基于相关性分析和长短时记忆网络的稳态电压质量指标预测

doi:10.12204/j.issn.1000-7229.2021.04.002

电力建设 ›› 2021, Vol. 42 ›› Issue (4): 9-16.doi: 10.12204/j.issn.1000-7229.2021.04.002

• 配电网智能化感知与供电质量提升关键技术 ·栏目主持唐巍教授· • 上一篇下一篇

基于相关性分析和长短时记忆网络的稳态电压质量指标预测

杨朝赟¹, 夏圣峰¹, 江南¹, 黄毅标¹, 李渴², 张逸²

1.国网福建省电力有限公司福州供电公司,福州市 350009
2.福州大学电气工程与自动化学院,福州市 350108

收稿日期:2020-08-03 出版日期:2021-04-01 发布日期:2021-03-30
作者简介:杨朝赟(1990),男,本科,工程师,主要研究方向为电能质量、智能配电网|夏圣峰(1985),男,硕士,高级工程师,主要从事电力生产经营管理工作|江南(1982),男,硕士,高级工程师,主要从事电力生产管理工作|黄毅标(1982),男,硕士,高级工程师,主要从事配电建设、运维工作|张逸(1984),男,博士,副教授,主要研究方向为电能质量、主动配电网及电力数据分析等
基金资助:
国网福建省电力有限公司科技项目“基于物联网技术的配电台区电能质量智能感知和治理决策技术研究及应用”(52131020011M)

Prediction of Steady-State Indices of Voltage Quality Based on Correlation Analysis and Long Short-Term Memory Network

YANG Chaoyun¹, XIA Shengfeng¹, JIANG Nan¹, HUANG Yibiao¹, LI Ke², ZHANG Yi²

1. Fuzhou Power Supply Company, State Grid Fujian Electric Power Co., Ltd., Fuzhou 350009, China
2. College of Electrical Engineering and Automation, Fuzhou University, Fuzhou 350108, China

Received:2020-08-03 Online:2021-04-01 Published:2021-03-30
Supported by:
Science and Technology Project “Research and application of intelligent perception and governance decision-making technology for power quality in power distribution station area based on Internet of Things technology”of State Grid Fujian Electric Power Co., Ltd.(52131020011M)

摘要/Abstract

摘要：

随着电力网络规模日益增大,多种负荷接入配电网带来诸多稳态电能质量问题。对配电台区电压质量监测数据进行预测,有助于掌握电能质量水平变化趋势,对电能质量预警和治理具有重要意义。为了有效分析稳态指标数据变化规律并提高电能质量水平,文章提出一种基于长短时记忆(long short-term memory,LSTM)网络的稳态电压质量指标预测方法,挖掘并利用不同时序数据的关联关系,优化稳态指标预测效果。首先,分析有功功率与电压质量指标的关联性,通过时序相关性匹配用户有功功率数据和实际稳态指标的时间序列特征;其次,用LSTM网络对筛选出的用户有功功率序列和稳态电压质量监测数据之间的关联关系进行建模;最后,利用LSTM模型对福建电网某个区域内稳态电压质量数据进行预测。通过实测数据验证,结合特定用户用电行为因素构建的预测模型,在用户日用电行为相对恒定和发生变化两种情况下,均能够提升稳态电压质量指标短期预测精度,且后者场景下长期预测效果更为显著。

关键词: 电压质量, 稳态指标, 相关性分析, 时间序列, 长短时记忆网络

Abstract:

