CN113077357A - Power time sequence data abnormity detection method and filling method thereof - Google Patents
Power time sequence data abnormity detection method and filling method thereof Download PDFInfo
- Publication number
- CN113077357A CN113077357A CN202110335947.8A CN202110335947A CN113077357A CN 113077357 A CN113077357 A CN 113077357A CN 202110335947 A CN202110335947 A CN 202110335947A CN 113077357 A CN113077357 A CN 113077357A
- Authority
- CN
- China
- Prior art keywords
- data
- abnormal
- missing
- value
- power time
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 45
- 238000000034 method Methods 0.000 title claims abstract description 33
- 230000002159 abnormal effect Effects 0.000 claims abstract description 71
- 230000003068 static effect Effects 0.000 claims abstract description 5
- 230000005856 abnormality Effects 0.000 claims description 24
- 238000010586 diagram Methods 0.000 description 4
- 238000004519 manufacturing process Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/21—Design, administration or maintenance of databases
- G06F16/215—Improving data quality; Data cleansing, e.g. de-duplication, removing invalid entries or correcting typographical errors
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Theoretical Computer Science (AREA)
- Economics (AREA)
- Databases & Information Systems (AREA)
- Health & Medical Sciences (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Marketing (AREA)
- Quality & Reliability (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Data Mining & Analysis (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Supply And Distribution Of Alternating Current (AREA)
- Testing And Monitoring For Control Systems (AREA)
Abstract
The invention discloses an electric power time sequence data abnormity detection method, which comprises the steps of obtaining electric power time sequence data to be analyzed and judging missing values; filling missing data and judging abnormity; and finally, performing power time series data abnormity detection according to the abnormity judgment result. The invention also discloses a filling method comprising the power time sequence data abnormity detection method. The method judges whether the data needs to be filled according to the missing condition of the power time series data, and fills the data by using a moving average method of daytime data; meanwhile, whether the filled data is filled abnormally is considered, and abnormal point detection is carried out by adopting an abnormal detection method of combining a transverse short-term ring ratio with a dynamic threshold and combining a longitudinal same-ratio amplitude with a static threshold; finally, correcting the data with abnormal detection by using a moving average method of data in the day; therefore, the method can realize the abnormal detection and filling of the power time sequence data, and has the advantages of stability, reliability, high accuracy and higher efficiency.
Description
Technical Field
The invention belongs to the field of electrical automation, and particularly relates to a power time sequence data abnormity detection method and a filling method thereof.
Background
With the development of economic technology and the improvement of living standard of people, electric energy becomes essential secondary energy in production and life of people, and brings endless convenience to production and life of people. Therefore, stable and reliable operation of the power system becomes one of the most important tasks of the power system.
In the electric power system, a terminal acquisition device acquires electric power time sequence data such as active power, electricity consumption and the like according to set time frequency, and stores and uploads the acquired data in a unified manner. However, due to uncontrollable factors such as equipment failure and transmission channel interference, an abnormal situation of power data loss is inevitable. Due to the irreproducibility of power grid data acquisition, missing power data needs to be filled and abnormal detection is carried out.
However, in the existing power system, there is no complete, reliable and efficient detection method for power time series data abnormality and corresponding filling method.
Disclosure of Invention
The invention aims to provide a method for detecting the abnormality of power time sequence data, which is stable, reliable, high in accuracy and high in efficiency.
Another object of the present invention is to provide a filling method including the power sequence data abnormality detection method.
The invention provides a power time sequence data abnormity detection method, which comprises the following steps:
s1, acquiring power time sequence data to be analyzed;
s2, carrying out missing value judgment on the data acquired in the step S1;
s3, aiming at the missing value determined in the step S2, filling missing data by adopting a daytime data moving average algorithm;
s4, aiming at the filled data obtained in the step S3, performing data abnormity judgment by adopting a short-term cycle ratio abnormity detection algorithm;
s5, aiming at the filled data obtained in the step S3, carrying out abnormal judgment on the data by adopting a same-ratio amplitude abnormal detection algorithm;
and S6, performing final power time series data abnormity detection according to the abnormity judgment results obtained in the steps S4 and S5.
