CN110824292A - Power distribution network voltage loss fault intelligent identification method based on feature classification - Google Patents
Power distribution network voltage loss fault intelligent identification method based on feature classification Download PDFInfo
- Publication number
- CN110824292A CN110824292A CN201810893938.9A CN201810893938A CN110824292A CN 110824292 A CN110824292 A CN 110824292A CN 201810893938 A CN201810893938 A CN 201810893938A CN 110824292 A CN110824292 A CN 110824292A
- Authority
- CN
- China
- Prior art keywords
- voltage
- voltage loss
- fault
- loss
- trend
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000012549 training Methods 0.000 claims abstract description 17
- 238000012360 testing method Methods 0.000 claims abstract description 16
- 230000000694 effects Effects 0.000 claims abstract description 12
- 238000011156 evaluation Methods 0.000 claims abstract description 11
- 238000004140 cleaning Methods 0.000 claims abstract description 7
- 230000006837 decompression Effects 0.000 claims description 18
- 230000002159 abnormal effect Effects 0.000 claims description 9
- 238000013145 classification model Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 4
- 150000001875 compounds Chemical class 0.000 claims description 3
- 238000012937 correction Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 230000003247 decreasing effect Effects 0.000 claims description 2
- 230000007812 deficiency Effects 0.000 claims description 2
- 238000012417 linear regression Methods 0.000 claims description 2
- 230000002459 sustained effect Effects 0.000 claims description 2
- 230000009191 jumping Effects 0.000 claims 1
- 238000001514 detection method Methods 0.000 abstract description 4
- 230000002265 prevention Effects 0.000 abstract description 4
- 230000001737 promoting effect Effects 0.000 abstract 1
- 238000004422 calculation algorithm Methods 0.000 description 7
- 238000004458 analytical method Methods 0.000 description 6
- 238000007726 management method Methods 0.000 description 6
- 230000000630 rising effect Effects 0.000 description 4
- 238000007635 classification algorithm Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000005611 electricity Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006698 induction Effects 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000007637 random forest analysis Methods 0.000 description 2
- 230000005856 abnormality Effects 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 238000003066 decision tree Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000003203 everyday effect Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000007619 statistical method Methods 0.000 description 1
- 238000013024 troubleshooting Methods 0.000 description 1
- 230000003442 weekly effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/08—Locating faults in cables, transmission lines, or networks
- G01R31/088—Aspects of digital computing
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Theoretical Computer Science (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a power distribution network voltage loss fault intelligent identification method based on feature classification, which comprises the following steps: firstly, learning data and prediction data are obtained and data cleaning is carried out; then, constructing a fault evaluation index system based on the type of the voltage loss fault: forming a learning sample and a prediction sample by the sudden change voltage loss characteristic, the trend voltage loss characteristic and the jump voltage loss characteristic; dividing the learning sample into a training set and a testing set, learning a voltage loss fault recognition model by using the training set, and evaluating the model effect based on the testing set; and finally, taking the prediction sample as the input quantity of the voltage loss fault recognition model, outputting the suspected voltage loss fault coefficient of each user, and locking the suspected fault user. The invention refines the types of the voltage loss faults, including sudden voltage loss, trend voltage loss and jump voltage loss, analyzes the suspicion of the voltage loss faults of users aiming at different types, and can realize effective and accurate identification of the voltage loss faults, thereby promoting the fault detection management mode to the management level of 'prevention in advance and control in the process'.
Description
Technical Field
The invention relates to the technical field of power distribution network voltage loss fault identification, in particular to a power distribution network voltage loss fault intelligent identification method based on feature classification.
Background
The electric energy metering device in the three-phase power supply system has various faults due to factors such as fuse and the like, wherein the voltage loss fault is the most prominent, and the metering abnormality of the electric energy metering device caused by the faults not only causes electric quantity loss for electric power enterprises, but also brings certain difficulty for follow-up compensation of subsequent electric quantity. In order to avoid and reduce such losses, metering faults need to be found quickly and accurately, and timely and effective processing is needed to ensure reliable and normal operation of the metering device.