With the increasing scale of power network, many steady-state power quality (PQ) problems are caused by multiple loads connected to the distribution network. The prediction of voltage quality according to the monitoring data in distribution stations is helpful to grasp the change trend of power quality level, which is of great significance for early warning and governance of power quality. In order to effectively analyze the change law of steady-state data and improve the level of power quality, a new method for voltage quality steady-state indices prediction is proposed. The method is based on long short-term memory (LSTM) network. It mines and uses the correlation of different time series data and optimizes the prediction effect of steady-state indices. Firstly, the correlation between active power and voltage quality steady-state indices is analyzed, and then their time series characteristics are matched by temporal correlation. Secondly, LSTM network is used to model correlation relationship between selected active power and actual steady-state monitoring data for voltage quality. Finally, the forecasting model is used to predict the voltage quality steady-state data of an area of Fujian power grid. According to experimental results, the DTW-LSTM model constructed by combining with the specific user's electricity consumption behavior can improve the short-term prediction accuracy of voltage quality steady-state indices under both relatively constant and changing daily electricity consumption behavior of users. Moreover, the long-term predictions are more effective in the second case.

Key words: voltage quality, steady-state indices, correlation analysis, time series, long short-term memory network

中图分类号:

TM73

杨朝赟, 夏圣峰, 江南, 黄毅标, 李渴, 张逸. 基于相关性分析和长短时记忆网络的稳态电压质量指标预测[J]. 电力建设, 2021, 42(4): 9-16.

YANG Chaoyun, XIA Shengfeng, JIANG Nan, HUANG Yibiao, LI Ke, ZHANG Yi. Prediction of Steady-State Indices of Voltage Quality Based on Correlation Analysis and Long Short-Term Memory Network[J]. ELECTRIC POWER CONSTRUCTION, 2021, 42(4): 9-16.