In step S2, the missing value determination is performed on the data acquired in step S1, specifically, the following steps are adopted to perform the missing value determination:
if the power time sequence data is complete, missing value abnormal filling is not needed;
if the missing proportion of the power time sequence data is higher than a set threshold value, directly discarding the group of power time sequence data;
if the power sequence data missing ratio is within the set threshold value range, the following judgment is carried out again:
if the missing data is NA, then judging the missing value according to the area of the transformer substation and the actual situation of the collected data;
if the missing data is 0, judging according to the granularity division of the data acquisition source again:
if the data is acquired from the platform area or the fine-grained equipment, the data is determined to be normal, and the missing abnormal condition processing is not required;
if the data are acquired from various transformer substations with different voltage levels or equipment with a coarser granularity, the data are determined to be abnormal, and the missing abnormal condition is required to be processed.
In step S3, for the missing value determined in step S2, a daytime data moving average algorithm is used to perform missing data padding, specifically, the following steps are used to perform the processing:
and filling missing data by adopting data at the same time in the historical data:
middle valfIs a fill value; valhThe data values of the historical data corresponding to the missing data at the same moment in the previous h days are obtained; t ishIs the time window size of the historical data.
In step S4, for the filled data obtained in step S3, a short-term cycle ratio anomaly detection algorithm is used to determine anomalies in the data, specifically, the following steps are used to determine anomalies:
setting a time window T, and comparing the filled data with each data in the time window T: if the number of times that the difference between the filled data and the data in the time window T is greater than the set difference threshold exceeds the set number threshold, the filled data is determined as an abnormal point:
in the formula, H is a first determination value of an abnormal point, H is determined to be normal when H is 0, and H is determined to be abnormal when H is 1; count () is an operation of determining the number of times the expression in parentheses is established; valcThe data after being filled; valiThe data value of the ith data in the time window T; n is the total number of data in the time window T; t is tdIs a set difference threshold; nums is a set number threshold.
In step S5, the filled data obtained in step S3 is subjected to abnormality determination by using a unity-ratio amplitude abnormality detection algorithm, specifically, the following steps are performed:
and performing amplitude anomaly detection on the data at the time c in the filled data and the data at the same time c of the past several days by adopting the following formula:
wherein V is a second determination value of the abnormal point, and V is determined to be normal when V is 0 and is determined to be abnormal when V is 1; valc(t) data for time c on day t; valc(t-1) is data of the time c of day t-1; n is the last n days; α is an amplitude, andtsis a static threshold, and tsThe average value of the day data of the value to be detected is MIN (MAX-AVG, AVG-MIN), MAX is the maximum value of the day data of the value to be detected, AVG is the average value of the day data of the value to be detected, and MIN is the minimum value of the day data of the value to be detected.
In step S6, the final power time series data abnormality detection is performed according to the abnormality determination results obtained in steps S4 and S5, specifically, the following rules are adopted to perform abnormality determination:
if the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 are both 0, the determined power sequence data is determined to be normal data;
if any one of the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 is 1, the determined power sequence data is regarded as abnormal data.
The invention also discloses a filling method comprising the power time sequence data abnormity detection method, which further comprises the following steps:
s7, if the abnormal data is judged, the value val obtained by the following formulamAnd replacing the abnormal data so as to complete the filling of the abnormal data:
middle valmFor padded data; valtHistorical data of the time corresponding to the abnormal data of the previous t days; t istIs the size of the set time window.
The invention provides an abnormal detection method and a filling method for power time series data, which are used for judging whether the data needs to be filled according to the missing condition of the power time series data and filling the data by using a moving average method of daytime data; meanwhile, whether the filled data is filled abnormally is considered, and abnormal point detection is carried out by adopting an abnormal detection method of combining a transverse short-term ring ratio with a dynamic threshold and combining a longitudinal same-ratio amplitude with a static threshold; finally, correcting the data with abnormal detection by using a moving average method of data in the day; therefore, the method can realize the abnormal detection and filling of the power time sequence data, and has the advantages of stability, reliability, high accuracy and higher efficiency.
