CN109342630B - Transformer oil chromatographic online monitoring abnormal data diagnosis method - Google Patents
Transformer oil chromatographic online monitoring abnormal data diagnosis method Download PDFInfo
- Publication number
- CN109342630B CN109342630B CN201811409813.0A CN201811409813A CN109342630B CN 109342630 B CN109342630 B CN 109342630B CN 201811409813 A CN201811409813 A CN 201811409813A CN 109342630 B CN109342630 B CN 109342630B
- Authority
- CN
- China
- Prior art keywords
- data
- frequency
- distribution sequence
- frequency distribution
- transformer oil
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N30/00—Investigating or analysing materials by separation into components using adsorption, absorption or similar phenomena or using ion-exchange, e.g. chromatography or field flow fractionation
- G01N30/02—Column chromatography
- G01N30/86—Signal analysis
- G01N30/8675—Evaluation, i.e. decoding of the signal into analytical information
Abstract
The invention discloses a method for diagnosing abnormal data of transformer oil chromatographic online monitoring, which comprises the following steps: acquiring historical data of transformer oil chromatographic monitoring, and performing attribute division on the historical data according to gas types according to a preset time period; preprocessing historical data corresponding to each type of gas to obtain a preprocessed sample; then obtaining a unary frequency distribution sequence and a binary frequency distribution sequence of the preprocessed sample; then establishing an abnormal data diagnosis model; updating the unary frequency distribution sequence and the binary frequency distribution sequence by using new monitoring data; and finally, judging the unary frequency and the binary frequency of the monitored data to be detected according to the updated unary frequency distribution sequence and the binary frequency distribution sequence, and substituting the unary frequency and the binary frequency into an abnormal data diagnosis model for diagnosis. The invention is beneficial to improving the diagnosis accuracy and is convenient to identify a novel abnormal mode.
Description
Technical Field
The invention relates to the technical field of transformer maintenance, in particular to a transformer oil chromatogram online monitoring abnormal data diagnosis method.
Background
The online monitoring of transformer oil chromatography is an important means for detecting the internal defects of the transformer, and has been widely applied to the transformers with the voltage class of 220 kilovolt and above. In field application, oil chromatography on-line monitoring generally determines whether to issue an alarm by judging whether real-time measurement data exceeds a preset value, but practice shows that an on-line monitoring device is easily influenced by factors such as complex strong electromagnetic environment, severe operation working conditions, equipment quality and the like of high-voltage equipment to generate a data abnormal phenomenon, so that a trigger device generates a false alarm, which aggravates field operation maintenance workload on one hand, and submerges correct alarm on the other hand. Therefore, the research on the diagnosis of the abnormal data of the transformer oil chromatogram has important significance for reducing the field workload and improving the monitoring accuracy of the equipment state. However, the existing transformer oil chromatogram abnormal data diagnosis method mostly utilizes a mode identification method, identifies abnormal data by using regular characterization features, has strong pertinence, cannot identify a novel abnormal mode, and has low diagnosis accuracy.
Disclosure of Invention
The invention aims to provide a method for diagnosing abnormal data of transformer oil chromatography on-line monitoring, which is beneficial to improving the diagnosis accuracy and identifying a novel abnormal mode.
In order to solve the technical problems, the technical scheme provided by the invention is as follows:
a transformer oil chromatogram on-line monitoring abnormal data diagnosis method comprises the following steps:
1) acquiring historical data of transformer oil chromatographic monitoring, and performing attribute division on the historical data according to gas types according to a preset time period;
2) preprocessing historical data corresponding to each type of gas in a preset time period to obtain a preprocessed sample, and defining data in the preprocessed sample as preprocessed data;
3) performing statistical analysis on the preprocessed samples corresponding to each type of gas to obtain a unitary frequency distribution sequence and a binary frequency distribution sequence of the preprocessed samples corresponding to each type of gas;
4) establishing an abnormal data diagnosis model according to historical data, a unary frequency distribution sequence and a binary frequency distribution sequence of transformer oil chromatographic monitoring;
5) acquiring a group of new monitoring data, calculating the unitary frequency and the binary frequency of the group of new monitoring data, and updating the unitary frequency distribution sequence and the binary frequency distribution sequence in the step 3) according to the unitary frequency and the binary frequency of the group of new monitoring data;
6) and judging the unary frequency and the binary frequency of the monitored data to be detected according to the unary frequency distribution sequence and the binary frequency distribution sequence updated in the step 5), and substituting the unary frequency and the binary frequency into an abnormal data diagnosis model for diagnosis.