图/表 11

图1 不同用户同一天内有功功率曲线

Fig.1 Curves of active power of different users on the same day

图2 稳态电压质量指标曲线

Fig.2 Curves of voltage quality steady-state indices

图3 LSTM内部结构示意图

Fig.3 Schematic of LSTM cell structure

图4 电压总谐波畸变率时间序列

Fig.4 Time series of total harmonic distortion of the voltage

图5 多用户有功功率时间序列

Fig.5 Time-series of active power of the users

表1 THDu与有功功率的相关性测度

Table 1 Correlation measure between THDu and active power

表2 情况一中THDu的预测误差统计

Table 2 Forecast error statistics of THDu in the first case

表3 情况一中电压偏差的预测误差统计

Table 3 Forecast error statistics of voltage deviation in the first case

图6 某用户有功功率时间序列

Fig.6 Time series of active power of a power user

表4 情况二中THDu的预测误差统计

Table 4 Forecast error statistics of THDu in the second case

表5 情况二中电压偏差的预测误差统计

Table 5 Forecast error statistics of voltage deviation in the second case

参考文献 23

[1]	刘健, 张志华, 黄炜, 等. 分布式电源接入对配电网故障定位及电压质量的影响分析[J]. 电力建设, 2015,36(1):115-121.
	LIU Jian, ZHANG Zhihua, HUANG Wei, et al. Influence of distributed generation on fault location and voltage quality of distribution network[J]. Electric Power Construction, 2015,36(1):115-121.
[2]	贡晓旭, 金强, 王基, 等. 典型功能区分布式光伏接入对配电网建设改造的影响[J]. 中国电力, 2016,49(3):166-171.
	GONG Xiaoxu, JIN Qiang, WANG Ji, et al. Effects of distributed PV connected to grid on distribution network constructions in typical functional zones[J]. Electric Power, 2016,49(3):166-171.
[3]	ZHOU A H, ZHU L P, QIU H B, et al. Detection of abnormal trends in electrical data[C]// 2015 IEEE International Conference on Progress in Informatics and Computing (PIC). December 18-20, 2015, Nanjing, China. IEEE, 2015: 247-251.
[4]	林顺富, 汤继开, 汤波, 等. 典型电能质量稳态指标预测模型研究[J]. 电网技术, 2018,42(2):614-620.
	LIN Shunfu, TANG Jikai, TANG Bo, et al. Study on forecasting model of typical power quality steady state indices[J]. Power System Technology, 2018,42(2):614-620.
[5]	费思源. 大数据技术在配电网中的应用综述[J]. 中国电机工程学报, 2018,38(1):85-96,345.
	FEI Siyuan. Overview of application of big data technology in power distribution system[J]. Proceedings of the CSEE, 2018,38(1):85-96,345.
[6]	SALARVAND A, MIRZAEIAN B, MOALLEM M. Obtaining a quantitative index for power quality evaluation in competitive electricity market[J]. IET Generation, Transmission & Distribution, 2010,4(7):810. doi: 10.1049/iet-gtd.2009.0479 URL
[7]	PAN A Q, ZHOU J, ZHANG P, et al. Predicting of power quality steady state index based on chaotic theory using least squares support vector machine[J]. Energy and Power Engineering, 2017,9(4):713-724. doi: 10.4236/epe.2017.94B077 URL
[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.
	DING Zejun, LIU Ping, OUYANG Sen, et al. Mechanism of power quality forecast and early warning and their application[J]. Proceedings of the CSU-EPSA, 2015,27(10):87-92.
[9]	李奇. 电能质量变化趋势预测与预警方法研究及其应用[D]. 广州: 华南理工大学, 2015.
	LI Qi. Research on the method of power quality change trend forecasting and early-warning and its application[D]. Guangzhou: South China University of Technology, 2015.
[10]	刘启斌, 尹温硕, 胡卫华, 等. 基于LSTM算法的电力谐波监测数据预测[J]. 电力电容器与无功补偿, 2019,40(5):139-145.
	LIU Qibin, YIN Wenshuo, HU Weihua, et al. Prediction of power harmonic monitoring data based on LSTM algorithm[J]. Power Capacitor & Reactive Power Compensation, 2019,40(5):139-145.
[11]	VANTUCH T, MIŠÁK S, JEŽOWICZ T, et al. The power quality forecasting model for off-grid system supported by multiobjective optimization[J]. IEEE Transactions on Industrial Electronics, 2017,64(12):9507-9516. doi: 10.1109/TIE.2017.2711540 URL
[12]	BAI J J, GU W, YUAN X D, et al. Power quality prediction, early warning, and control for points of common coupling with wind farms[J]. Energies, 2015,8(9):9365-9382. doi: 10.3390/en8099365 URL
[13]	苏卫卫. 基于数据挖掘技术的稳态电能质量数据预测预警研究[D]. 北京: 华北电力大学, 2014.
	SU Weiwei. Research of steady-state power quality forecasting and early warning based on data mining technology[D]. Beijing: North China Electric Power University, 2014.
[14]	秦浩庭, 李群湛, 刘燕, 等. 基于Monte Carlo的电铁电能质量预测方法[J]. 电力系统保护与控制, 2011,39(13):64-70.
	QIN Haoting, LI Qunzhan, LIU Yan, et al. Prediction of the power quality caused by electrified railways based on Monte Carlo[J]. Power System Protection and Control, 2011,39(13):64-70.
[15]	张逸, 王攸然, 刘航, 等. 基于监测数据相关性分析的用户谐波责任划分方法[J]. 电力系统自动化, 2020,44(2):189-197.
	ZHANG Yi, WANG Youran, LIU Hang, et al. Determination method of user harmonic responsibility based on correlation analysis of monitoring data[J]. Automation of Electric Power Systems, 2020,44(2):189-197.
[16]	赵洪山, 寿佩瑶, 马利波. 低压台区缺失数据的张量补全方法[J]. 中国电机工程学报, 2020,40(22):7328-7336.
	ZHAO Hongshan, SHOU Peiyao, MA LiBo. A tensor completion method of missing data in transformer district[J]. Proceeding of the CSEE, 2020,40(22):7328-7336.
[17]	王芳, 顾伟, 袁晓冬, 等. 面向智能电网的新一代电能质量管理平台[J]. 电力自动化设备, 2012,32(7):134-139.
	WANG Fang, GU Wei, YUAN Xiaodong, et al. Power quality management platform for smart grid[J]. Electric Power Automation Equipment, 2012,32(7):134-139.
[18]	ZHANG Z, TAVENARD R, BAILLY A, et al. Dynamic time warping under limited warping path length[J]. Information Sciences, 2017,393:91-107. doi: 10.1016/j.ins.2017.02.018 URL
[19]	HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997,9(8):1735-1780. doi: 10.1162/neco.1997.9.8.1735 URL
[20]	GREFF K, SRIVASTAVA R K, KOUTNÍK J, et al. LSTM: A search space odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017,28(10):2222-2232. doi: 10.1109/TNNLS.2016.2582924 URL
[21]	徐京京. 基于聚类和神经网络的异常数据识别算法研究[D]. 北京: 华北电力大学, 2019.
	XU Jingjing. Research on anomaly data recognition algorithm based on clustering and neural network[D]. Beijing: North China Electric Power University, 2019.
[22]	TANG X L, DAI Y Y, WANG T, et al. Short-term power load forecasting based on multi-layer bidirectional recurrent neural network[J]. IET Generation, Transmission & Distribution, 2019,13(17):3847-3854. doi: 10.1049/gtd2.v13.17 URL
[23]	曹梦舟, 张艳. 基于卷积-长短期记忆网络的电能质量扰动分类[J]. 电力系统保护与控制, 2020,48(2):86-92.
	CAO Mengzhou, ZHANG Yan. Classification for power quality disturbances based on CNN-LSTM network[J]. Power System Protection and Control, 2020,48(2):86-92.