Drawings
FIG. 1 is a schematic method flow diagram of the anomaly detection method of the present invention.
FIG. 2 is a schematic diagram of the abnormality determination of the method of the present invention.
Fig. 3 is a schematic method flow diagram of the filling method of the present invention.
Detailed Description
Fig. 1 is a schematic flow chart of the method of the anomaly detection method of the present invention: the invention provides a power time sequence data abnormity detection method, which comprises the following steps:
s1, acquiring power time sequence data to be analyzed;
after data is acquired, the number of related variables needs to be determined, and missing data processing operation can be directly carried out on single-variable time sequence data; however, for multi-variable, it is necessary to perform variable unity processing to prevent multiple variables from affecting each other in the missing data stuffing processing operation;
s2, carrying out missing value judgment on the data acquired in the step S1; specifically, the method comprises the following steps of:
if the power time sequence data is complete, missing value abnormal filling is not needed;
if the missing proportion of the power time sequence data is higher than a set threshold value, directly discarding the group of power time sequence data;
if the power sequence data missing ratio is within the set threshold value range, the following judgment is carried out again:
if the missing data is NA, then judging the missing value according to the area of the transformer substation and the actual situation of the collected data;
if the missing data is 0, judging according to the granularity division of the data acquisition source again:
if the data is acquired from the platform area or the fine-grained equipment, the data is determined to be normal, and the missing abnormal condition processing is not required;
if the data are acquired from various transformer substations with different voltage levels or equipment with a coarser granularity, the data are determined to be abnormal, and the missing abnormal condition is required to be processed;
s3, aiming at the missing value determined in the step S2, filling missing data by adopting a daytime data moving average algorithm; the method specifically comprises the following steps:
and filling missing data by adopting data at the same time in the historical data:
middle valfIs a fill value; valhThe data values of the historical data corresponding to the missing data at the same moment in the previous h days are obtained; t ishA time window size for historical data;
s4, aiming at the filled data obtained in the step S3, performing data abnormity judgment by adopting a short-term cycle ratio abnormity detection algorithm; specifically, the abnormality determination is carried out by adopting the following steps:
setting a time window T, and comparing the filled data with each data in the time window T: if the number of times that the difference between the filled data and the data in the time window T is greater than the set difference threshold exceeds the set number threshold, the filled data is determined as an abnormal point:
in the formula, H is a first determination value of an abnormal point, H is determined to be normal when H is 0, and H is determined to be abnormal when H is 1; count () is an operation of determining the number of times the expression in parentheses is established; valcThe data after being filled; valiThe data value of the ith data in the time window T; n is the total number of data in the time window T; t is tdIs a set difference threshold; nums is a set frequency threshold;
s5, aiming at the filled data obtained in the step S3, carrying out abnormal judgment on the data by adopting a same-ratio amplitude abnormal detection algorithm; specifically, the abnormality determination is carried out by adopting the following steps:
and performing amplitude anomaly detection on the data at the time c in the filled data and the data at the same time c of the past several days by adopting the following formula:
wherein V is a second determination value of the abnormal point, and V is determined to be normal when V is 0 and is determined to be abnormal when V is 1; valc(t) data for time c on day t; valc(t-1) is data of the time c of day t-1; n is the last n days; α is an amplitude, andtsis a static threshold, and tsMIN (MAX-AVG, AVG-MIN), wherein MAX is the maximum value of the day all-day data of the value to be detected, AVG is the average value of the day all-day data of the value to be detected, and MIN is the minimum value of the day all-day data of the value to be detected;
the detection concepts of S4 and S5, as shown in FIG. 2;
s6, performing final power time sequence data abnormity detection according to the abnormity judgment results obtained in the steps S4 and S5; specifically, the following rules are adopted for abnormality determination:
if the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 are both 0, the determined power sequence data is determined to be normal data;
if any one of the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 is 1, the determined power sequence data is regarded as abnormal data;
expressed by the formula, namely: is or operation; YC ═ 0 indicates that the power time-series data is normal data, and YC ═ 1 indicates that the power time-series data is normal data.
