)和人工蜂群的配电测试系统状态估计[J]. 中国测试, 2018, 44(7): 7-12+41
[6] 岳勇, 陈雯婷, 聂伟. LMD和ARMA组合风速预测方法[J]. 中国测试, 2020, 46(8): 126-130
[7] 徐严军, 吴蒙, 白佳灵, 等. 多特征提取与深度学习关口计量装置异常事件识别方法[J]. 中国测试, 2021, 47(5): 104-111
[8] 丁泽俊, 刘平, 欧阳森, 等. 电能质量预测与预警机制及其应用[J]. 电力系统及其自动化学报, 2015, 27(10): 87-92
[9] 肖斐, 艾芊. 考虑天气因素的区域配网扰动事件预估[J]. 电网技术, 2018, 42(4): 1132-1139
[10] 汪颖, 谢佳妮, 邓凌峰, 等. 基于典型波形特征与改进DBSCAN的电压暂降同源识别方法[J]. 电力系统自动化, 2021, 45(11): 126-135
[11] Electromagnetic compatibility (EMC)-Part 4-30: Testing and measurement techniques-power quality measurement methods: IEC 61000-4-30: 2015[S], 2015.
[12] WANG Y, DENG L, BOLLEN M, et al. Calculation of the point-on-wave for voltage dips in three-phase systems[J]. IEEE Transaction on Power Delivery, 2020, 35(4): 2068-2079
[13] 徐文远, 雍静. 电力扰动数据分析学——电能质量监测数据的新应用[J]. 中国电机工程学报, 2013, 33(19): 93-101+15
[14] 张欣雨, 李茹, 王晋宇. 改进最大Lyapunov指数的瓦斯时间序列预测研究[J]. 计算机与现代化, 2014(10): 119-122
[15] 刘丁源, 裴磊, 魏炯, 等. 基于模糊C均值算法的电力变压器聚类分析[J]. 能源与环保, 2021, 43(6): 224-228
[16] 陈春俊, 聂锡成, 张洁. 基于ARMA模型的高速列车隧道压力波预测研究[J]. 中国测试, 2013, 39(6): 5-9