基于相关性分析和长短时记忆网络的稳态电压质量指标预测

Prediction of Steady-State Indices of Voltage Quality Based on Correlation Analysis and Long Short-Term Memory Network

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 11

参考文献 23

相关文章 12

编辑推荐

Metrics

本文评价

[1]	朱刘柱, 王绪利, 马静, 陈庆会, 齐先军. 基于小波包分解与循环神经网络的综合能源系统短期负荷预测[J]. 电力建设, 2020, 41(12): 133-140.
[2]	江雪辰，朱俊澎，袁越，王跃峰，黄阮明. 基于新型场景划分与考虑时序相关性的光伏出力时间序列模拟方法[J]. 电力建设, 2018, 39(10): 63-70.
[3]	赵宇，肖白，顾兵，王逍祎，张钰，王成龙. 基于改进马尔科夫链的风电功率时间序列模型[J]. 电力建设, 2017, 38(7): 18-.
[4]	张炜，王英洁，邬蓉蓉. 面向油中溶解气体监测时序数据压缩的改进方法[J]. 电力建设, 2017, 38(5): 98-.
[5]	裴湉, 齐冬莲. 基于时间序列提取和维诺图的电力数据异常检测方法 [J]. 电力建设, 2017, 38(5): 105-.
[6]	宋梦琪，陶顺，陈鹏伟. 基于主动配电网运行特性的电压质量评估[J]. 电力建设, 2016, 37(12): 119-.
[7]	陈伟彪，陈亦平，姚伟，孙雁斌，文劲宇. 基于多源信息时间序列匹配的电网故障诊断方法[J]. 电力建设, 2016, 37(12): 134-.
[8]	邱金鹏，牛东晓. 基于小波-时间序列组合模型的风电功率预测[J]. 电力建设, 2016, 37(1): 125-130.
[9]	陈彬，唐波，曹红英，彭友仙，李昱. 短波频段输电线路无源干扰谐振影响因子的关联性分析[J]. 电力建设, 2015, 36(6): 53-58.
[10]	付学谦，陈皓勇，牛铭，金小明. 地区电网短期负荷预测系统设计[J]. 电力建设, 2014, 35(2): 13-17.
[11]	陈尚发 . 我国电力系统终端用户的无功管理与电价[J]. 电力建设, 2008, 29(4): 89-91.
[12]	梁平,樊福梅,曾东平 . 电力负荷非线性预报与分析[J]. 电力建设, 2004, 25(3): 4-0.