Fig. 3 is a schematic diagram of a method flow of the filling method of the present invention: the filling method provided by the invention comprises the power time sequence data abnormity detection method, and further comprises the following steps:
s7, if the judgment is madeFor abnormal data, the value val calculated by the following formulamAnd replacing the abnormal data so as to complete the filling of the abnormal data:
middle valmFor padded data; valtHistorical data of the time corresponding to the abnormal data of the previous t days; t istIs the size of the set time window.
Claims (7)
1. A power time series data abnormity detection method comprises the following steps:
s1, acquiring power time sequence data to be analyzed;
s2, carrying out missing value judgment on the data acquired in the step S1;
s3, aiming at the missing value determined in the step S2, filling missing data by adopting a daytime data moving average algorithm;
s4, aiming at the filled data obtained in the step S3, performing data abnormity judgment by adopting a short-term cycle ratio abnormity detection algorithm;
s5, aiming at the filled data obtained in the step S3, carrying out abnormal judgment on the data by adopting a same-ratio amplitude abnormal detection algorithm;
and S6, performing final power time series data abnormity detection according to the abnormity judgment results obtained in the steps S4 and S5.
2. The method for detecting abnormality in power time series data according to claim 1, wherein the step S2 is performed to determine missing values of the data acquired in the step S1, specifically, the following steps are performed to determine missing values:
if the power time sequence data is complete, missing value abnormal filling is not needed;
if the missing proportion of the power time sequence data is higher than a set threshold value, directly discarding the group of power time sequence data;
if the power sequence data missing ratio is within the set threshold value range, the following judgment is carried out again:
if the missing data is NA, then judging the missing value according to the area of the transformer substation and the actual situation of the collected data;
if the missing data is 0, judging according to the granularity division of the data acquisition source again:
if the data is acquired from the platform area or the fine-grained equipment, the data is determined to be normal, and the missing abnormal condition processing is not required;
if the data are acquired from various transformer substations with different voltage levels or equipment with a coarser granularity, the data are determined to be abnormal, and the missing abnormal condition is required to be processed.
3. The method for detecting abnormality in power time series data according to claim 1 or 2, wherein the missing data filling is performed by using a daytime data moving average algorithm for the missing value determined in step S2 in step S3, and specifically, the following steps are performed:
and filling missing data by adopting data at the same time in the historical data:
middle valfIs a fill value; valhThe data values of the historical data corresponding to the missing data at the same moment in the previous h days are obtained; t ishIs the time window size of the historical data.
4. The method for detecting abnormality in power time series data according to claim 3, wherein the step S4 is to determine abnormality of the data by using a short-term cycle ratio abnormality detection algorithm with respect to the filled data obtained in the step S3, specifically, the method comprises the following steps:
setting a time window T, and comparing the filled data with each data in the time window T: if the number of times that the difference between the filled data and the data in the time window T is greater than the set difference threshold exceeds the set number threshold, the filled data is determined as an abnormal point:
in the formula, H is a first determination value of an abnormal point, H is determined to be normal when H is 0, and H is determined to be abnormal when H is 1; count () is an operation of determining the number of times the expression in parentheses is established; valcThe data after being filled; valiThe data value of the ith data in the time window T; n is the total number of data in the time window T; t is tdIs a set difference threshold; nums is a set number threshold.