The method for preprocessing the historical data corresponding to each type of gas in the step 2) comprises the following steps: and eliminating negative values in the historical data, merging continuous and unchangeable numerical values in a preset time period into a numerical value, and taking the sampling time of the last continuous and unchangeable numerical value as the sampling time of the merged numerical value.
The method for preprocessing the historical data corresponding to each gas type further comprises the following steps: and keeping the data value of each historical data to three decimal places.
The step 3) is specifically as follows: dividing a plurality of data intervals, wherein the interval length of each data interval is the same, judging the data interval in which the preprocessed data in each preprocessed sample is located, sequencing the data intervals according to the sampling time of the preprocessed data, then calculating unary frequency and binary frequency, wherein the values of all unary frequency of each preprocessed sample form an unary frequency distribution sequence, and the values of all binary frequency of each preprocessed sample form a binary frequency distribution sequence;
the unitary frequency is the frequency of each data interval appearing in a preset time period, and the binary frequency is the frequency of two adjacent data intervals appearing repeatedly in the preset time period.
The following steps are also carried out after the step 6): and analyzing the diagnosis result of the abnormal data diagnosis model, and optimizing the abnormal data diagnosis model according to the analysis result.
The abnormal data diagnosis model in the step 4) is established by adopting a logistic regression model.
The gas types include hydrogen, methane, ethane, ethylene, acetylene, carbon monoxide, carbon dioxide, oxygen, nitrogen, and total hydrocarbons.
The invention has the following beneficial effects: according to the method for diagnosing the abnormal data of the transformer oil chromatogram in the online monitoring manner, the abnormal data diagnosis model is established for diagnosing the abnormal data, the abnormal data diagnosis model is convenient to optimize, the diagnosis accuracy is improved, and a novel abnormal mode can be conveniently identified; the expansion capability of the abnormal data diagnosis model is improved.
Drawings
FIG. 1 is a schematic flow chart of the online monitoring abnormal data diagnosis method for transformer oil chromatography.
Detailed Description
The present invention is further described below in conjunction with the following figures and specific examples so that those skilled in the art may better understand the present invention and practice it, but the examples are not intended to limit the present invention.
As shown in fig. 1, the embodiment discloses a method for diagnosing abnormal data of online monitoring of transformer oil chromatography, which comprises the following steps:
1) acquiring historical data of oil chromatography monitoring of a transformer, and performing attribute division on the historical data according to gas types according to a preset time period; the historical data of each type of gas in a preset time period is classified into one group, if the number of the gas types is ten, the gas types are classified into ten groups, and each gas type corresponds to one group;
further, the historical data is gas content, and the gas types include hydrogen, methane, ethane, ethylene, acetylene, carbon monoxide, carbon dioxide, oxygen, nitrogen, and total hydrocarbons.
The preset time period may be three years.
2) Preprocessing historical data corresponding to each type of gas in a preset time period to obtain a preprocessed sample, and defining data in the preprocessed sample as preprocessed data;
3) performing statistical analysis on the pretreatment sample corresponding to each type of gas to obtain a unitary frequency distribution sequence and a binary frequency distribution sequence of the pretreatment sample corresponding to each type of gas; wherein each type of gas corresponds to a pretreatment sample;
the step 3) is specifically as follows: dividing a plurality of data intervals, wherein the interval length of each data interval is the same, judging the data interval in which the preprocessed data in each preprocessed sample is located, sequencing the data intervals according to the sampling time of the preprocessed data, then calculating unary frequency and binary frequency, wherein the values of all unary frequency of each preprocessed sample form an unary frequency distribution sequence, and the values of all binary frequency of each preprocessed sample form a binary frequency distribution sequence;
the unitary frequency is a frequency of each data interval appearing in a preset time period, and the binary frequency is a frequency of two adjacent data intervals appearing repeatedly in the preset time period, that is, a frequency of each data interval appearing after another data interval.