The traditional measurement fault finding means mainly comprise weekly inspection, first inspection, meter reading personnel on site meter reading, electricity consumption inspection and the like when the electricity quantity accounting is abnormal. The measures are strong in passivity and low in efficiency, measurement faults are long in duration and cannot be found, losses are caused to both power supply and power utilization, and even safety power supply is affected. With the progress of science and technology, the wide application of the power grid metering automation system can realize the real-time acquisition, monitoring, analysis and processing of the metering electric energy information, but the analysis of the metering faults still remains in a simple statistical analysis layer, and the defect of high false alarm rate exists.
In the existing research for identifying the voltage loss fault based on the data mining technology, although classification models such as a neural network are used, fault users and normal users can be classified at a certain level, the voltage loss is generally classified into one type for analysis, and a high rate of missing judgment exists. Based on the background, the invention provides an intelligent identification method for the voltage loss fault of the power distribution network based on feature classification. The method refines the types of the voltage loss faults, including three types of sudden change voltage loss, trend voltage loss and jump voltage loss; analyzing user faults according to different types of characteristics and possible causes of different fault types, such as voltage jump phenomenon caused by poor contact of a voltage loop; furthermore, suspicion of the user voltage loss fault and possible fault reasons are analyzed, accurate identification of the voltage loss fault can be effectively achieved, lower rate of missing judgment of an intelligent identification model is guaranteed, meanwhile, impending faults can be early warned, and therefore the fault detection management mode is improved to the management level of 'prevention in advance and control in the process'.
Disclosure of Invention
The invention relates to a power distribution network voltage loss fault intelligent identification method based on feature classification, which mainly comprises the following steps:
s1: acquiring learning data and prediction data and cleaning the data;
s2: constructing a decompression fault evaluation index system based on the decompression fault type, wherein the decompression fault evaluation index system comprises three characteristics of sudden change decompression, trend decompression and jump decompression, and a learning sample and a prediction sample are formed;
s3: dividing the learning sample into a training set and a testing set, learning a voltage loss fault recognition model by using the training set, and evaluating the model effect based on the testing set;
s4: and taking the prediction sample as the input quantity of the voltage loss fault identification model, outputting the suspected voltage loss fault coefficient of each user, and locking the suspected fault user.
The learning data and the prediction data comprise user files and load data, wherein the load data takes 24 integral point data values of the load data per day; the learning data must include normal user samples and voltage loss fault user samples, and the test data samples only include part of users in the power distribution network.
The data cleaning specifically comprises the following substeps:
s1.1: identifying missing values in the three-phase voltage data;
s1.2: based on the deficiency value, counting the integrity rate of three-phase voltage (including A phase, B phase and C phase, and three-phase three-wire users neglect B phase voltage), and removing users who do not meet the threshold value of the integrity rate, namely, not identifying the failure, wherein the threshold value can be set to 90% according to industry knowledge;
s1.3: calculating the normal voltage of the user to judge whether the wiring mode recorded in the file is wrong or not and correcting;
s1.4: comparing the user normal voltage with an abnormal threshold value, and performing mean value correction processing on abnormal data larger than the threshold value, wherein the threshold value can be set to be 1.2 times of a rated voltage value according to a professional;
s1.5: interpolating the missing value identified in step S1.1, with the interpolation rule: and taking the average value of 5 values before and after the missing value as an interpolation value.
The user normal voltage in step S1.3 represents a voltage value of the user in a normal state, and the voltage fluctuation is generally small in the power consumption process, so that the average voltage value of 2 months is taken as the normal voltage value.