5. The method for detecting abnormality in power time series data according to claim 4, wherein the filled data obtained in step S3 is subjected to abnormality determination by a geometric amplitude abnormality detection algorithm in step S5, specifically, the following steps are adopted:
and performing amplitude anomaly detection on the data at the time c in the filled data and the data at the same time c of the past several days by adopting the following formula:
wherein V is a second determination value of the abnormal point, and V is determined to be normal when V is 0 and is determined to be abnormal when V is 1; valc(t) data for time c on day t; valc(t-1) is data of the time c of day t-1; n is the last n days; α is an amplitude, andtsis a static threshold, and tsMAX (MAX-AVG, AVG-MIN), MAX being the maximum value of the day data of the value to be detected, AVG being the average value of the day data of the value to be detectedAnd MIN is the minimum value of the data of the whole day of the day where the value to be detected is located.
6. The method for detecting abnormality in power time series data according to claim 5, wherein the final detection of abnormality in power time series data is performed according to the abnormality determination results obtained in steps S4 and S5 in step S6, specifically, the abnormality determination is performed according to the following rules:
if the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 are both 0, the determined power sequence data is determined to be normal data;
if any one of the first determination value H of the abnormal point obtained in step S4 and the second determination value V of the abnormal point obtained in step S5 is 1, the determined power sequence data is regarded as abnormal data.
7. A filling method comprising the power time series data abnormity detection method of any claim 1-6, characterized by further comprising the following steps:
s7, if the abnormal data is judged, the value val obtained by the following formulamAnd replacing the abnormal data so as to complete the filling of the abnormal data:
middle valmFor padded data; valtHistorical data of the time corresponding to the abnormal data of the previous t days; t istIs the size of the set time window.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110335947.8A CN113077357B (en) | 2021-03-29 | 2021-03-29 | Power time sequence data anomaly detection method and filling method thereof |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110335947.8A CN113077357B (en) | 2021-03-29 | 2021-03-29 | Power time sequence data anomaly detection method and filling method thereof |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113077357A true CN113077357A (en) | 2021-07-06 |
CN113077357B CN113077357B (en) | 2023-11-28 |
Family
ID=76611538
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110335947.8A Active CN113077357B (en) | 2021-03-29 | 2021-03-29 | Power time sequence data anomaly detection method and filling method thereof |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113077357B (en) |
Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012089057A (en) * | 2010-10-22 | 2012-05-10 | Hitachi Engineering & Services Co Ltd | Facility state monitoring method, apparatus therefor and facility state monitoring program |
WO2016012972A1 (en) * | 2014-07-25 | 2016-01-28 | Suez Environnement | Method for detecting anomalies in a distribution network, in particular for drinking water |
CN107463633A (en) * | 2017-07-17 | 2017-12-12 | 中国航天系统科学与工程研究院 | A kind of real time data rejecting outliers method based on EEMD neutral nets |
CN107563539A (en) * | 2017-07-24 | 2018-01-09 | 佛山市顺德区中山大学研究院 | Short-term and long-medium term power load forecasting method based on machine learning model |
CN109299170A (en) * | 2018-10-25 | 2019-02-01 | 南京大学 | A kind of complementing method for tape label time series data |
