CN111999695B - State evaluation and abnormity diagnosis method for metering device of transformer substation - Google Patents
State evaluation and abnormity diagnosis method for metering device of transformer substation Download PDFInfo
- Publication number
- CN111999695B CN111999695B CN202011168103.0A CN202011168103A CN111999695B CN 111999695 B CN111999695 B CN 111999695B CN 202011168103 A CN202011168103 A CN 202011168103A CN 111999695 B CN111999695 B CN 111999695B
- Authority
- CN
- China
- Prior art keywords
- state evaluation
- time
- metering device
- diagnosis
- abnormal
- 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
- 238000003745 diagnosis Methods 0.000 title claims abstract description 91
- 238000011156 evaluation Methods 0.000 title claims abstract description 78
- 238000000034 method Methods 0.000 title claims abstract description 46
- 239000013598 vector Substances 0.000 claims abstract description 79
- 230000005856 abnormality Effects 0.000 claims abstract description 45
- 238000012549 training Methods 0.000 claims abstract description 24
- 230000002159 abnormal effect Effects 0.000 claims description 71
- 238000004364 calculation method Methods 0.000 claims description 24
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 238000007635 classification algorithm Methods 0.000 claims description 3
- 238000012423 maintenance Methods 0.000 abstract description 6
- 230000006870 function Effects 0.000 description 12
- 230000005611 electricity Effects 0.000 description 7
- 230000009286 beneficial effect Effects 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000012360 testing method Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000005070 sampling Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000010248 power generation Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R35/00—Testing or calibrating of apparatus covered by the other groups of this subclass
- G01R35/04—Testing or calibrating of apparatus covered by the other groups of this subclass of instruments for measuring time integral of power or current
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
The invention relates to a transformer substation metering device state evaluation and abnormity diagnosis method, which comprises the following steps: acquiring electrical parameter data, and calculating an electrical parameter deviation degree or an unbalance degree index at each moment based on the electrical parameter data; acquiring the statistical characteristics of the electrical parameter deviation degree or unbalance degree indexes in different time periods; acquiring data integrity characteristics of the current symbol characteristics and the electrical parameter data at different time intervals, and establishing a characteristic vector at different time intervals based on the statistical characteristics, the current symbol characteristics and the data integrity characteristics; respectively training a time-interval state evaluation and abnormality diagnosis model by using the characteristic vectors in the time-interval training data set and the corresponding states of the transformer substation metering devices; inputting the characteristic vector of the substation metering device to be subjected to state evaluation and abnormality diagnosis into a state evaluation and abnormality diagnosis model in a corresponding time period and then outputting a diagnosis result; the method can timely and accurately find the abnormity and diagnose the abnormity type, and provides powerful support for field maintenance.
Description
Technical Field
The invention belongs to the technical field of intelligent power grid metering device abnormity diagnosis, and particularly relates to a transformer substation metering device state evaluation and abnormity diagnosis method.
Background
The transformer substation in the core part of the power grid not only undertakes power dispatching, but also is responsible for metering and settling the electric energy among various regions and between power generation and power utilization. Under the condition of huge electricity consumption, any abnormity of the metering device of the transformer substation causes immeasurable influence and loss on inter-regional electricity charge settlement. The existing transformer substation monitoring system pays more attention to the working state of a transformer, monitoring and management are carried out on the running state of a metering device of a transformer substation without necessary technical means, data analysis and field inspection are mainly carried out manually, the efficiency is very low, and the operation and maintenance requirements of a power grid cannot be met at all. Because the working principle of the substation metering device is similar to that of a high-voltage user, the state evaluation and the abnormality diagnosis of the substation metering device based on the technical means are mostly improved and optimized on the basis of a method for identifying abnormal electricity consumption of the high-voltage user. Because abnormal electricity utilization behaviors are a long-term process and mainly influenced by human factors, the abnormal electricity utilization behaviors are different from metering devices of the transformer substation to a certain extent. Therefore, the existing method has the following defects:
1) only normal and abnormal conditions can be identified, and the reason of the abnormal condition cannot be diagnosed, which is not beneficial to on-site investigation and maintenance;
2) the timeliness of the abnormity diagnosis is poor, and the abnormity diagnosis is usually performed on the basis of data of a longer time period, so that the abnormity cannot be found and eliminated in time;
3) the method has higher requirements on data quality, and diagnosis cannot be carried out once data are missing;
4) often, only one abnormal type can be output, but for example, voltage loss or power failure can coexist with other abnormal types, and a single diagnosis result is not beneficial to on-site overhaul and maintenance.