For example, a preprocessed sample is {0.02, 0.15, 0.17, 0.06, 0.12, 0.13, 0.19, 0.20, 0.16, 0.26}, each preprocessed data in the sample is the gas content of hydrogen and is ordered according to the sequence of data sampling time, three data intervals are defined, the length of the data interval is 0.1, each data interval is {0-0.1}, {0.1-0.2}, and {0.2-0.3}, and the preprocessed data in the preprocessed sample are in the interval sequence of {0-0.1, 0.1-0.2, 0.1-0.2, 0-0.1, 0.1-0.2, 0.1-0.2, 0.1-0.2, 0.2-0.3, 0.1-0.2, 0.3, and the frequency of the frequency interval of { 0.1-0.1 } is two times of the unary data in the interval of { 0.1-0.1, 0.1-0.3 }, the primary frequency is the ratio of the number of occurrences (twice) to the total number (ten times); the frequency of the data interval {0-0.1} and {0.1-0.2} adjacent to each other is two times, that is, the frequency of the data interval {0.1-0.2} adjacent to each other after {0-0.1} is two times, and the binary frequency is the ratio of the number of times (two times) of the two intervals adjacent to each other to the total number of times (ten times).
4) Establishing an abnormal data diagnosis model according to historical data, a unitary frequency distribution sequence and a binary frequency distribution sequence of transformer oil chromatographic monitoring; it will be appreciated that each gas type corresponds to an anomaly data diagnostic model.
5) Acquiring a group of new monitoring data, calculating the unary frequency and the binary frequency of the group of new monitoring data, and updating the unary frequency distribution sequence and the binary frequency distribution sequence in the step 3) according to the unary frequency and the binary frequency of the group of new monitoring data;
6) and 5) judging the unitary frequency and the binary frequency of the monitoring data to be detected according to the unitary frequency distribution sequence and the binary frequency distribution sequence updated in the step 5), substituting the unitary frequency and the binary frequency into an abnormal data diagnosis model for diagnosis, and diagnosing whether the monitoring data is abnormal data.
Step 6) is followed by the following steps: and the diagnosis result of the abnormal data diagnosis model is manually analyzed, and the abnormal data diagnosis model is optimized according to the analysis result, so that the novel abnormal mode can be identified, the expansion capability of the abnormal data diagnosis model is further improved, and the diagnosis accuracy is also improved.
In one embodiment, the method for preprocessing the historical data corresponding to each type of gas in step 2) includes: and eliminating negative values in the historical data, merging continuous and unchangeable values in a preset time period into a value, and taking the sampling time of the last continuous and unchangeable value as the sampling time of the merged value during merging, namely taking the sampling time corresponding to the value with the most leaner sampling time in the continuous and unchangeable values as the sampling time of the merged value.
In one embodiment, the method for preprocessing the historical data corresponding to each gas type further comprises: and (4) reserving the data value of each historical data to three digits after the decimal point, wherein the data value is less than three digits after the decimal point, and complementing the three digits after the decimal point by adopting a zero padding mode.
In one embodiment, the abnormal data diagnosis model in step 4) is established by using a logistic regression model. The step 4) is specifically as follows: and diagnosing and marking the history data of the transformer oil chromatographic monitoring, and establishing an abnormal data diagnosis model by a logistic stewart regression method according to the marking result, the unary frequency distribution sequence and the binary frequency distribution sequence.
According to the transformer oil chromatogram online monitoring abnormal data diagnosis method, through preprocessing of historical data and establishment of the unitary frequency distribution sequence and the binary frequency sequence, an abnormal data diagnosis model is optimized, diagnosis accuracy is improved, meanwhile, the unitary frequency distribution sequence and the binary frequency sequence are convenient to update, a novel abnormal mode is convenient to recognize, and the diagnosis accuracy can be further improved; the method is also easy to expand to other types of abnormal data diagnosis and analysis of monitoring quantity, and the expansion capability of the transformer oil chromatogram online monitoring abnormal data diagnosis model can be obviously improved.
The above-mentioned embodiments are merely preferred embodiments for fully illustrating the present invention, and the scope of the present invention is not limited thereto. The equivalent substitution or change made by the technical personnel in the technical field on the basis of the invention is all within the protection scope of the invention. The protection scope of the invention is subject to the claims.