CN109947812A (en) * | 2018-07-09 | 2019-06-28 | 平安科技(深圳)有限公司 | Consecutive miss value fill method, data analysis set-up, terminal and storage medium |
US20190391574A1 (en) * | 2018-06-25 | 2019-12-26 | Nec Laboratories America, Inc. | Early anomaly prediction on multi-variate time series data |
CN110865929A (en) * | 2019-11-26 | 2020-03-06 | 携程旅游信息技术(上海)有限公司 | Abnormity detection early warning method and system |
CN111340288A (en) * | 2020-02-25 | 2020-06-26 | 武汉墨锦创意科技有限公司 | Urban air quality time sequence prediction method considering space-time correlation |
CN111507412A (en) * | 2020-04-20 | 2020-08-07 | 南京工程学院 | Voltage missing value filling method based on historical data auxiliary scene analysis |
CN111695620A (en) * | 2020-06-08 | 2020-09-22 | 中国电力科学研究院有限公司 | Method and system for detecting and correcting abnormal data of time sequence of power system |
CN111967509A (en) * | 2020-07-31 | 2020-11-20 | 北京赛博星通科技有限公司 | Method and device for processing and detecting data acquired by industrial equipment |
CN111984514A (en) * | 2020-09-02 | 2020-11-24 | 大连大学 | Prophet-bLSTM-DTW-based log anomaly detection method |
CN112101482A (en) * | 2020-10-26 | 2020-12-18 | 西安交通大学 | Method for detecting abnormal parameter mode of missing satellite data |
CN112365070A (en) * | 2020-11-18 | 2021-02-12 | 深圳供电局有限公司 | Power load prediction method, device, equipment and readable storage medium |
CN112527788A (en) * | 2020-12-17 | 2021-03-19 | 北京中恒博瑞数字电力科技有限公司 | Method and device for detecting and cleaning abnormal value of transformer monitoring data |
-
2021
- 2021-03-29 CN CN202110335947.8A patent/CN113077357B/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2012089057A (en) * | 2010-10-22 | 2012-05-10 | Hitachi Engineering & Services Co Ltd | Facility state monitoring method, apparatus therefor and facility state monitoring program |
WO2016012972A1 (en) * | 2014-07-25 | 2016-01-28 | Suez Environnement | Method for detecting anomalies in a distribution network, in particular for drinking water |
CN107463633A (en) * | 2017-07-17 | 2017-12-12 | 中国航天系统科学与工程研究院 | A kind of real time data rejecting outliers method based on EEMD neutral nets |
CN107563539A (en) * | 2017-07-24 | 2018-01-09 | 佛山市顺德区中山大学研究院 | Short-term and long-medium term power load forecasting method based on machine learning model |
US20190391574A1 (en) * | 2018-06-25 | 2019-12-26 | Nec Laboratories America, Inc. | Early anomaly prediction on multi-variate time series data |
CN109947812A (en) * | 2018-07-09 | 2019-06-28 | 平安科技(深圳)有限公司 | Consecutive miss value fill method, data analysis set-up, terminal and storage medium |
CN109299170A (en) * | 2018-10-25 | 2019-02-01 | 南京大学 | A kind of complementing method for tape label time series data |
CN110865929A (en) * | 2019-11-26 | 2020-03-06 | 携程旅游信息技术(上海)有限公司 | Abnormity detection early warning method and system |
CN111340288A (en) * | 2020-02-25 | 2020-06-26 | 武汉墨锦创意科技有限公司 | Urban air quality time sequence prediction method considering space-time correlation |
CN111507412A (en) * | 2020-04-20 | 2020-08-07 | 南京工程学院 | Voltage missing value filling method based on historical data auxiliary scene analysis |
CN111695620A (en) * | 2020-06-08 | 2020-09-22 | 中国电力科学研究院有限公司 | Method and system for detecting and correcting abnormal data of time sequence of power system |
CN111967509A (en) * | 2020-07-31 | 2020-11-20 | 北京赛博星通科技有限公司 | Method and device for processing and detecting data acquired by industrial equipment |
CN111984514A (en) * | 2020-09-02 | 2020-11-24 | 大连大学 | Prophet-bLSTM-DTW-based log anomaly detection method |
CN112101482A (en) * | 2020-10-26 | 2020-12-18 | 西安交通大学 | Method for detecting abnormal parameter mode of missing satellite data |
CN112365070A (en) * | 2020-11-18 | 2021-02-12 | 深圳供电局有限公司 | Power load prediction method, device, equipment and readable storage medium |
CN112527788A (en) * | 2020-12-17 | 2021-03-19 | 北京中恒博瑞数字电力科技有限公司 | Method and device for detecting and cleaning abnormal value of transformer monitoring data |