The voltage loss fault types comprise three types of sudden change voltage loss, trend voltage loss and jump voltage loss, the voltage of the sudden change voltage loss suddenly drops in a change amplitude exceeding a certain range at a certain moment, and the change keeps small-amplitude fluctuation not recovered; the trend voltage loss refers to that the voltages in the period are all smaller than a rated value, and the general trend is shown; the jump voltage loss refers to the voltage which is repeated between a normal state and an abnormal state, and the duration time is longer than a certain threshold value.
The voltage loss fault evaluation system comprises three characteristics of sudden change voltage loss, trend voltage loss and jump voltage loss, wherein the sudden change voltage loss characteristics comprise voltage drop alarmAnd maximum number of sustained alarms for voltage drop
The trend voltage loss characteristic comprises an average voltage coefficient
And number of voltage trend decreases
The jump voltage-loss characteristic comprises the voltage sudden increase timesAnd number of voltage drops
。
The sudden voltage loss characteristic is voltage drop alarmMaximum number of continuous alarm times of voltage dropThe construction comprises the following steps:
s2.1: the voltage drop alarm
The calculation formula is as follows:
wherein,the content of the compound is as a percentage,
rated voltage for the user;
s2.2: the maximum continuous alarm frequency of the voltage drop
By statistics
The maximum number of consecutive 1.
Average voltage coefficient in the trend voltage loss characteristicAnd number of voltage trend decreases
The construction of the index comprises the following steps:
s2.3: the average voltage coefficient
Indicates from the current time point
(typically zero) start, move forwardIn the time period of a day, the ratio of the average value of the three-phase voltage to the normal voltage is calculated by the following formula:
wherein,
is the current time pointIs pushed forward toThe average voltage of the individual points in time,
is a normal voltage;
s2.4: the number of voltage trend decreasesIndicates from the current time point
Begin to push forward
In the time period of day, the voltage of each phase voltage at the previous time point and the next time pointThe number of times that the voltage of the point shows a downward trend compared to the voltage of the point is calculated as follows:
wherein,indicating the current time point
Is pushed forward to
Whether the voltage of the time point is lower than the first
The voltage of the time point is whether to show a decreasing trend.
Is a logic variable, and takes the value of 0 or 1:
wherein,
is the current time point
Is pushed forward toThe voltage variation trend of the time point can be represented by a voltage slope, namely, a unitary linear regression equation is constructed, and the least square method is utilized to solve the following steps:
wherein,
。
the jump voltage loss characteristic is the voltage sudden increase times
And number of voltage drops
The voltage of each phase voltage at the current moment in the period is compared with the voltage at the last moment in the period, and the rising or falling percentage of the voltage exceeds the rising threshold value
Or a falling threshold value
The number of times.
The under-voltage fault recognition model is trained by adopting different classification models aiming at a training set, and the fault condition of a test set is predicted according to the model obtained by training, so that the accuracy rate is further constructedAnd the misjudgment rate
Rate of sum and miss
Linear combination of (2)
To evaluate the recognition effect of the model:
and has the following components:
wherein,
the number of users is actually normal and is predicted to be normal at the same time;
the number of users who are actually normal but predicted to be faulty;
the number of users is predicted as the number of actual faults and the number of users with faults;
the number of users is predicted to be normal, though it is an actual failure.
According to the power distribution network voltage loss fault identification method based on feature classification, the voltage loss fault types including sudden change voltage loss, trend voltage loss and jump voltage loss are refined according to the voltage loss fault rule; the possibility of voltage loss of the user is analyzed according to different types from a plurality of angles of voltage change trend of the user in a normal state, voltage change trend of the user in a normal state to a fault state and voltage change trend of the user in the fault state, induction analysis is carried out according to fault reasons causing different voltage loss types, and the user fault condition can be accurately identified while impending faults can be early warned. Meanwhile, in order to ensure the recognition effect of the models, model training is carried out on different classification algorithms, the models are evaluated by utilizing the accuracy, the misjudgment rate and the misjudgment rate, and the models with high accuracy and low misjudgment rate are screened out, so that the finally formed voltage loss fault recognition model is optimal, and the fault detection management mode is promoted to the management level of 'prevention in advance and control in the event'.