Claims (7)
1. A transformer oil chromatogram on-line monitoring abnormal data diagnosis method is characterized by comprising the following steps:
1) acquiring historical data of transformer oil chromatographic monitoring, and performing attribute division on the historical data according to gas types according to a preset time period;
2) preprocessing historical data corresponding to each type of gas in a preset time period to obtain a preprocessed sample, and defining data in the preprocessed sample as preprocessed data;
3) performing statistical analysis on the preprocessed samples corresponding to each type of gas to obtain a unitary frequency distribution sequence and a binary frequency distribution sequence of the preprocessed samples corresponding to each type of gas;
4) establishing an abnormal data diagnosis model according to historical data, a unitary frequency distribution sequence and a binary frequency distribution sequence of transformer oil chromatographic monitoring;
5) acquiring a group of new monitoring data, calculating the unary frequency and the binary frequency of the group of new monitoring data, and updating the unary frequency distribution sequence and the binary frequency distribution sequence in the step 3) according to the unary frequency and the binary frequency of the group of new monitoring data;
6) and judging the unitary frequency and the binary frequency of the monitoring data to be detected according to the unitary frequency distribution sequence and the binary frequency distribution sequence updated in the step 5), and substituting the unitary frequency and the binary frequency into an abnormal data diagnosis model for diagnosis.
2. The method for diagnosing the abnormal data of the transformer oil chromatogram online monitoring as claimed in claim 1, wherein the method for preprocessing the historical data corresponding to each type of gas in the step 2) comprises the following steps: and eliminating negative values in the historical data, merging continuous and unchangeable numerical values in a preset time period into a numerical value, and taking the sampling time of the last continuous and unchangeable numerical value as the sampling time of the merged numerical value.
3. The method for diagnosing the abnormal data of the transformer oil chromatogram in the online monitoring process, as claimed in claim 2, wherein the method for preprocessing the historical data corresponding to each gas type further comprises: and keeping the data value of each historical data to three decimal places.
4. The method for diagnosing the abnormal data of the transformer oil chromatogram in the on-line monitoring process as claimed in claim 1, wherein the step 3) is specifically as follows: dividing a plurality of data intervals, wherein the interval length of each data interval is the same, judging the data interval in which the preprocessed data in each preprocessed sample is located, sequencing the data intervals according to the sampling time of the preprocessed data, calculating unitary frequency and binary frequency, wherein the values of all the unitary frequencies of each preprocessed sample form a unitary frequency distribution sequence, and the values of all the binary frequencies of each preprocessed sample form a binary frequency distribution sequence;
the unitary frequency is the frequency of each data interval appearing in a preset time period, and the binary frequency is the frequency of two adjacent data intervals appearing repeatedly in the preset time period.
5. The method for diagnosing the abnormal data of the transformer oil chromatogram in the on-line monitoring process as claimed in claim 1, wherein the following steps are further carried out after the step 6): and analyzing the diagnosis result of the abnormal data diagnosis model, and optimizing the abnormal data diagnosis model according to the analysis result.
6. The method for diagnosing the abnormal data of the transformer oil chromatogram in the online monitoring manner as claimed in claim 1, wherein the abnormal data diagnosis model in the step 4) is established by using a logistic stewart regression model.
