变压器油色谱在线监测数据甄别与清洗技术
|
摘要
油色谱在线监测系统能够连续、实时监测主变内部绝缘状况,其监测数据是否准确、可靠直接关系到主变的安全稳定运行。经调查,福建省电网在用主变油色谱在线监测系统的运行现状,发现系统存在报警信号滞后、传感器装置故障频次较高造成数据不可靠等问题。为此,首先提出采用滑动窗口的形式实时采集在线监测数据,利用基于密度模式的空间数据聚类算法将滑动窗口内数据逐一甄别;其次,针对不同数据类型的在线监测系统"脏数据",采取不同清洗手段,并利用小波神经网络算法预测和填补清洗数据点;最后,通过实例验证所提方法能够准确甄别在线监测数据类型,有效地清洗在线监测"脏数据",提升在线监测数据的可靠性和数据处理及时性。
The on-line monitoring system of oil chromatography can continuously monitor the insulation status of transformer in real time. Whether the monitoring data is accurate and reliable is directly related to the safe and stable operation of transformer. Based on the investigation of the operation status of transformer oil chromatography on-line monitoring system used in Fujian power grid,it is found that the system has some problems,such as delayed alarm signal,high failure frequency of sensor device and unreliable data. In the light of this,this paper firstly proposes to collect the online monitoring data in real time in the form of sliding window,and uses the spatial data clustering algorithm based on density mode to identify the data in sliding window one by one. Secondly,for the"dirty data"of online monitoring system of different data types,different cleaning methods are adopted,and the wavelet neural network algorithm is used to predict and fill in the cleaning data points. Finally,an example verifies that the method proposed in this paper can accurately identify the online monitoring data types,effectively clean the online monitoring "dirty data",and improve the reliability of online monitoring data and data processing timeliness.
The on-line monitoring system of oil chromatography can continuously monitor the insulation status of transformer in real time. Whether the monitoring data is accurate and reliable is directly related to the safe and stable operation of transformer. Based on the investigation of the operation status of transformer oil chromatography on-line monitoring system used in Fujian power grid,it is found that the system has some problems,such as delayed alarm signal,high failure frequency of sensor device and unreliable data. In the light of this,this paper firstly proposes to collect the online monitoring data in real time in the form of sliding window,and uses the spatial data clustering algorithm based on density mode to identify the data in sliding window one by one. Secondly,for the"dirty data"of online monitoring system of different data types,different cleaning methods are adopted,and the wavelet neural network algorithm is used to predict and fill in the cleaning data points. Finally,an example verifies that the method proposed in this paper can accurately identify the online monitoring data types,effectively clean the online monitoring "dirty data",and improve the reliability of online monitoring data and data processing timeliness.
引文
[1]梁永亮,李可军,赵建国,等.变压器油色谱在线监测周期动态调整策略研究[J].中国电机工程学报,2014,34(9):1446-1453.
[2]赵婉芳,王慧芳,邱剑,等.基于油色谱监测数据的变压器动态可靠性分析[J].电力系统自动化,2014,38(22):38-42.
[3]郝思鹏,张济韬,张仰飞,等.融合在线监测数据的变压器状态评估[J].电力自动化设备,2017,37(11):176-181.
[4]郭雅辉.变压器油色谱在线检测与故障诊断研究[J].科技风,2018(13):161.
[5]杨小涛.在线油色谱装置监测变压器运行技术分析[J].通讯世界,2018,25(12):167-168.
[6]范竞敏,汪沨,蒋勤稷,等.固体氧化物燃料电池传感器检测变压器油中溶解气体定量特性[J].电工技术学报,2017,32(21):186-193.
[7]翟少磊,曹敏,沈鑫,等.变电站在线监测多维信息聚合技术[J].高电压技术,2015,41(12):3973-3979.
[8]潘玮,郑鹏,黄锦泉,等.基于数据清洗“DEAN”流程的健康信息领域研究热点探测[J].现代情报,2018,38(10):73-77.
[9]沈晓东.基于小波模极大值变压器油色谱在线异常数据识别[J].新型工业化,2018,8(6):26-31.
[10]李恩文,王力农,宋斌,等.基于改进模糊聚类算法的变压器油色谱分析[J].电工技术学报,2018,33(19):4594-4602.
[11]康乐,刘勇志.变压器油色谱在线监测数据有效性评估[J].建筑工程技术与设计,2018(30):3387.
[12]黄晓璐,李莉.变压器油色谱在线监测装置的运维管理[J].数字化用户,2018,24(23):41.
[13]唐红,郎雪淞,郑维刚,等.油色谱在线监测装置准确性校验方法的研究[J].东北电力技术,2018,39(4):11-13.
[14]刁赢龙,盛万兴,刘科研,等.大规模配电网负荷数据在线清洗与修复方法研究[J].电网技术,2015,39(11):3134-3140.
[15]康振华.大型电力变压器油色谱在线监测技术与案例分析[J].华电技术,2018,40(3):17-20.
[16]曲朝阳,张艺竞,王永文,等.基于Spark框架的能源互联网电力能源大数据清洗模型[J].电测与仪表,2018,55(2):39-44.
[2]赵婉芳,王慧芳,邱剑,等.基于油色谱监测数据的变压器动态可靠性分析[J].电力系统自动化,2014,38(22):38-42.
[3]郝思鹏,张济韬,张仰飞,等.融合在线监测数据的变压器状态评估[J].电力自动化设备,2017,37(11):176-181.
[4]郭雅辉.变压器油色谱在线检测与故障诊断研究[J].科技风,2018(13):161.
[5]杨小涛.在线油色谱装置监测变压器运行技术分析[J].通讯世界,2018,25(12):167-168.
[6]范竞敏,汪沨,蒋勤稷,等.固体氧化物燃料电池传感器检测变压器油中溶解气体定量特性[J].电工技术学报,2017,32(21):186-193.
[7]翟少磊,曹敏,沈鑫,等.变电站在线监测多维信息聚合技术[J].高电压技术,2015,41(12):3973-3979.
[8]潘玮,郑鹏,黄锦泉,等.基于数据清洗“DEAN”流程的健康信息领域研究热点探测[J].现代情报,2018,38(10):73-77.
[9]沈晓东.基于小波模极大值变压器油色谱在线异常数据识别[J].新型工业化,2018,8(6):26-31.
[10]李恩文,王力农,宋斌,等.基于改进模糊聚类算法的变压器油色谱分析[J].电工技术学报,2018,33(19):4594-4602.
[11]康乐,刘勇志.变压器油色谱在线监测数据有效性评估[J].建筑工程技术与设计,2018(30):3387.
[12]黄晓璐,李莉.变压器油色谱在线监测装置的运维管理[J].数字化用户,2018,24(23):41.
[13]唐红,郎雪淞,郑维刚,等.油色谱在线监测装置准确性校验方法的研究[J].东北电力技术,2018,39(4):11-13.
[14]刁赢龙,盛万兴,刘科研,等.大规模配电网负荷数据在线清洗与修复方法研究[J].电网技术,2015,39(11):3134-3140.
[15]康振华.大型电力变压器油色谱在线监测技术与案例分析[J].华电技术,2018,40(3):17-20.
[16]曲朝阳,张艺竞,王永文,等.基于Spark框架的能源互联网电力能源大数据清洗模型[J].电测与仪表,2018,55(2):39-44.