Drawings
FIG. 1 is a main flow chart of an intelligent identification method for a power distribution network voltage loss fault based on feature classification.
Detailed Description
The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and examples.
In this embodiment, a power distribution network voltage loss fault intelligent recognition model based on feature classification makes a prediction on a suspected user of a voltage loss fault of a certain power grid company, and the model establishment and solving process shown in fig. 1 is combined, and the specific steps are as follows:
step 1: acquiring learning data and prediction data and cleaning the data;
step 2: constructing a decompression fault evaluation index system based on the decompression fault type, wherein the decompression fault evaluation index system comprises three characteristics of sudden change decompression, trend decompression and jump decompression, and a learning sample and a prediction sample are formed;
and step 3: dividing the learning sample into a training set and a testing set, learning a voltage loss fault recognition model by using the training set, and evaluating the model effect based on the testing set;
and 4, step 4: and taking the prediction sample as the input quantity of the voltage loss fault identification model, outputting the suspected voltage loss fault coefficient of each user, and locking the suspected fault user.
The step 1 is specifically explained as follows:
the data related to the embodiment is derived from user profile (user number, wiring mode, rated voltage) and load (user number, date and voltage) data of a distribution network in a region controlled by a certain power grid company from 2016 (9) to 2016 (11) and the load data takes 24 integral points of a day. The data of the voltage-loss fault users and part of normal users in 2016 (9 th month) to 2016 (10 th month) are learning data, and the data of part of users in 2016 (11 th month) is prediction data. The process can be described as:
s1.1: and (5) extracting learning data. Two types of data of voltage loss fault users and partial normal users in 2016 and 10 under the jurisdiction of the grid company are extracted from a metering automation system of the grid company.
S1.2: and extracting the prediction data. The 2016 month 11 portion of user data is extracted as forecast data.
S1.3: the data cleaning specifically comprises the following 5 sub-steps:
1) identifying missing values in three-phase voltages (including an A phase, a B phase and a C phase, and a three-phase three-wire user ignores a B phase voltage) in the load data, wherein the missing values comprise data with the numerical values of-1000 and NA;
2) based on the missing values, respectively counting the three-phase voltage integrity rate of each user, and removing the users which do not meet the threshold value of the integrity rate, namely, not identifying the faults of the users, wherein the threshold value is set to be 90% according to industry knowledge;
3) respectively calculating the voltage mean value of each user within 2 months as normal voltage, judging the wiring mode of the user according to the section of the normal voltage, and comparing whether the wiring mode recorded in the file is wrong or not and correcting;
4) comparing the user normal voltage with an abnormal threshold value, and performing mean value correction processing on abnormal data larger than the threshold value, wherein the threshold value is set to be 1.2 times of a rated voltage value according to a professional;
5) interpolating the missing values identified in step 1), wherein the interpolation rule is as follows: and taking the average value of 5 values before and after the missing value as an interpolation value.
The step 2 is specifically explained as follows:
s2.1: aiming at the washed learning and prediction data, the voltage drop alarm index of each user every day is calculated：
Wherein,
the content of the compound is as a percentage,
rated for the user.
S2.2: by means of statistical meters
The maximum number of continuous 1 is obtained, and the maximum continuous alarm number of voltage drop of each user per day is obtained
。
S2.3: calculate each userAverage voltage coefficient of each phase voltage of day
：
Wherein,
，
is the current time point
Is pushed forward to
The average voltage of the individual points in time,
is a normal voltage.
S2.4: calculate each user
Number of voltage trend reductions of each phase voltage of day
：
Wherein,
，
。
s2.5: calculate each user
Number of voltage surges of each phase voltage of day
And number of voltage drops
I.e. by
The rising or falling percentage of the voltage at the current time point in the day is more than the rising threshold value compared with the voltage at the last time point
Or a falling threshold value
Number of times, here
And
all values are 0.2.
S2.6: learning samples and prediction samples are formed. For the learning sample, labeling each record according to whether the user has a voltage loss fault at a certain date, wherein 1 represents voltage loss, and 0 represents no voltage loss, and finally forming the learning sample and the prediction sample as shown in table 1 and table 2 by way of example:
watch (A)Learning sample example
Table 2 prediction sample example
The step 3 is specifically explained as follows:
s3.1: and dividing a training set and a testing set according to the learning sample, wherein the ratio is 7: 3.
S3.2: and constructing a classification model by using different classification algorithms by using a training set for training, wherein the classification model comprises a BP neural network algorithm, a CART decision tree algorithm, a naive Bayes algorithm and a random forest algorithm.
S3.3: and evaluating the effect of each model by using the test set. Taking the test set as the input of each model, predicting the suspected probability of the voltage loss fault of each user in the test set in a certain day, and taking 0.5 as a threshold value for dividing whether the voltage is lost; definition of，
，Calculating the evaluation value of each model
The results are shown in table 3 below:
in combination with services, generally, a classification model has excellent classification effect, which is reflected in that the classification model has higher accuracy, lower rate of missing judgment and lower rate of false judgment, so that the comprehensive evaluation quantity of the classification model, the high accuracy, the low rate of missing judgment and the low rate of false judgment is obtained
The larger the size, the better the classification effect of the model. As can be known by combining the table 3, the classification effect of the model constructed by the random forest algorithm is better than that of other algorithms, so that the algorithm is selected to construct the voltage-loss fault identification model for fault prediction.
The step 4 is specifically explained as follows:
and taking the prediction sample as the input quantity of the voltage loss fault recognition model, outputting the suspected voltage loss fault coefficient of each user, and locking suspected users to perform accurate fault troubleshooting.
By combining the analysis, the power distribution network voltage loss fault identification method based on the characteristic classification provided by the invention refines the voltage loss fault types including sudden voltage loss, trend voltage loss and jump voltage loss according to the voltage loss fault rule; the possibility of voltage loss of the user is analyzed according to different types from a plurality of angles of voltage change trend of the user in a normal state, voltage change trend of the user in a normal state to a fault state and voltage change trend of the user in the fault state, induction analysis is carried out according to fault reasons causing different voltage loss types, and the user fault condition can be accurately identified while impending faults can be early warned. Meanwhile, in order to ensure the recognition effect of the models, model training is carried out on different classification algorithms, the models are evaluated by utilizing the accuracy, the misjudgment rate and the misjudgment rate, and the models with high accuracy and low misjudgment rate are screened out, so that the finally formed voltage loss fault recognition model is optimal, and the fault detection management mode is promoted to the management level of 'prevention in advance and control in the event'.
Claims (10)
1. A power distribution network voltage loss fault intelligent identification method based on feature classification is characterized by comprising the following steps:
s1: acquiring learning data and prediction data and cleaning the data;
s2: constructing a decompression fault evaluation index system based on the decompression fault type, wherein the decompression fault evaluation index system comprises three characteristics of sudden change decompression, trend decompression and jump decompression, and a learning sample and a prediction sample are formed;
s3: dividing the learning sample into a training set and a testing set, learning a voltage loss fault recognition model by using the training set, and evaluating the model effect based on the testing set;
s4: and taking the prediction sample as the input quantity of the voltage loss fault identification model, outputting the suspected voltage loss fault coefficient of each user, and locking the suspected fault user.
2. The intelligent identification method for the voltage-loss fault of the power distribution network based on the characteristic classification as claimed in claim 1, wherein the learning data and the prediction data comprise user files and load data; the learning data must include normal user samples and voltage loss fault user samples, and the test data samples only include part of users in the power distribution network.
3. The intelligent identification method for the voltage loss fault of the power distribution network based on the characteristic classification as claimed in claim 1, wherein the data cleaning specifically comprises the following substeps:
s1.1: identifying missing values in the three-phase voltage data;
s1.2: based on the deficiency value, counting the integrity rate of three-phase voltage (including A phase, B phase and C phase, and three-phase three-wire users neglect B phase voltage), and removing users who do not meet the threshold value of the integrity rate, namely, not identifying the failure, wherein the threshold value can be set to 90% according to industry knowledge;
s1.3: calculating the normal voltage of the user to judge whether the wiring mode recorded in the file is wrong or not and correcting;
s1.4: comparing the user normal voltage with an abnormal threshold value, and performing mean value correction processing on abnormal data larger than the threshold value, wherein the threshold value can be set to be 1.2 times of a rated voltage value according to a professional;
s1.5: interpolating the missing value identified in step S1.1, with the interpolation rule: and taking the average value of 5 values before and after the missing value as an interpolation value.
4. The intelligent identification method for the voltage loss fault of the power distribution network based on the feature classification as claimed in claim 3, wherein the normal voltage of the user in the step S1.3 represents the voltage value of the user in the normal state, and the voltage fluctuation is generally small in the power utilization process, so that the average voltage value of 2 months is taken as the normal voltage value.
5. The intelligent identification method for the voltage loss faults of the power distribution network based on the characteristic classification as claimed in claim 1, wherein the voltage loss fault types comprise three types of sudden change voltage loss, trend voltage loss and jump voltage loss, the sudden change voltage loss refers to that the voltage suddenly drops in a change amplitude exceeding a certain range at a certain moment, and the change keeps small-amplitude fluctuation not recovered; the trend voltage loss refers to that the voltages in the period are all smaller than a rated value, and the general trend is shown; the jump voltage loss refers to the voltage which is repeated between a normal state and an abnormal state, and the duration time is longer than a certain threshold value.
6. The intelligent identification method for the voltage loss faults of the power distribution network based on the characteristic classification as claimed in claim 1, wherein the voltage loss fault evaluation system comprises three characteristics of sudden change voltage loss, trend voltage loss and jump voltage loss, and the sudden change voltage loss characteristics comprise voltage drop alarmAnd maximum number of sustained alarms for voltage drop
The trend voltage loss characteristic comprises an average voltage coefficient
And number of voltage trend decreases
The jump voltage-loss characteristic comprises the voltage sudden increase times
And number of voltage drops
。
7. The intelligent identification method for the voltage loss faults of the power distribution network based on the characteristic classification as claimed in claim 1 or 6, wherein the voltage drop alarm in the sudden change voltage loss characteristicAnd maximum continuation notice of voltage dropNumber of alarm times
The construction comprises the following steps:
s2.1: the voltage drop alarmThe calculation formula is as follows:
wherein,the content of the compound is as a percentage,
rated voltage for the user;
s2.2: the maximum continuous alarm frequency of the voltage dropBy statisticsThe maximum number of consecutive 1.
8. The method for intelligently identifying the voltage-loss fault of the power distribution network based on the feature classification as claimed in claim 1 or 6, wherein the average voltage coefficient in the trend voltage-loss featureAnd number of voltage trend decreases
The construction of the index comprises the following steps:
s2.3: the average voltage coefficient
Indicates from the current time point
(typically zero) start, move forward
The ratio of the average value of the voltage to the normal voltage in the time period of days is calculated by the following formula:
wherein,
is the current time pointIs pushed forward to
The average voltage of the individual points in time,
is a normal voltage;
s2.4: the number of voltage trend decreases
Indicates from the current time point
Begin to push forward
The number of times that the voltage at the previous time point is in a descending trend compared with the voltage at the next time point in the time period of the day is calculated according to the following formula:
wherein,indicating the current time point
Is pushed forward to
Whether the voltage of the time point is lower than the firstThe voltage of the time point is whether to show a decreasing trend.
Is a logic variable, and takes the value of 0 or 1:
wherein,
is the current time point
Is pushed forward to
The voltage variation trend of the time point canThe voltage slope representation is considered, namely, the voltage slope representation is obtained by constructing a unary linear regression equation and solving by using a least square method:
wherein,
。
9. the intelligent identification method for the voltage loss faults of the power distribution network based on the feature classification as claimed in claim 1 or 6, wherein the number of voltage sudden increases in the jumping voltage loss features
And number of voltage drops
Indicating that the percentage of rise or fall of the voltage at the current time in the cycle compared with the voltage at the previous time exceeds the rise threshold
Or a falling threshold valueThe number of times.
10. The intelligent identification method for the voltage loss faults of the power distribution network based on the feature classification as claimed in claim 1, wherein the voltage loss fault identification model is trained by adopting different classification models aiming at a training set, and the fault conditions of a test set are predicted according to the model obtained by training, so that the accuracy is further constructed
False alarm rate
Rate of missed report
Linear combination of (2)
To evaluate the recognition effect of the model:
and has the following components:
wherein,
the number of users is actually normal and is predicted to be normal at the same time;
the number of users who are actually normal but predicted to be faulty;
the number of users is predicted as the number of actual faults and the number of users with faults;
the number of users is predicted to be normal, though it is an actual failure.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810893938.9A CN110824292A (en) | 2018-08-08 | 2018-08-08 | Power distribution network voltage loss fault intelligent identification method based on feature classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810893938.9A CN110824292A (en) | 2018-08-08 | 2018-08-08 | Power distribution network voltage loss fault intelligent identification method based on feature classification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110824292A true CN110824292A (en) | 2020-02-21 |
Family
ID=69533875
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810893938.9A Pending CN110824292A (en) | 2018-08-08 | 2018-08-08 | Power distribution network voltage loss fault intelligent identification method based on feature classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110824292A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113140096A (en) * | 2021-04-23 | 2021-07-20 | 广东电网有限责任公司 | Method, device, equipment and storage medium for monitoring and judging station loss of voltage |
| CN114971264A (en) * | 2022-05-23 | 2022-08-30 | 广东电网有限责任公司广州供电局 | Method for identifying different voltage levels of live-line work of power grid equipment in random forest |
| CN116500379A (en) * | 2023-05-15 | 2023-07-28 | 珠海中瑞电力科技有限公司 | Accurate positioning method for voltage drop of STS device |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
| CN106650797A (en) * | 2016-12-07 | 2017-05-10 | 广东电网有限责任公司江门供电局 | Distribution network electricity stealing suspected user intelligent recognition method based on integrated ELM (Extreme Learning Machine) |
| CN107895137A (en) * | 2017-10-10 | 2018-04-10 | 浙江工业大学 | SAR image target interbehavior recognition methods based on coupled hidden markov model |
| CN108110893A (en) * | 2017-11-22 | 2018-06-01 | 国电南瑞科技股份有限公司 | A kind of electric network fault grade automatic identifying method of transmission ＆ distribution collaboration |
| CN108227678A (en) * | 2018-01-04 | 2018-06-29 | 广东电网有限责任公司电力科学研究院 | A kind of power distribution network no-voltage fault diagnostic method and device based on metering automation system |
-
2018
- 2018-08-08 CN CN201810893938.9A patent/CN110824292A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106555788A (en) * | 2016-11-11 | 2017-04-05 | 河北工业大学 | Application of the deep learning based on Fuzzy Processing in hydraulic equipment fault diagnosis |
| CN106650797A (en) * | 2016-12-07 | 2017-05-10 | 广东电网有限责任公司江门供电局 | Distribution network electricity stealing suspected user intelligent recognition method based on integrated ELM (Extreme Learning Machine) |
| CN107895137A (en) * | 2017-10-10 | 2018-04-10 | 浙江工业大学 | SAR image target interbehavior recognition methods based on coupled hidden markov model |
| CN108110893A (en) * | 2017-11-22 | 2018-06-01 | 国电南瑞科技股份有限公司 | A kind of electric network fault grade automatic identifying method of transmission ＆ distribution collaboration |
| CN108227678A (en) * | 2018-01-04 | 2018-06-29 | 广东电网有限责任公司电力科学研究院 | A kind of power distribution network no-voltage fault diagnostic method and device based on metering automation system |
Non-Patent Citations (1)
| Title |
|---|
| 张长胜 等: "基于粗糙集理论的恶意代码特征分析", 《电子科技》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113140096A (en) * | 2021-04-23 | 2021-07-20 | 广东电网有限责任公司 | Method, device, equipment and storage medium for monitoring and judging station loss of voltage |
| CN113140096B (en) * | 2021-04-23 | 2022-09-30 | 广东电网有限责任公司 | Method, device, equipment and storage medium for monitoring and judging station loss of voltage |
| CN114971264A (en) * | 2022-05-23 | 2022-08-30 | 广东电网有限责任公司广州供电局 | Method for identifying different voltage levels of live-line work of power grid equipment in random forest |
| CN116500379A (en) * | 2023-05-15 | 2023-07-28 | 珠海中瑞电力科技有限公司 | Accurate positioning method for voltage drop of STS device |
| CN116500379B (en) * | 2023-05-15 | 2024-03-08 | 珠海中瑞电力科技有限公司 | Accurate positioning method for voltage drop of STS device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109583680B (en) | Power stealing identification method based on support vector machine | |
| Correcher et al. | Intermittent failure dynamics characterization | |
| CN110414155A (en) | A kind of detection of fan part temperature anomaly and alarm method with single measuring point | |
| CN112561736A (en) | Fault diagnosis system and method for relay protection device of intelligent substation | |
| CN111191878A (en) | Abnormal analysis based station area and electric energy meter state evaluation method and system | |
| CN110824292A (en) | Power distribution network voltage loss fault intelligent identification method based on feature classification | |
| CN106022951A (en) | Electricity consumption abnormity analysis method and apparatus | |
| CN109977535A (en) | Line loss abnormity diagnosis method, device, equipment and readable storage medium | |
| CN113420992A (en) | Power system network risk assessment method and system | |
| CN105588995A (en) | Line loss abnormity detection method for electric power metering automation system | |
| CN113032454A (en) | Interactive user power consumption abnormity monitoring and early warning management cloud platform based on cloud computing | |
| CN109491339B (en) | Big data-based substation equipment running state early warning system | |
| CN115454778A (en) | Intelligent monitoring system for abnormal time sequence indexes in large-scale cloud network environment | |
| CN109542742A (en) | Database server hardware health evaluating method based on expert model | |
| CN114997313B (en) | Abnormality detection method for ocean on-line monitoring data | |
| CN111612019A (en) | Method for identifying and analyzing fault abnormality of intelligent electric meter based on big data model | |
| CN111275295A (en) | Distributed photovoltaic fault diagnosis method based on inverse distance weight interpolation | |
| CN115809805A (en) | Power grid multi-source data processing method based on edge calculation | |
| CN116882804A (en) | Intelligent power monitoring method and system | |
| CN102621502A (en) | Fault monitoring and forecasting method of direct current (DC)-direct current (DC) power source system | |
| CN110968703A (en) | Method and system for constructing abnormal metering point knowledge base based on LSTM end-to-end extraction algorithm | |
| CN117391357B (en) | Scheduling self-checking system for power grid scheduling operation management based on big data | |
| CN118132931A (en) | Daily electrical appliance sub-health detection method based on machine learning | |
| CN117703690A (en) | Wind generating set health state assessment method and system | |
| CN112418662A (en) | Power distribution network operation reliability analysis method using artificial neural network |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200221 |
|
| WD01 | Invention patent application deemed withdrawn after publication |