7. The transformer oil chromatographic on-line monitoring abnormal data diagnosis method as claimed in claim 1, wherein the gas types include hydrogen, methane, ethane, ethylene, acetylene, carbon monoxide, carbon dioxide, oxygen, nitrogen and total hydrocarbons.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811409813.0A CN109342630B (en) | 2018-11-23 | 2018-11-23 | Transformer oil chromatographic online monitoring abnormal data diagnosis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811409813.0A CN109342630B (en) | 2018-11-23 | 2018-11-23 | Transformer oil chromatographic online monitoring abnormal data diagnosis method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109342630A CN109342630A (en) | 2019-02-15 |
| CN109342630B true CN109342630B (en) | 2022-09-20 |
Family
ID=65317592
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811409813.0A Active CN109342630B (en) | 2018-11-23 | 2018-11-23 | Transformer oil chromatographic online monitoring abnormal data diagnosis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109342630B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114675010B (en) * | 2022-05-31 | 2022-09-13 | 卡松科技股份有限公司 | Intelligent analysis method for oxidation resistance of lubricating oil |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104677997A (en) * | 2015-02-02 | 2015-06-03 | 华北电力大学 | Transformer oil chromatography online monitoring differential early warning method |
| CN104764869A (en) * | 2014-12-11 | 2015-07-08 | 国家电网公司 | Transformer gas fault diagnosis and alarm method based on multidimensional characteristics |
| CN105044499A (en) * | 2015-07-01 | 2015-11-11 | 国家电网公司 | Method for detecting transformer state of electric power system equipment |
| CN108646124A (en) * | 2018-03-08 | 2018-10-12 | 南京工程学院 | A kind of oil chromatography online monitoring data variation tendency detection method based on small echo maximum |
-
2018
- 2018-11-23 CN CN201811409813.0A patent/CN109342630B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104764869A (en) * | 2014-12-11 | 2015-07-08 | 国家电网公司 | Transformer gas fault diagnosis and alarm method based on multidimensional characteristics |
| CN104677997A (en) * | 2015-02-02 | 2015-06-03 | 华北电力大学 | Transformer oil chromatography online monitoring differential early warning method |
| CN105044499A (en) * | 2015-07-01 | 2015-11-11 | 国家电网公司 | Method for detecting transformer state of electric power system equipment |
| CN108646124A (en) * | 2018-03-08 | 2018-10-12 | 南京工程学院 | A kind of oil chromatography online monitoring data variation tendency detection method based on small echo maximum |
Non-Patent Citations (3)
| Title |
|---|
| Analysis of Box Transformer Oil Chromatographic Abnormal Data in a 35kV Wind Farm;Wang Qinghao 等;《Applied Mechanics and Materials》;20130227;第310卷;395-398 * |
| Determination of phenol, m-cresol and o-cresol in transformer oil by HPLC method;Dijana Vrsaljko;《Electric Power Systems Research》;20120728;第93卷;24-31 * |
| 基于油色谱数据的变压器可靠度评估方法;张钰宁 等;《电力科学与工程》;20130731;第29卷(第7期);1-6 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109342630A (en) | 2019-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111508216B (en) | Intelligent early warning method for dam safety monitoring data | |
| WO2022252505A1 (en) | Device state monitoring method based on multi-index cluster analysis | |
| CN111460167A (en) | Method for positioning pollution discharge object based on knowledge graph and related equipment | |
| CN102486471A (en) | Oil chromatogram diagnosis method of transformer internal fault and defect degree | |
| CN109446189A (en) | A kind of technological parameter outlier detection system and method | |
| JP2000512766A (en) | Statistical pattern analysis method for partial discharge measurement in high voltage insulation | |
| CN108681801B (en) | Method and system for monitoring operation index in production process | |
| US20190056446A1 (en) | Method and system for assessment of fault severity, risk exposure, and gassing status for liquid-filled high-voltage apparatus | |
| CN105044499A (en) | Method for detecting transformer state of electric power system equipment | |
| CN103389430A (en) | Oil-immersed type transformer fault detection method based on Bayesian discrimination theory | |
| US20210073204A1 (en) | Method, system, and device for processing operation and maintenance data | |
| CN109342630B (en) | Transformer oil chromatographic online monitoring abnormal data diagnosis method | |
| CN116882735A (en) | Chemical risk correction early warning method based on process safety index | |
| CN116742799A (en) | Auxiliary power distribution monitoring and early warning system based on Internet of things technology | |
| CN116304957A (en) | On-line identification method for monitoring state mutation of power supply and transformation equipment | |
| CN113420162A (en) | Equipment operation chain state monitoring method based on knowledge graph | |
| KR20190053035A (en) | A method for defects | |
| CN117074852A (en) | Power distribution network electric energy monitoring and early warning management method and system | |
| CN116187861A (en) | Isotope-based water quality traceability monitoring method and related device | |
| CN115659271A (en) | Sensor abnormality detection method, model training method, system, device, and medium | |
| CN114740159A (en) | Natural gas energy metering component acquisition method and Internet of things system | |
| CN115600695A (en) | Fault diagnosis method of metering equipment | |
| CN115511238A (en) | Petrochemical device maintenance decision method and system based on operation condition evaluation | |
| CN113177362A (en) | Furnace temperature prediction method and device based on furnace temperature soft measurement model | |
| CN111651446B (en) | Data repetition frequency monitoring method based on big data |
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 |