Therefore, the existing abnormal electricity utilization detection substation metering device state evaluation and abnormal diagnosis method cannot meet the requirements of power grid operation and management. In order to ensure the accuracy of the metering of the transformer substation, a state evaluation and abnormity diagnosis method for a metering device of the transformer substation is urgently needed, and the abnormity can be quickly positioned and accurately checked after the metering device is abnormal.
Disclosure of Invention
The invention provides a method for evaluating the state and diagnosing the abnormality of a metering device of a transformer substation aiming at the technical problems in the prior art, and solves the problem that the state evaluation of the metering device of the transformer substation is inaccurate in the prior art.
The technical scheme for solving the technical problems is as follows: a transformer substation metering device state evaluation and abnormity diagnosis method comprises the following steps:
step 1, acquiring electrical parameter data, and calculating an electrical parameter deviation degree or an unbalance degree index at each moment based on the electrical parameter data; acquiring the statistical characteristics of the electrical parameter deviation degree or unbalance degree indexes in different time periods;
step 2, acquiring current symbol characteristics and data integrity characteristics of the electrical parameter data at different time intervals, and establishing characteristic vectors at different time intervals based on the statistical characteristics, the current symbol characteristics and the data integrity characteristics;
the current sign characteristics represent different proportions of split-phase current signs;
step 3, respectively training a state evaluation and abnormity diagnosis model of each time period by using the characteristic vector in the training data set of each time period and the state of the corresponding substation metering device;
and 4, inputting the characteristic vector of the substation metering device to be subjected to state evaluation and abnormity diagnosis into the trained state evaluation and abnormity diagnosis model in the corresponding time period, and outputting a diagnosis result by the state evaluation and abnormity diagnosis model. .
The invention has the beneficial effects that: the method comprises the steps that a state evaluation and abnormity diagnosis model is established on the basis of the electrical parameter data of a transformer substation metering device in one day, main abnormity types are better covered, data phenomena expressed by abnormity of the metering device are further highlighted by extracting statistical characteristics, and the statistical characteristics comprise electrical parameter statistical characteristics, current sign characteristics and data integrity characteristics, so that evaluation and abnormity diagnosis of the metering state are accurately completed; the model can be used for timely and accurately finding and diagnosing the abnormal type, provides powerful support for field maintenance, and improves the operation efficiency of the power company.
On the basis of the technical scheme, the invention can be further improved as follows.
Further, the electrical parameter data in step 1 includes: split-phase voltage, split-phase current, split-phase active power and total active power, split-phase power factor and total power factor.
Further, the electrical parameter deviation or imbalance indicator in step 1 includes: voltage deviation, voltage unbalance, current deviation, current unbalance, active power difference, active power unbalance and phase angle unbalance.
Further, the voltage deviation degree at the time iVD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C three-phase voltages at the time i, and min () representing a minimum function;
the voltage unbalance degree of the time iVB i The calculation formula of (2) is as follows:
max () represents a get max function;
the current deviation degree at time iCD i The calculation formula of (2) is as follows:
wherein,a, B, C represent the currents of three phases at time i, respectively, and avg () represents the averaging function;
the current imbalance at time iCB i The calculation formula of (2) is as follows:
the power difference at time iPD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the active power of the three phases at time i,representing the total active power at time i, abs () represents an absolute value function,represents a summation;
the active imbalance at time iPB i The calculation formula of (2) is as follows:
the degree of phase angle imbalance at time iFB i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the power factor of the three phases at time i,representing the total power factor at time i, arccos () represents the inverse cosine function.
Further, the statistical characteristics in the step 1 include a minimum value, a lower quantile, a middle quantile, an upper quantile, a maximum value and an average value;
the time periods of the sub-periods include daytime periods and nighttime periods.
Further, the calculation formula of the current sign characteristic in the step 2 is as follows:
wherein CS is a current symbol characteristic, N is the number of data items collected in a set time period,a, B, C represent the currents of the three phases at time i, respectively, sign () represents a sign function defined as:。
further, the state of the substation metering device in the step 3 includes: normal, voltage loss, current loss, phase loss, voltage unbalance, current unbalance, power failure, abnormal power factor and wrong wiring;
the classification algorithm of the state evaluation and abnormity diagnosis model adopts an XGboost algorithm, and the state evaluation and abnormity diagnosis model comprises a night XGboost model and a daytime XGboost model.
Further, the step 4 comprises:
after the feature vectors are input into the trained night XGboost model or the daytime XGboost model in the corresponding time period, the night XGboost model or the daytime XGboost model respectively outputs probability vectors representing the probabilities of the predicted types, and the probability vectors respectively output by the night XGboost model or the daytime XGboost model are added according to the element mode to obtain probabilities and vectors;
and when the probability of the normal type in the probability and the vector is the maximum, performing state evaluation on the substation metering device to be subjected to state evaluation and abnormal diagnosis, otherwise, performing abnormal diagnosis on the substation metering device to be subjected to state evaluation and abnormal diagnosis.
Further, the process of performing state evaluation on the substation metering device to be subjected to state evaluation and abnormality diagnosis includes:
and when the probabilities of normal types in the probability vectors respectively output by the night XGboost model or the daytime XGboost model are the maximum, evaluating the state of the transformer substation metering device to be subjected to state evaluation and abnormality diagnosis to be healthy, and otherwise evaluating the transformer substation metering device to be subjected to state evaluation and abnormality diagnosis to be sub-healthy.
Further, the process of performing abnormality diagnosis on the substation metering device to be subjected to state evaluation and abnormality diagnosis includes:
and when the abnormal type with the maximum positive probability value in the probability vectors respectively output by the night XGboost model or the daytime XGboost model is the same, the abnormal type of the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis is diagnosed as the abnormal type with the maximum probability and the maximum abnormal type, otherwise, the abnormal type is diagnosed as the two abnormal types with the maximum probability and the maximum abnormal type.
The beneficial effect of adopting the further scheme is that: by dividing the natural day into the daytime and the nighttime for classification and judgment respectively, the method not only can further improve the timeliness of the method for diagnosing the metering abnormality, but also can ensure that the state evaluation and the abnormality diagnosis result are more stable and reliable; the invention also has the capability of diagnosing at most two abnormal types in one natural day, and is easier to capture the abnormal types which have short duration time and possibly coexist with other abnormalities, such as voltage loss, power failure and the like.
Drawings
Fig. 1 is a schematic diagram of a framework of a method for evaluating a state and diagnosing an abnormality of a metering device of a transformer substation according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a feature vector structure according to an embodiment of the present invention;
FIG. 3 is a flow chart of a state estimation and anomaly diagnosis model training process according to an embodiment of the present invention;
fig. 4 is a flow chart of status evaluation and anomaly diagnosis according to an embodiment of the present invention.
Detailed Description
The principles and features of this invention are described below in conjunction with the following drawings, which are set forth by way of illustration only and are not intended to limit the scope of the invention.
The invention provides a transformer substation metering device state evaluation and abnormity diagnosis method, which comprises the following steps:
step 1, acquiring electrical parameter data, and calculating an electrical parameter deviation degree or an unbalance degree index at each moment based on the electrical parameter data; and acquiring the statistical characteristics of the electrical parameter deviation degree or the unbalance degree index in different time periods.
And 2, acquiring data integrity characteristics of the current symbol characteristics and the electrical parameter data in different time periods, and establishing a characteristic vector in different time periods based on the statistical characteristics, the current symbol characteristics and the data integrity characteristics.
The current sign characteristics indicate the different proportions of the split-phase current sign.
And 3, respectively training the state evaluation and abnormality diagnosis models of the time intervals by using the characteristic vectors in the training data sets of the time intervals and the states of the corresponding substation metering devices.
And 4, inputting the characteristic vector of the substation metering device to be subjected to state evaluation and abnormality diagnosis into the trained state evaluation and abnormality diagnosis model in the corresponding time period, and outputting a diagnosis result by the state evaluation and abnormality diagnosis model.
The invention provides a method for evaluating and diagnosing the state of a metering device of a transformer substation, which is characterized in that a state evaluation and abnormality diagnosis model is established on the basis of the electrical parameter data of the metering device of the transformer substation in one day, so that the main abnormality types are better covered, the data phenomenon shown by the abnormality of the metering device is further highlighted by extracting statistical characteristics, and the statistical characteristics comprise electrical parameter statistical characteristics, current sign characteristics and data integrity characteristics, so that the evaluation and abnormality diagnosis of the metering state are accurately finished; the model can be used for timely and accurately finding and diagnosing the abnormal type, provides powerful support for field maintenance, and improves the operation efficiency of the power company.
Example 1
Embodiment 1 provided by the present invention is an embodiment of a method for evaluating a state and diagnosing an abnormality of a substation metering device provided by the present invention, and as shown in fig. 1, is a schematic diagram of a framework of the method for evaluating a state and diagnosing an abnormality of a substation metering device provided by the embodiment of the present invention, as can be seen from fig. 1, the embodiment includes:
step 1, acquiring electrical parameter data, and calculating an electrical parameter deviation degree or an unbalance degree index at each moment based on the electrical parameter data; and acquiring the statistical characteristics of the electrical parameter deviation degree or the unbalance degree index in different time periods.
Specifically, the electrical parameter data includes: split-phase voltage, split-phase current, split-phase active power and total active power, split-phase power factor and total power factor.
The electrical parameter data is time-spanned by natural days, and the preferred acquisition frequency of all electrical parameters is 96 sampled values per natural day.
The electrical parameter deviation or imbalance indicators include: and 7 items of voltage deviation, voltage unbalance, current deviation, current unbalance, active power difference, active power unbalance, phase angle unbalance and the like.
In particular, the voltage deviation at time iVD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the voltages of the three phases at time i, and min () representing the function of taking the minimum value.
Voltage unbalance degree at time iVB i The calculation formula of (2) is as follows:
max () represents a maximum function.
Current deviation degree at time iCD i The calculation formula of (2) is as follows:
wherein,representing A, B, C the currents of the three phases at time i, respectively, and avg () representing the averaging function.
Current imbalance at time iCB i The calculation formula of (2) is as follows:
power difference at time iPD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the active power of the three phases at time i,representing the total active power at time i, abs () represents an absolute value function,indicating a summation.
Active imbalance at time iPB i The calculation formula of (2) is as follows:
degree of phase angle imbalance at time iFB i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the power factor of the three phases at time i,representing the total power factor at time i, arccos () represents the inverse cosine function.
The time interval division can be performed according to the electricity utilization characteristics of the user, for example, a natural day is divided into a night time interval and a day time interval, wherein the night time interval comprises 0 to 7 points and 20 to 23 points, and the day time interval comprises 8 to 19 points.
By dividing the natural day into the daytime and the nighttime for classification and judgment respectively, the method not only can further improve the timeliness of the method for diagnosing the metering abnormality, but also can ensure that the state evaluation and the abnormality diagnosis result are more stable and reliable; the invention also has the capability of diagnosing at most two abnormal types in one natural day, and is easier to capture the abnormal types which have short duration time and possibly coexist with other abnormalities, such as voltage loss, power failure and the like.
Specifically, the statistical characteristics of the electrical parameter deviation or unbalance index include a minimum value, a lower quantile, a middle quantile, an upper quantile, a maximum value, and an average value.
The method comprises the steps of calculating electrical parameter statistical characteristics of 6 statistical characteristics including a minimum value, a lower quantile point, a middle quantile point, an upper quantile point, a maximum value, an average value and the like of night time and day time aiming at 7 electrical parameter indexes including a voltage deviation degree, a voltage unbalance degree, a current deviation degree, a current unbalance degree, an active power difference, an active power unbalance degree, a phase angle unbalance degree and the like of the night time and the day time respectively.
And 2, acquiring data integrity characteristics of the current symbol characteristics and the electrical parameter data in different time periods, and establishing a characteristic vector in different time periods based on the statistical characteristics, the current symbol characteristics and the data integrity characteristics. Fig. 2 is a schematic diagram of a feature vector structure according to an embodiment of the present invention.
The current sign characteristics indicate the different proportions of the split-phase current sign.
Specifically, the calculation formula of the current sign characteristic is as follows:
wherein CS is a current symbol feature, N is a number of data items collected within a set time period, sign () represents a symbol function, and is defined as:。
and extracting data integrity characteristics of the night time period and the day time period, and respectively counting the proportion of missing data items in the night time period and the day time period in the same natural day. The statistics of the missing data items cover all split-phase voltages, split-phase currents, split-phase active power and total active power, split-phase power factors and total power factors in the time period.
And 3, respectively training the state evaluation and abnormality diagnosis models of the time intervals by using the characteristic vectors in the training data sets of the time intervals and the states of the corresponding substation metering devices.
The state of the substation metering device in the step 3 comprises the following steps: and 9 types including normal, voltage loss, current loss, phase loss, voltage unbalance, current unbalance, power failure, power factor abnormity, wrong wiring and the like, wherein 1 type is normal, and 8 types are abnormal.
The classification algorithm of the state evaluation and abnormality diagnosis model preferably adopts an XGBoost (extreme gradient spanning tree) algorithm, but is not limited to the XGBoost algorithm, and the state evaluation and abnormality diagnosis model includes a nighttime XGBoost model and a daytime XGBoost model.
In a specific implementation, as shown in fig. 3, which is a flowchart of a state evaluation and abnormality diagnosis model training provided by the embodiment of the present invention, as can be seen from fig. 3, step 3 may include:
The invention supports the diagnosis of abnormal types such as voltage loss, current loss, phase failure, voltage unbalance, current unbalance, power failure, abnormal power factor, wrong wiring and the like.
And step 302, taking the natural day as a time span, dividing the electrical parameter data in the case data into samples, and calibrating the samples according to the types corresponding to the cases, wherein 0 represents normal, and 1-8 represent abnormal types such as voltage loss, current loss, phase loss, voltage unbalance, current unbalance, power failure, power factor abnormity, wrong wiring and the like.
And step 303, extracting the night feature vector and the daytime feature vector of each sample according to the steps 1 to 2.
And step 304, establishing a night characteristic vector training data set and a day characteristic vector training data set by using the night characteristic vectors and the day characteristic vectors of all case data samples.
And 305, training a nighttime XGboost model by using the nighttime feature vector training data set independently until the nighttime XGboost model converges.
And step 306, training the data set by using the daytime feature vector, and independently training the daytime XGboost model until the daytime XGboost model converges.
And 4, inputting the characteristic vector of the substation metering device to be subjected to state evaluation and abnormality diagnosis into the trained state evaluation and abnormality diagnosis model in the corresponding time period, and outputting a diagnosis result by the state evaluation and abnormality diagnosis model.
Step 4 comprises the following steps:
after the feature vectors are input into a trained night XGboost model or a daytime XGboost model in a corresponding time period, the night XGboost model or the daytime XGboost model respectively outputs probability vectors representing the probabilities of predicted types, and the probability vectors respectively output by the night XGboost model or the daytime XGboost model are added in an element mode to obtain probabilities and vectors.
And when the probability and the probability of the normal type in the vector are the maximum, performing state evaluation on the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis, and otherwise, performing abnormal diagnosis on the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis.
Further, the process of performing state evaluation on the substation metering device to be subjected to state evaluation and abnormality diagnosis includes:
when the probabilities of normal types in the probability vectors respectively output by the night XGboost model or the daytime XGboost model are the maximum, the state of the transformer substation metering device to be subjected to state evaluation and abnormality diagnosis is evaluated to be healthy, and otherwise, the state is evaluated to be sub-healthy.
The process of carrying out abnormity diagnosis on the transformer substation metering device to be subjected to state evaluation and abnormity diagnosis comprises the following steps:
when the abnormal type with the maximum positive probability value in the probability vectors respectively output by the nighttime XGboost model or the daytime XGboost model is the same, the abnormal type of the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis is diagnosed as the abnormal type with the maximum probability, otherwise, the abnormal type is diagnosed as the two abnormal types with the maximum probability.
Specifically, as shown in fig. 4, which is a flow chart of state evaluation and abnormality diagnosis provided by the embodiment of the present invention, as can be seen from fig. 4, step 4 may include: step 401, obtaining data of the smart meter to be tested, including phase splitting voltage, phase splitting current, phase splitting active power and total active power, phase splitting power factor and total power factor of 96 sampling values.
And step 404, adding the night probability vector and the day probability vector output by the night XGboost model and the day XGboost model according to an element mode to obtain a probability sum vector. The night probability vector, the day probability vector, and the probability sum vector are all 9-dimensional vectors, and represent the above 9 types.
In step 405, if the probability value of the normal type is the maximum in the probability sum vector, step 406 is executed, otherwise, step 409 is executed.
In step 406, if the probability values of the normal types in the night probability vector and the day probability vector are both maximum values, step 407 is executed, otherwise, step 408 is executed.
And step 407, evaluating the state of the metering device of the transformer substation to be healthy, outputting an evaluation result, and ending the process.
And step 408, evaluating the state of the metering device of the transformer substation to be sub-health, outputting an evaluation result and ending the process.
In step 409, if the anomaly type with the maximum probability value is the same in the night probability vector and the day probability vector, step 410 is executed, otherwise, step 411 is executed.
And step 410, diagnosing the abnormal type of the metering device of the transformer substation into the abnormal type with the highest probability and the highest abnormal type, outputting a diagnosis result and ending the process.
And 411, diagnosing the abnormal type of the metering device of the transformer substation into two abnormal types with the highest probability and the highest probability, outputting a diagnosis result, and ending the process.
Specifically, the method for evaluating the state and diagnosing the abnormality of the substation metering device based on the construction method of the feature vector given in fig. 2 includes:
(1) a training phase for obtaining case data, comprising: the case data comprises 9 types of normal, voltage loss, current loss, phase loss, voltage unbalance, current unbalance, power failure, power factor abnormity, wrong wiring and the like.
(2) And dividing the case data into samples taking natural days as time units, and completing type calibration of the samples, wherein 0 represents normal, and 1-8 represent abnormal types such as voltage loss, current loss, phase failure, voltage unbalance, current unbalance, power failure, power factor abnormity, wrong wiring and the like.
(3) Dividing the divided data of the intelligent electric meter into two parts according to {0 point-7 point, 20 point-23 point } and {8 point-19 point } of a natural day, and respectively extracting a night feature vector FN and a daytime feature vector FD; and processing all samples and combining the type numbers of the samples to further establish a nighttime feature vector training data set FNDS and a daytime feature vector training data set FDDS.
(4) The nighttime XGBoost model and the daytime XGBoost model are trained with the training data sets FNDS and FDDS, respectively, until they each converge separately.
(5) And (5) in the testing stage, acquiring the data of the intelligent electric meter on the day to be tested, and extracting the characteristic vectors FN and FD at night and in the daytime respectively according to the step (3).
(6) And (4) respectively inputting the night feature vector FN and the daytime feature vector FD extracted in the step (5) into a night XGboost model and a daytime XGboost model to complete prediction, and collecting probability vectors output by the models.
(7) And performing element addition on the night probability vector and the daytime probability vector to obtain a probability sum vector.
(8) If the probability sum of the normal types in the probability sum vectors is maximum, further judging whether the probability corresponding to the normal types in the two probability vectors at night and in the daytime is the maximum value in the respective probability vectors; if so, outputting the running state of the metering device as 'healthy', otherwise, outputting the running state of the metering device as 'sub-healthy', and finishing the state evaluation of the metering device; otherwise, step (9) is performed.
(9) If the maximum value of the probability and the vector is the abnormal type, further judging whether the abnormal type in the two probability vectors at night and in the daytime is the maximum value, if so, outputting the corresponding abnormal type number, and finishing the diagnosis of the single abnormal type; otherwise, the probability and the numbers of the first two abnormal types with the maximum are output, and the diagnosis of the multiple abnormal types is completed.
Example 2
Embodiment 2 provided by the invention is a specific application embodiment of the substation metering device state evaluation and abnormality diagnosis method provided by the invention, and in the application embodiment, 543 gateway substation metering devices provided by a certain power-saving company are used as cases to be trained and verified. The data acquisition frequency of each user is 96 acquisition points per day, and each data acquisition point comprises electrical parameter data such as split-phase voltage, split-phase current, split-phase active power, total active power, split-phase power factor, total power factor and the like. The gateway transformer substation covers four voltage levels of 220kV, 110kV, 35kV, 10kV and the like, and comprises two wiring modes of three-phase three-wire and three-phase four-wire. The time span of the data is 31 days, and except for normal cases, the data also comprises 9 types of voltage loss, current loss, phase failure, voltage unbalance, current unbalance, power factor abnormity, wrong wiring and the like. The power factor fault diagnosis device comprises 368 normal cases, 16 voltage loss cases, 19 current loss cases, 15 phase failure cases, 10 power failure cases, 23 voltage unbalance cases, 34 current unbalance cases, 30 power factor abnormal cases and 28 miswiring cases. In the application case, each exception case only contains one exception type.
According to the data sample structure provided by the invention, the electrical parameter data made by all metering devices are divided. Considering that part of abnormal cases are not all abnormal and belong to a mixture of normal and abnormal cases, normal samples in the abnormal cases are removed. And finally, randomly dividing the training data set and the testing data set by each type of sample according to a 7:3 mode.
The nighttime XGboost model and the daytime XGboost model adopt the same parameters, wherein the number of trees is 1000, the maximum depth of the trees is 7, and the learning rate is 0.01.
The transformer substation metering device state evaluation and abnormity diagnosis method provided by the invention is verified, and the specific test results are shown in the following table:
from the results in the table, the substation metering device state evaluation and abnormality diagnosis method provided by the invention has better state evaluation and abnormality diagnosis performance.
The present invention is not limited to the above preferred embodiments, and any modifications, equivalent replacements, improvements, etc. within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (7)
1. A transformer substation metering device state evaluation and abnormity diagnosis method is characterized by comprising the following steps:
step 1, acquiring electrical parameter data, and calculating an electrical parameter deviation degree or an unbalance degree index at each moment based on the electrical parameter data; acquiring the statistical characteristics of the electrical parameter deviation degree or unbalance degree indexes in different time periods;
step 2, acquiring current symbol characteristics and data integrity characteristics of the electrical parameter data at different time intervals, and establishing characteristic vectors at different time intervals based on the statistical characteristics, the current symbol characteristics and the data integrity characteristics;
the current sign characteristics represent different proportions of split-phase current signs;
step 3, respectively training a state evaluation and abnormity diagnosis model of each time period by using the characteristic vector in the training data set of each time period and the state of the corresponding substation metering device;
step 4, inputting the characteristic vector of the substation metering device to be subjected to state evaluation and abnormity diagnosis into the trained state evaluation and abnormity diagnosis model in the corresponding time period, and outputting a diagnosis result by the state evaluation and abnormity diagnosis model;
the state evaluation and abnormality diagnosis model comprises a night XGboost model and a daytime XGboost model;
the step 4 comprises the following steps:
after the feature vectors are input into the trained night XGboost model or the daytime XGboost model in the corresponding time period, the night XGboost model or the daytime XGboost model respectively outputs probability vectors representing the probabilities of the predicted types, and the probability vectors respectively output by the night XGboost model or the daytime XGboost model are added according to the element mode to obtain probabilities and vectors;
when the probability of the normal type in the probability and the vector is the maximum, performing state evaluation on the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis, otherwise, performing abnormal diagnosis on the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis;
the process of carrying out state evaluation on the transformer substation metering device to be subjected to state evaluation and abnormity diagnosis comprises the following steps:
when the probabilities of normal types in the probability vectors respectively output by the night XGboost model or the daytime XGboost model are the maximum, the state of the transformer substation metering device to be subjected to state evaluation and abnormality diagnosis is evaluated to be healthy, otherwise, the state of the transformer substation metering device to be subjected to state evaluation and abnormality diagnosis is evaluated to be sub-healthy;
the process of carrying out abnormity diagnosis on the transformer substation metering device to be subjected to state evaluation and abnormity diagnosis comprises the following steps:
and when the abnormal type with the maximum positive probability value in the probability vectors respectively output by the night XGboost model or the daytime XGboost model is the same, the abnormal type of the transformer substation metering device to be subjected to state evaluation and abnormal diagnosis is diagnosed as the abnormal type with the maximum probability and the maximum abnormal type, otherwise, the abnormal type is diagnosed as the two abnormal types with the maximum probability and the maximum abnormal type.
2. The method of claim 1, wherein the electrical parameter data in step 1 comprises: split-phase voltage, split-phase current, split-phase active power and total active power, split-phase power factor and total power factor.
3. The method of claim 1, wherein the electrical parameter deviation or imbalance indicator of step 1 comprises: voltage deviation, voltage unbalance, current deviation, current unbalance, active power difference, active power unbalance and phase angle unbalance.
4. Method according to claim 3, characterized in that the voltage deviation at time iVD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C three-phase voltages at the time i, and min () representing a minimum function;
the voltage unbalance degree of the time iVB i The calculation formula of (2) is as follows:
max () represents a get max function;
the current deviation degree at time iCD i The calculation formula of (2) is as follows:
wherein,a, B, C represent the currents of three phases at time i, respectively, and avg () represents the averaging function;
the current imbalance at time iCB i The calculation formula of (2) is as follows:
the power difference at time iPD i The calculation formula of (2) is as follows:
wherein,respectively representing A, B, C the active power of the three phases at time i,representing the total active power at time i, abs () represents an absolute value function,represents a summation;
the active imbalance at time iPB i The calculation formula of (2) is as follows:
the degree of phase angle imbalance at time iFB i The calculation formula of (2) is as follows:
5. The method according to claim 1, wherein the statistical features in step 1 comprise a minimum value, a lower quantile, a middle quantile, an upper quantile, a maximum value and a mean value;
the time periods of the sub-periods include daytime periods and nighttime periods.
6. The method according to claim 1, wherein the calculation formula of the current sign characteristic in the step 2 is as follows:
7. the method of claim 1, wherein the status of the substation metering device in step 3 comprises: normal, voltage loss, current loss, phase loss, voltage unbalance, current unbalance, power failure, abnormal power factor and wrong wiring;
the classification algorithm of the state evaluation and abnormality diagnosis model adopts an XGboost algorithm.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011168103.0A CN111999695B (en) | 2020-10-28 | 2020-10-28 | State evaluation and abnormity diagnosis method for metering device of transformer substation |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202011168103.0A CN111999695B (en) | 2020-10-28 | 2020-10-28 | State evaluation and abnormity diagnosis method for metering device of transformer substation |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111999695A CN111999695A (en) | 2020-11-27 |
CN111999695B true CN111999695B (en) | 2021-01-12 |
Family
ID=73475181
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202011168103.0A Active CN111999695B (en) | 2020-10-28 | 2020-10-28 | State evaluation and abnormity diagnosis method for metering device of transformer substation |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111999695B (en) |
Families Citing this family (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112748390B (en) * | 2020-12-23 | 2022-02-15 | 南方电网电力科技股份有限公司 | Method and device for evaluating state of electric energy meter |
CN113204894B (en) * | 2021-05-28 | 2024-06-14 | 华中科技大学 | Construction method and application of electric energy metering abnormality diagnosis model |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105557070A (en) * | 2013-09-20 | 2016-05-04 | 西门子公司 | Method for setting an LED signal transmitter in terms of day and night mode, and LED signal transmitter |
CN110245846A (en) * | 2019-05-28 | 2019-09-17 | 国网浙江省电力有限公司绍兴供电公司 | Current break reason method of discrimination based on XGBoost algorithm |
CN110717537A (en) * | 2019-10-08 | 2020-01-21 | 支付宝(杭州)信息技术有限公司 | Method and device for training user classification model and executing user classification prediction |
CN111551888A (en) * | 2020-04-23 | 2020-08-18 | 宁夏隆基宁光仪表股份有限公司 | Electric energy meter metering data fault analysis method based on improved AdaBoost algorithm |
US10753969B2 (en) * | 2018-04-13 | 2020-08-25 | General Electric Company | Methods and systems for predicting failures in X-ray tubes |
-
2020
- 2020-10-28 CN CN202011168103.0A patent/CN111999695B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105557070A (en) * | 2013-09-20 | 2016-05-04 | 西门子公司 | Method for setting an LED signal transmitter in terms of day and night mode, and LED signal transmitter |
US10753969B2 (en) * | 2018-04-13 | 2020-08-25 | General Electric Company | Methods and systems for predicting failures in X-ray tubes |
CN110245846A (en) * | 2019-05-28 | 2019-09-17 | 国网浙江省电力有限公司绍兴供电公司 | Current break reason method of discrimination based on XGBoost algorithm |
CN110717537A (en) * | 2019-10-08 | 2020-01-21 | 支付宝(杭州)信息技术有限公司 | Method and device for training user classification model and executing user classification prediction |
CN111551888A (en) * | 2020-04-23 | 2020-08-18 | 宁夏隆基宁光仪表股份有限公司 | Electric energy meter metering data fault analysis method based on improved AdaBoost algorithm |
Also Published As
Publication number | Publication date |
---|---|
CN111999695A (en) | 2020-11-27 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102170124B (en) | Early warning method of stable-state index of power quality | |
CN106405475A (en) | Electric energy meter abnormity diagnosis method | |
CN107886171B (en) | PMU data-based breaker state online diagnosis method and system | |
CN108683180B (en) | Three-phase low-voltage power distribution network topology reconstruction method | |
CN111999695B (en) | State evaluation and abnormity diagnosis method for metering device of transformer substation | |
CN110703009B (en) | Abnormal analysis and processing method for line loss rate of transformer area | |
CN113032454A (en) | Interactive user power consumption abnormity monitoring and early warning management cloud platform based on cloud computing | |
CN107543989B (en) | Method for judging line loss abnormity based on voltage loss and phase loss of electric energy meter | |
CN106786531B (en) | Power distribution terminal optimal configuration method | |
CN110928866A (en) | Line topology checking method and system based on power failure information big data analysis | |
CN105652118A (en) | Load instantaneous energy feature-based power grid electric energy load monitoring method | |
CN112288303A (en) | Method and device for determining line loss rate | |
CN111398859B (en) | User low-voltage cause big data analysis method and system | |
CN115270974A (en) | Intelligent electricity stealing detection system based on big data analysis | |
CN116823226A (en) | Electric power district fault monitoring system based on big data | |
CN108596450B (en) | Power grid risk early warning method and system | |
Liu et al. | Identification of major power quality disturbance sources in regional grid based on monitoring data correlation analysis | |
CN111638401A (en) | Capacitor online monitoring system and method | |
CN114675212B (en) | Method for studying and judging abnormal connection of neutral point of distribution transformer | |
CN116502149A (en) | Low-voltage power distribution network user-transformation relation identification method and system based on current characteristic conduction | |
CN114156865B (en) | Low-voltage distribution network topology generation and fault prediction method considering state perception | |
CN115733258A (en) | Control method of all-indoor intelligent substation system based on Internet of things technology | |
CN113552527A (en) | Multi-meter set line loss meter-setting device and line loss meter-setting method | |
Huang et al. | Extracting substation cyber-physical architecture through intelligent electronic devices' data | |
CN111986469A (en) | Intelligent diagnosis method for field terminal fault |
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 | ||
CP01 | Change in the name or title of a patent holder | ||
CP01 | Change in the name or title of a patent holder |
Address after: Room 1803-1805, building 2-07, guanggu.core center, 303 Guanggu Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000 Patentee after: Wuhan Gelanruo Intelligent Technology Co.,Ltd. Address before: Room 1803-1805, building 2-07, guanggu.core center, 303 Guanggu Avenue, Donghu New Technology Development Zone, Wuhan City, Hubei Province, 430000 Patentee before: WUHAN GLORY ROAD INTELLIGENT TECHNOLOGY Co.,Ltd. |