CN115000973A - Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering - Google Patents
Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering Download PDFInfo
- Publication number
- CN115000973A CN115000973A CN202210605202.3A CN202210605202A CN115000973A CN 115000973 A CN115000973 A CN 115000973A CN 202210605202 A CN202210605202 A CN 202210605202A CN 115000973 A CN115000973 A CN 115000973A
- Authority
- CN
- China
- Prior art keywords
- power
- reactive
- phase
- data
- reactive power
- 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
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J3/00—Circuit arrangements for ac mains or ac distribution networks
- H02J3/18—Arrangements for adjusting, eliminating or compensating reactive power in networks
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00002—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by monitoring
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02J—CIRCUIT ARRANGEMENTS OR SYSTEMS FOR SUPPLYING OR DISTRIBUTING ELECTRIC POWER; SYSTEMS FOR STORING ELECTRIC ENERGY
- H02J13/00—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network
- H02J13/00006—Circuit arrangements for providing remote indication of network conditions, e.g. an instantaneous record of the open or closed condition of each circuitbreaker in the network; Circuit arrangements for providing remote control of switching means in a power distribution network, e.g. switching in and out of current consumers by using a pulse code signal carried by the network characterised by information or instructions transport means between the monitoring, controlling or managing units and monitored, controlled or operated power network element or electrical equipment
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
- Y02E40/00—Technologies for an efficient electrical power generation, transmission or distribution
- Y02E40/30—Reactive power compensation
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
A reactive compensation characteristic quantity sample library is established on the basis of daily metering data of months with month average power factors not smaller than a set threshold value, which are selected from historical electricity consumption data of power consumers, the reactive compensation characteristic quantity sample library and data corresponding to days to be detected are clustered together, and if the data corresponding to the days to be detected form an individual cluster, the low-voltage reactive compensation device of the user on the detection day is judged to be abnormal. Therefore, the established detection method for clustering the reactive compensation characteristic quantity sample library and the data corresponding to the days to be detected together can find out abnormality in time, and the power regulation fee penalty caused by the fact that the low-voltage reactive compensation device of the user fails to normally compensate reactive power is reduced. The method is suitable for detecting the fault of the reactive power compensation device for the power users equipped with the low-voltage reactive power compensation device.
Description
Technical Field
The invention belongs to the field of power device abnormity detection, and relates to a method for detecting whether a low-voltage reactive power compensation device of a power user on a day to be detected has faults or not by adopting an unsupervised learning clustering method based on reactive power compensation characteristic quantities.
Background
The method for improving the electricity utilization efficiency is the most economical way for reducing carbon emission, and the most common method for improving the electricity utilization efficiency of users at present is to improve the power factor of an electricity load by adopting a reactive compensation device. In order to improve the electricity utilization efficiency of the users, a power supply company collects punitive force-regulating electricity fee fines according to the percentages of the electricity quantity and the electricity fee and the required amount and the electricity fee for the users who do not meet the power factor assessment requirement, and gives force-regulating electricity fee rewards to the users who exceed the power factor requirement. The reactive power absorbed by the user from the power grid is reduced, the electricity expense of the user can be reduced, the running loss of the power grid can be reduced, and the method is a low-carbon emission reduction means capable of realizing multi-win.
The power utilization behavior of power consumers is large in discreteness, and traditionally, power systems mainly require the users to configure reactive compensation according to 20% -40% of the installation capacity to guarantee power utilization efficiency. Because the reactive power compensation device at the user side is wide in points, power supply enterprises are difficult to effectively monitor whether the reactive power compensation device has faults or not, the users are mainly depended on to find the faults of the reactive power compensation device according to the device alarm or to be reminded to pay attention to related problems in a monthly power charging fine mode. Power supply enterprises mainly pay attention to optimal configuration and operation and maintenance of reactive power compensation in transformer substations. Under the condition that the warning information of the reactive compensation device cannot be observed remotely, an electric power technician mainly judges whether the reactive compensation device breaks down or not on the basis of the reactive compensation effect represented by the user power factor index. Although the abnormal fault of the low-voltage reactive power compensation device can not only cause the electricity cost loss of a user, but also increase the running loss of a power grid, a remote fault diagnosis method aiming at the low-voltage reactive power compensation device of a power user is fresh at present.
The low-voltage reactive power compensation device faults are related to the configuration and the reliability of the capacitor and the control protection device thereof:
the capacitor is affected by factors such as ambient temperature, harmonic waves, overvoltage, and the like, and the capacitor bulges and leaks oil after aging, and the capacitance value is reduced or opened, so that the compensation effect is affected.
Common capacitor-switching switches are contactors, thyristor switches and compound switches. Repeated switching times of the contactor are limited, switching reliability and sensitivity are reduced inevitably along with the increase of the switching times, and contacts of the contactor are melted due to long-term switching inrush current. The thyristor type switching switch is applied to a fast switching scene, and not only can certain power loss be caused by continuous operation, but also the aging and damage can be accelerated due to heating.
In addition, factors such as wiring errors and parameter setting errors may also cause the capacitor not to be able to switch the compensation reactive power according to the requirements of the power factor controller.
When the reactive power compensation device has a fault and is abnormal, the low-voltage reactive power compensation device can not be put into or withdrawn under severe conditions, so that a large amount of reactive power is absorbed from a power grid or is reversely delivered, and power regulation fee fine and power grid line loss are increased.
Disclosure of Invention
The invention aims to provide a remote fault detection method for a low-voltage reactive power compensation device of a power consumer based on reactive compensation characteristic quantity clustering, aiming at the problem that a considerable number of reactive power compensation devices on the user side operate with diseases for a long time due to the lack of a remote fault detection method for the low-voltage reactive power compensation device on the power consumer side at present.
In order to achieve the purpose, the invention adopts the technical scheme that: a remote fault detection method for a low-voltage reactive power compensation device of a power consumer based on reactive power compensation characteristic quantity clustering comprises the following steps:
step 1: acquiring historical electricity consumption data of an electricity consumer, selecting daily metering data of months with the monthly average power factor being larger than or equal to a set threshold value as sample database data, wherein the number of samples of the sample database data is larger than or equal to 180 days;
because the power users comprise large industrial users, common industrial and non-industrial users, commercial users, agricultural users and the like, and users who do not reach the power factor power grid assessment standard of the type of the power user need to carry out power adjustment and fine payment, the invention sets the set threshold value as: the power factor power grid evaluation standard of the type of the power user is less than or equal to a set threshold and less than 1, and monthly daily metering data with a monthly average power factor greater than or equal to the set threshold is used as a sample library, so that the data of the sample library can be ensured to correspond to the metering data of the reactive power compensation device in a normal operation state as far as possible.
Meanwhile, the method takes 180 days as the lower limit of the sample number of the sample database data, and can traverse all the electricity consumption behavior rules of the user. Because the distribution conditions of the three-phase active power of the user in one day are similar under the power consumption behavior rule of the same user, the three-phase active power and the three-phase reactive power of the user are in an approximate positive correlation relationship, and the reactive power absorbed from the power grid after compensation has approximate distribution characteristics under the same power consumption behavior.
Step 2: acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling every same time period every day from the sample database data of the user; and taking three-phase active power, three-phase reactive power and three-phase power factors as characteristic quantities, selecting any two combinations as reactive compensation characteristic quantities, and establishing a reactive compensation characteristic quantity sample library based on the obtained three characteristic quantity data. For example, with a combination of three-phase reactive power and three-phase power factor as the reactive compensation characteristic quantity, the three-phase reactive power time series Y i ＝{Q A1 ，…，Q AM ，…，Q C1 ，…，Q CM }, three-phase power factor time series Z i ＝{COSθ A1 ，…，COSθ AM ，…COSθ C1 ，…，COSθ CM And combining two characteristic quantities of three-phase reactive power and three-phase power factor together to serve as a reactive compensation characteristic quantity X i ＝{Q A1 ，…，Q AM ，…，Q C1 ，…，Q CM ，COSθ A1 ，…，COSθ AM ，…COSθ C1 ，…，COSθ CM Establishing a reactive compensation characteristic quantity sample library; wherein i is a data sample serial number in days in the established sample library, and i is 1, 2, …, n; m is the total number of sampling points divided by the same time period of each phase every day.
During sampling, recording three-phase active power, three-phase reactive power or three-phase power factor at the same time interval every day, wherein the time interval or time period can be 1-60 min; a typical time interval is 15min, where three-phase active power, three-phase reactive power or three-phase power factor data of 96 collection points are contained each day, and M is 96. The three-phase active power, three-phase reactive power and three-phase power factor data of the time intervals of one day can be obtained from a power utilization information acquisition system or a marketing system.
The method selects the three-phase active power, the three-phase reactive power and the three-phase power factor as characteristic quantities, can identify the abnormity of the reactive power compensation device better compared with one or a combination mode of a plurality of three-phase power, three-phase power consumption, three-phase power factor, three-phase current or three-phase harmonic content, and has obvious rationality. Because: the three-phase harmonic is mainly related to voltage deviation, the fault of the reactive power compensation device is not a main reason causing the three-phase harmonic, and the three-phase harmonic is not suitable to be used as a characteristic quantity for judging the fault operation of the reactive power compensation device; the three-phase current is mainly determined by load and voltage level, and the fault of the reactive compensation device cannot cause the violent fluctuation of the three-phase current, so the three-phase current is not suitable to be used as the reactive compensation characteristic quantity; and no coupling relation exists between the three-phase power consumption and the three-phase power factor, and the combination of the three-phase power consumption and the three-phase power factor cannot reflect the running state of the reactive power compensation device.
Moreover, when the reactive power absorbed from the network increases, the value of the reactive power Q increases and the value of the voltage U decreases, according to the formula
It is known that the current I value increases. By the formula
It is known that the Q value is increased, resulting inThe power factor cos θ decreases. Normally, P > Q, the value of I is mainly determined by active power P, and fluctuation of the active power has small influence on change of the value of I. The power factor cos θ and the change in the reactive power Q value are relatively more significantly affected by the reactive compensation fault. Because any one of the power factor cos theta and the reactive power Q is independently selected as the reactive compensation characteristic quantity, the normal state and the abnormal state of the low-voltage reactive compensation device cannot be accurately distinguished, namely when one characteristic quantity is independently selected, the condition that misjudgment or misjudgment occurs because the data is clustered into the same class as the data during normal switching of the equipment exists. When the switching is normal, the power factor is higher; in abnormal operation, the power factor is low. When the reactive power corresponding to normal switching and the reactive power corresponding to abnormal switching are clustered into the same class, the data of normal switching and abnormal operation can be distinguished by adding a power factor characteristic. Therefore, a combination of the three-phase power factor cos θ and the three-phase reactive power Q can be used as the reactive compensation characteristic quantity. In addition, the active power P and the reactive power Q are in a substantially positive correlation, and different active power levels correspond to different reactive power levels. In the event of a fault, the distribution relationship between the active power and the reactive power under normal conditions is destroyed, i.e. the combination of the three-phase active power P and the three-phase reactive power Q can also identify the fault. In addition, since fault detection is carried out according to the power factor cos theta alone, the abnormity can be judged to be normal when the load is heavy, the combination of the three-phase active power P and the three-phase power factor cos theta can avoid the situation that misjudgment or missed judgment is normal simply caused by low power factor, and the detection accuracy is higher than that of judgment carried out through the three-phase power factor cos theta alone. Namely, the combination of any two of three-phase active power P, three-phase power factor cos theta and three-phase reactive power Q is selected as the reactive compensation characteristic quantity, and the abnormal state of reactive compensation can be effectively reflected.
And step 3: acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling is carried out every time period the same as that in the step 2 on the day to be detected (namely the sampling time interval of the day to be detected is the same as that in the step 2), clustering the reactive compensation characteristic quantity sample library data established in the step 2 together with the acquired data of the day to be detected, wherein the corresponding characteristic quantity in the reactive compensation characteristic quantity sample library in the step 2 is the same as the corresponding characteristic quantity of the data of the day to be detected in clustering (namely if the sample library in the step 2 selects the three-phase reactive power and the three-phase power factor as reactive compensation characteristic quantities, the data selected on the day to be detected is also the three-phase reactive power and the three-phase power factor data), and if the data of the day to be detected is clustered independently, judging that the low-voltage reactive power compensation device on the detection day of the user has a fault abnormality, and prompting to carry out field inspection, otherwise, the low-voltage reactive power compensation device of the user on the detection day is normal, and whether the low-voltage reactive power compensation device of the user on the next detection day is abnormal or not is detected.
In the step 3, since the historical electricity consumption behavior rule of the user is not determined, the method performs cluster analysis by using a method without presetting cluster number. The input variable of the method is one-dimensional multivariable, a distance-based neighbor propagation algorithm clustering algorithm is selected, and the closer the distance between two objects is, the greater the similarity is. And clustering algorithms such as DBSCAN, mean-shift and the like based on density clustering can also be selected.
According to the method, the data corresponding to the day to be detected and the established reactive compensation characteristic quantity sample base are subjected to clustering analysis by means of a clustering algorithm, and whether the low-voltage reactive compensation device is abnormal or not is judged by judging whether the data of the day to be detected are clustered or not to form a new cluster. If the data corresponding to the day to be detected independently form a new cluster, judging that the low-voltage reactive power compensation device is abnormal; and if no single cluster is formed, considering that the low-voltage reactive power compensation device of the user on the detection day normally works. The method can find out whether the low-voltage reactive power compensation device is abnormal or not in time on the detection day, reduces the power adjustment fee penalty caused by the fact that the low-voltage reactive power compensation device of a user cannot normally compensate reactive power due to faults, and is suitable for carrying out reactive power compensation device fault detection on power users (including residential public transformer users, large industrial users, common industrial and non-industrial users, commercial users, agricultural users and the like) provided with the low-voltage reactive power compensation device.
Drawings
Fig. 1 is a distribution diagram of three-phase reactive power and three-phase power factor when the reactive power compensation device is in normal operation.
Fig. 2 is a distribution diagram of three-phase reactive power and three-phase power factor when a partial loop of the reactive power compensation device is not in fault.
Fig. 3 is a diagram of three-phase reactive power and three-phase power factor distribution of a sample in a sample library.
Fig. 4 is a distribution diagram of three-phase reactive power and three-phase power factor of a fault sample in which a certain drain is determined to be normal.
Detailed Description
The smart electric meter at the power consumer side is popularized and applied, and can report high-density metering data by adopting broadband carrier communication. The power supply enterprise can carry out mining analysis according to historical power utilization data reported by power consumers, so that not only can the power utilization behavior habits of the consumers be analyzed, but also the fault abnormity of the reactive power compensation device at the consumer side can be mined and analyzed. The method is a method for constructing reactive compensation characteristic quantities based on historical electricity utilization data and remotely detecting whether the low-voltage reactive compensation device has fault abnormality or not through clustering.
The invention relates to a remote fault detection method for a low-voltage reactive power compensation device of a power consumer based on reactive power compensation characteristic quantity clustering, which comprises the following steps:
step 1: acquiring historical electricity consumption data of an electricity consumer, selecting daily metering data of months with the monthly average power factor being larger than or equal to a set threshold value as sample database data, wherein the number of samples of the sample database data is larger than or equal to 180 days; the power factor power grid assessment standard of the user type to which the power user belongs is less than or equal to a set threshold value < 1.
Step 2: acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling every same time period every day from the sample database data of the user; and taking the three-phase active power, the three-phase reactive power and the three-phase power factor as characteristic quantities, selecting any two combinations as reactive compensation characteristic quantities, and establishing a reactive compensation characteristic quantity sample library based on the three acquired characteristic quantity data.
During sampling, the three-phase active power, the three-phase reactive power or the three-phase power factor are recorded at the same time interval every day, and the time interval or the time period can be between 1min and 60min, such as 10min and 15 min.
And 3, step 3: acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling is carried out every time period the same as that in the step 2 on the day to be detected, clustering the reactive compensation characteristic quantity sample library data established in the step 2 together with the acquired data on the day to be detected, wherein the corresponding characteristic quantity in the sample library in the step 2 is the same as the corresponding characteristic quantity of the data on the day to be detected during clustering, judging that the low-voltage reactive compensation device of the user has fault abnormality if the data corresponding to the day to be detected is clustered into a cluster, prompting to carry out field inspection, otherwise judging that the detection day reactive compensation device of the user normally operates, and carrying out abnormal detection on the low-voltage reactive compensation device of the next detection day of the user.
Since the number of the clustered clusters is not determined in the step, the method selects a method without presetting the number of the clusters to perform clustering analysis. The clustering algorithm is conventional in the art.
Example 1
And (3) carrying out low-voltage reactive power compensation device abnormity detection on metering data of a certain user in the knitting or crocheting product manufacturing industry with monthly power adjustment fee fines. And screening daily metering data and establishing a sample library from the historical electricity consumption data of the user according to the monthly average power factor which is more than or equal to a set threshold, wherein the number of the sample libraries is more than 180 days.
With three-phase reactive power Y of each day in the sample library i ＝{Q A1 ，…，Q A96 ，…，Q C1 ，…，Q C96 } and the three-phase power factor Z i ＝{COSθ A1 ，…，COSθ A96 ，…COSθ C1 ，…，COSθ C96 Is combined into variable X i ＝{Q A1 ，…，Q A96 ，…，Q C1 ，…，Q C96 ，COSθ A1 ，…，COSθ A96 ，…COSθ C1 ，…，COSθ C96 And forming a reactive compensation characteristic quantity sample library. And selecting the three-phase reactive power-free and three-phase power factor data and the three-phase reactive power-free and power factor-free data of the real fault day acquired on siteThe three-phase reactive power-free and three-phase power factor data metering data under the normal operation condition of the power compensation device form a sample to be detected, the metering data of 10 days of normal operation in the samples (30) to be detected is a normal sample, and the metering data of 20 days of fault operation of the reactive power compensation device is a fault sample. And clustering the data of the samples to be detected one by one with the data of the reactive compensation characteristic quantity sample library, wherein the result is shown in table 1. Data corresponding to 20 fault samples in the samples to be detected are all clustered into a cluster independently, namely the low-voltage reactive power compensation device is abnormal in fault, and 10 normal samples are not clustered independently, namely the reactive power compensation device operates normally. The detection result of the method completely accords with the field acquisition result. The detection result shows that the method can accurately detect the fault date of the reactive power compensation device of the user without false alarm.
TABLE 1 test results
| Normal sample (10 days) | Failure sample (20 days) | |
| Detected as normal | 10 | 0 |
| Is detected as abnormal | 0 | 20 |
Further analysis is performed by combining fig. 1 and fig. 2, and fig. 1 and fig. 2 respectively show the distribution of three-phase reactive power and three-phase power factor when the reactive power compensation device is in normal operation under a certain power consumption behavior, and the distribution of three-phase reactive power and three-phase power factor when a partial loop of the reactive power compensation device is not in fault. The black box lines in fig. 1 are the envelope ranges of the three-phase reactive power and the three-phase power factor after 8 points. As is apparent from fig. 2, when the user reactive power compensation device is abnormal, both the three-phase reactive power and the three-phase power factor deviate from the distribution when the reactive power compensation is normal after 8 points, and respectively approach the upper and lower bounds of the envelope line.
Part of the loop is abnormal and can not be used in the fault sample, and the controller is used for the rest loop capacitor banks in order to achieve the three-phase power factor target. If the normal loop capacitor bank can not meet the control target of the three-phase power factor after being completely put into use, the capacity of clamping the reactive power of a user at a low position is lost, so that the reactive power absorbed from a power grid is increased, and the three-phase power factor is reduced.
When the user load is in low-order operation, the residual normal reactive power compensation loop is enough to compensate the reactive power demand of the user, and the difference between the reactive power and the power factor of the reactive power abnormal user before 8 points and the reactive power normal time is not large; under the condition of high load after 8 points, due to the fault of a part of reactive compensation loops, the reactive demand caused by the increase of the load exceeds the compensation capacity of a normal working capacitor bank, so that the three-phase power factor and the three-phase reactive power slightly break through the normal envelope interval.
Under the normal and abnormal operation conditions of the reactive power compensation device, the distribution characteristics of three-phase reactive power and three-phase power factors are different, so that corresponding metering data is not clustered into the same cluster with the data of the sample base when the reactive power compensation device is in fault, but a single cluster is formed.
Comparative example 1
The comparative example is basically the same as the example 1, except that the sample database data of the comparative example is the electricity consumption metering data directly selected from the historical electricity consumption data of the user, wherein the electricity consumption metering data has the same sample number as the example 1, and the monthly average power factor is not screened whether reaching the set threshold value. The 30 samples to be tested used in this comparative example are identical to those in example 1, and the test results obtained by the sample library selection method in this comparative example are shown in table 2.
TABLE 2 results of unscreened test on the sample library data
| Normal sample (10 days) | Failure sample (20 days) | |
| The detection result is normal | 10 | 4 |
| The detection result is abnormal | 0 | 16 |
As can be seen from the above table, the same samples to be detected as in example 1 were used, and the detection accuracy of the failure samples was significantly lower than that of example 1 without performing the screening of the sample library.
The reason for analyzing the reduction of the detection accuracy rate of the fault sample is as follows: because the sample library is not subjected to condition screening, a sample with a monthly average power factor which does not reach the standard possibly enters the sample library, at the moment, an abnormal sample is taken as a normal sample, a detection sample with a fault and the abnormal sample in the sample library are clustered into the same class, and therefore, a single class cluster is not generated on the abnormal detection day, and the abnormal detection day is judged to be normal, so that the judgment is missed.
Further combining fig. 3 and 4, it can be found from the distribution of the three-phase reactive power and the three-phase power factor in the figures that the three-phase reactive power and the three-phase power factor distribution of the samples in the sample library and the fault samples are substantially consistent. Therefore, it can be inferred that the reason why the faulty sample is missed to be judged as normal is that the selection of the sample library has no screening condition, so that the abnormal sample is taken as a normal sample in the sample library, and the faulty sample on the detection day and the abnormal sample introduced into the sample library are clustered into the same cluster, thereby causing the missed judgment. Compared with the detection result of the example 1, the sample library data has higher detection accuracy after conditional screening.
Comparative example 2
The comparative example is basically the same as example 1, except that the comparative example selects a combination of three-phase power consumption and three-phase power factor for clustering, instead of the reactive power compensation characteristic quantity provided by the present invention. The results are shown in table 3, 10 normal samples in 30 samples to be detected do not form an independent cluster, 20 fault samples do not form an independent cluster, and the fault samples are not judged to be abnormal. The analysis reason is that there is no coupling relation between the three-phase power consumption and the three-phase power factor, that is, there is no causal connection between the power consumption and the power factor, so the combination of the two cannot reflect the operation state of the reactive power compensation device, and the abnormality of the low-voltage reactive power compensation device cannot be detected by selecting the combination of the three-phase power consumption and the three-phase power factor. Compared with the detection result of the embodiment 1, the clustering based on the reactive compensation characteristic quantity provided by the invention is more effective in the abnormity detection of the low-voltage reactive compensation device.
TABLE 3 test results Table
| Normal sample (10 days) | Failure sample (20 days) | |
| The detection result is normal | 10 | 20 |
| The detection result is abnormal | 0 | 0 |
Claims (4)
1. A remote fault detection method for a low-voltage reactive power compensation device of a power consumer based on reactive power compensation characteristic quantity clustering is characterized by comprising the following steps:
step 1: acquiring historical electricity consumption data of an electricity consumer, selecting daily metering data of months with the monthly average power factor being larger than or equal to a set threshold value as sample database data, wherein the number of samples of the sample database data is larger than or equal to 180 days;
step 2: acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling every other same time period every day from the sample database data of the user; taking three-phase active power, three-phase reactive power and three-phase power factors as characteristic quantities, selecting any two combinations as reactive compensation characteristic quantities, and establishing a reactive compensation characteristic quantity sample library based on the obtained three characteristic quantity data;
and step 3: and (3) acquiring three-phase active power, three-phase reactive power and three-phase power factor data when sampling is carried out every time period the same as that in the step (2) on the day to be detected, clustering the reactive compensation characteristic quantity sample library data established in the step (2) together with the data on the day to be detected, wherein the corresponding characteristic quantity in the sample library in the step (2) is the same as the corresponding characteristic quantity of the data on the day to be detected during clustering, if the data on the day to be detected is clustered into a cluster independently, judging that the low-voltage reactive compensation device of the detection day of the user is abnormal, otherwise, judging that the low-voltage reactive compensation device is normal.
2. The method for remotely detecting the fault of the power consumer low-voltage reactive power compensation device based on the reactive compensation characteristic quantity clustering as claimed in claim 1, wherein the set threshold is selected between the power factor grid evaluation standard of the type of the user to which the power consumer belongs and 1.
3. The method for remotely detecting the fault of the power consumer low-voltage reactive power compensation device based on the reactive compensation characteristic quantity clustering as claimed in claim 1, wherein the time period in the step 2 is in a range of 1min-60 min.
4. The method for remotely detecting the faults of the low-voltage reactive power compensation devices of the power consumers based on the reactive power compensation characteristic quantity clustering, as claimed in claim 1, wherein the clustering in the step 3 is a clustering algorithm which does not need to preset the cluster number of clusters.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210605202.3A CN115000973A (en) | 2022-05-31 | 2022-05-31 | Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210605202.3A CN115000973A (en) | 2022-05-31 | 2022-05-31 | Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115000973A true CN115000973A (en) | 2022-09-02 |
Family
ID=83030750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210605202.3A Pending CN115000973A (en) | 2022-05-31 | 2022-05-31 | Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115000973A (en) |
-
2022
- 2022-05-31 CN CN202210605202.3A patent/CN115000973A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| de Souza et al. | Detection and identification of energy theft in advanced metering infrastructures | |
| Cavraro et al. | Power distribution network topology detection with time-series signature verification method | |
| US8000910B2 (en) | Automated voltage analysis in an electrical system using contextual data | |
| McBee et al. | Utilizing a smart grid monitoring system to improve voltage quality of customers | |
| Guo et al. | Online data validation for distribution operations against cybertampering | |
| Wischkaemper et al. | Application of waveform analytics for improved situational awareness of electric distribution feeders | |
| CN112034260B (en) | Accurate analysis and anti-electricity-stealing accurate positioning method for low-voltage line loss of distribution transformer area | |
| US11740266B2 (en) | Systems and methods for monitoring energy-related data in an electrical system | |
| US20110254557A1 (en) | Electromechanical relays including embedded sensors | |
| CN114383652A (en) | Method, system and device for identifying potential fault online risk of power distribution network | |
| CN114167156A (en) | System and method for managing voltage event alerts in an electrical system | |
| CN108512222A (en) | A kind of intelligent substation complex automatic system | |
| CN102289582A (en) | Method for recognizing risk state of power distribution network | |
| US20160216308A1 (en) | Diagnosis and position identification for remote capacitor banks | |
| CN115000973A (en) | Remote fault detection method for low-voltage reactive power compensation device of power consumer based on reactive power compensation characteristic quantity clustering | |
| CN117349624A (en) | Electric power energy monitoring method, system, terminal equipment and storage medium | |
| US20220200281A1 (en) | Systems and methods for evaluating electrical phasors to identify, assess, and mitigate power quality issues | |
| CN115267422A (en) | Sensing, diagnosing and positioning method for line faults of complex power distribution network | |
| CN104714205A (en) | Electricity meter misplacement detection system and method thereof | |
| Feng et al. | The method and application of electric energy meter status evaluation fused security region | |
| de Oliveira et al. | Voltage dips caused by faults in a transmission system: a monitoring case study of a sensitive industrial consumer | |
| Nadkarni et al. | Real‐time spatio‐temporal trend and level (T&L) filtering scheme for early detection of voltage instability | |
| US20230186409A1 (en) | Method for evaluating power distributing networks | |
| Tsioumpri et al. | Data analytics to support operational distribution network monitoring | |
| Nulty | Does waveform analysis hold the key to a predictive grid? |
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 |