Non-Patent Citations (9)
Title |
---|
IT瘾: "时间序列异常检测算法梳理", Retrieved from the Internet <URL:https://itindex.net/detail/60453-%E6%97%B6%E9%97%B4%E5%BA%8F%E5%88%97-%E5%BC%82%E5%B8%B8%E6%A3%80%E6%B5%8B-%E7%AE%97%E6%B3%95> * |
JASMINEXJF: "基于时间序列的异常检测算法小结", Retrieved from the Internet <URL:https://blog.csdn.net/Jasminexjf/article/details/88527966> * |
XIAOHUI WANG等: "Power Consumption Predicting and Anomaly Detection Based on Long Short-Term Memory Neural Network", 《2019 IEEE 4TH INTERNATIONAL CONFERENCE ON CLOUD COMPUTING AND BIG DATA ANALYSIS》, pages 487 - 491 * |
YUCHENG CHEN等: "Power Equipment Anomaly Detection Based on Spatiotemporal Clustering", 《2016 INTERNATIONAL CONFERENCE ON CONDITION MONITORING AND DIAGNOSIS(CMD)》, pages 392 - 395 * |
余斌: "电力感知数据的异常模式检测", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, pages 042 - 1682 * |
倪景峰: "基于最小二乘支持向量机算法的测量数据时序异常检测方法", 《华北电力大学学报(自然科学版)》, vol. 132, no. 3, pages 62 - 66 * |
小小统计员: "模型进阶(1):数据缺失的处理方法", Retrieved from the Internet <URL:https://zhuanlan.zhihu.com/p/33263613> * |
康旭: "时序遥测数据异常检测方法研究", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》, no. 2, pages 031 - 726 * |
裴丽鹊: "一种基于滑动窗口的时间序列异常检测算法", 《巢湖学院学报》, vol. 13, no. 3, pages 28 - 31 * |
Also Published As
Publication number | Publication date |
---|---|
CN113077357B (en) | 2023-11-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN116702081B (en) | Intelligent inspection method for power distribution equipment based on artificial intelligence | |
CN111047082A (en) | Early warning method and device for equipment, storage medium and electronic device | |
CN111024158B (en) | Kitchen electrical hazard intelligent monitoring method combined with edge calculation | |
CN108088916B (en) | A kind of method for improving and system of oil dissolved gas online monitoring data quality | |
CN108627720B (en) | Power equipment state monitoring method based on Bayesian algorithm | |
CN117129790B (en) | Fault diagnosis system for power system | |
CN103729804A (en) | On-line decision support method for responding to power quality early warning | |
CN117332215B (en) | High-low voltage power distribution cabinet abnormal fault information remote monitoring system | |
CN116069079A (en) | Intelligent heat dissipation control method and system for intelligent switch cabinet | |
CN116739829B (en) | Big data-based power data analysis method, system and medium | |
CN117411189A (en) | Monitoring data enhancement method of micro-grid coordination controller | |
El Chamie et al. | Physics-based features for anomaly detection in power grids with micro-pmus | |
CN114369849A (en) | Method and device for monitoring health degree of electrolytic cell and electrolytic cell monitoring system | |
CN116203351A (en) | Method and system for detecting abnormal line impedance | |
CN113077357A (en) | Power time sequence data abnormity detection method and filling method thereof | |
CN112731022B (en) | Photovoltaic inverter fault detection method, equipment and medium | |
CN112731827B (en) | Monitoring system for intelligent sensor for power equipment | |
CN117131321A (en) | Electric energy data acquisition method | |
CN111537819A (en) | Distribution automation monitoring method, system, terminal and storage medium | |
CN117031201A (en) | Multi-scene topology anomaly identification method and system for power distribution network | |
CN115248292A (en) | Transformer fault analysis and diagnosis method and system | |
CN115511156A (en) | Main transformer overload early warning method and system based on load prediction | |
CN112699104B (en) | Data processing method of intelligent sensor of power equipment and intelligent sensor | |
CN113418632A (en) | Conceptual drift detection method for oil temperature prediction of oil-immersed transformer | |
CN113627885A (en) | Power grid power supply abnormity monitoring system and monitoring method thereof |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |