Household photovoltaic anomaly identification method based on smart meter and geographic information clustering

Through the method of smart meter and geographic information grouping, combined with 3-sigma and box diagram methods, abnormal failures of household photovoltaic systems are identified, and the problems of difficulty and low accuracy in the existing technology are solved, and more efficient abnormalities are achieved Diagnosis and maintenance improve equipment utilization and user benefits.

CN112434822BActive Publication Date: 2024-11-29HUZHOU ELECTRIC POWER SUPPLY CO OF STATE GRID ZHEJIANG ELECTRIC POWER CO LTD +1
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202011096516.2
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-10-14
Publication Date
2024-11-29
Estimated Expiration
2040-10-14

AI Technical Summary

Technical Problem

The prior art is difficult to accurately identify and analyze abnormal failures of household photovoltaic systems, resulting in long diagnosis time and low accuracy, which affects the utilization rate of equipment and user benefits.

Method used

Using a method based on smart electricity meter and geographic information clustering, the average value and standard deviation of daily power generation are calculated, combined with 3-sigma and box graph methods, outliers are identified, and the intersection of abnormal users is finally identified. As a result, misjudgment and misjudgment are reduced.

Benefits of technology

It improves the accuracy of abnormal identification, shortens diagnosis time, reduces resource losses, and improves equipment utilization and user benefits.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112434822B_ABST
    Figure CN112434822B_ABST
Patent Text Reader

Abstract

The present invention discloses a household photovoltaic anomaly identification method based on smart meters and geographic information grouping, and relates to the field of photovoltaic anomaly identification. At present, the identification of abnormal photovoltaics requires large calculations and low accuracy. The present invention first finds out the three normal users closest to the faulty user, calculates the photocurrent value according to the single diode model formula of the photovoltaic module, and then calculates the power loss time series of the faulty user within a single day, and then can fit the power loss by the least squares method, and combine the output characteristics of the photovoltaic array to obtain the fault diagnosis result according to the fitting result. The technical solution makes a comprehensive judgment through two methods, with high accuracy, and reduces the resource loss caused by identification errors. And through the identification of abnormal users, the scope and calculation of abnormal diagnosis can be reduced, and the abnormal diagnosis time can be shortened, which is of great significance for the subsequent maintenance and repair of abnormal users; it improves the income and equipment utilization.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of photovoltaic anomaly identification, and in particular to a household photovoltaic anomaly identification method based on smart meters and geographic information grouping. Background Art

[0002] With the country's vigorous promotion of renewable clean energy and attention to natural ecology, photovoltaic power generation has been widely promoted and applied due to its pollution-free and sustainable characteristics. Around 2013, the policy subsidies for photovoltaic power generation increased, large-scale photovoltaic power generation was gradually connected to the grid, and the number of distributed photovoltaic users increased sharply, and the photovoltaic installed capacity has jumped to the first place in the world. After photovoltaic power generation has been widely deployed across the country and gradually stabilized, due to the particularity of photovoltaic installation locations, how to monitor the operating status of photovoltaic power generation and analyze the faults that occur has become the key to the healthy development of photovoltaic power generation.

[0003] Photovoltaic power generation in China is currently divided into two forms: centralized and distributed. The centralized form mainly appears in the form of photovoltaic power stations, while the distributed form mainly appears in the form of household photovoltaics. Among them, photovoltaic power stations are generally equipped with centralized monitoring systems due to their large capacity and highly concentrated geographical location, which can complete information interaction, intelligent operation and maintenance, and dispatch management in an integrated manner. Household photovoltaics are installed by residents themselves, and the photovoltaic modules are installed in places with sufficient sunshine in their own homes (usually rooftops), and are equipped with inverters to invert and boost the DC output of the photovoltaic modules to synchronize with the power grid. At the same time, smart meters are installed on the output side of the inverter to record the electricity generated by photovoltaics. Residential users can use this electricity for daily household electricity use, or they can sell it to power grid companies to make profits.

[0004] Household photovoltaics are huge in number, and the volume of a single unit is very small compared to photovoltaic power stations, but the total capacity ratio is very high. The installation location of photovoltaic modules is an open and unobstructed outdoor environment. In an unprotected state, there will be many reasons that will cause the failure of photovoltaic modules, resulting in a decrease in photovoltaic power generation efficiency. Due to its distributed characteristics, household positioning and low input-output ratio of operation and maintenance devices, it does not have the operation and maintenance conditions of photovoltaic power stations. The only basis for diagnosis that can be relied on is the readings recorded by the smart meter and the equipment information left by the household photovoltaics when it is installed in the power grid company. Among them, the data recorded by the smart meter are the output voltage, output current, output power and daily power generation of the photovoltaic array after the inverter is boosted. The data left by household photovoltaic users generally include the reported capacity, that is, the rated power of the photovoltaic system under standard testing conditions, the user's geographical location, that is, the user's address and longitude and latitude. At the same time, some users also record the manufacturer and model of photovoltaic modules and inverters.

[0005] Most household photovoltaic modules are connected in series and parallel, which has the advantage of increasing the output voltage while taking into account the output stability. At the same time, each photovoltaic module is equipped with a bypass diode in the junction box, which is cut off from the circuit when the module is in a certain shadow, which increases the output power while also protecting the module. The faults on the DC side include permanent short circuit and open circuit faults such as module delamination, internal bubbles, unit yellowing, scratches and hot spots, as well as temporary faults caused by factors such as local shadow coverage and dust pollution, and abnormal aging faults during the normal service life of the module. All of the above faults may lead to a decrease in the output power, efficiency and reliability of the household photovoltaic system. However, how to identify abnormal photovoltaics requires a lot of calculations and has low accuracy.

[0006] The identification of abnormal users is to reduce the scope and calculation amount of abnormal diagnosis and shorten the abnormal diagnosis time, which is of great significance for the subsequent maintenance and repair of abnormal users. At the same time, the identification of abnormal users can also provide users with corresponding abnormal alarm information, which helps household photovoltaic users to detect abnormalities in time and improve profits and equipment utilization. Summary of the invention

[0007] The technical problem to be solved and the technical task proposed by the present invention are to perfect and improve the existing technical solutions, and provide a household photovoltaic anomaly identification method based on smart meters and geographic information grouping to achieve the purpose of accurate identification; to this end, the present invention adopts the following technical solutions.

[0008] The household photovoltaic anomaly identification method based on smart meters and geographic information grouping includes the following steps:

[0009] 1) Obtaining raw data; the principle data includes the user’s geographical location, the photovoltaic load and power generation recorded by the smart energy meter;

[0010] 2) Clustering based on geographic location;

[0011] 3) Calculate the daily power generation per unit capacity of each photovoltaic user;

[0012] 4) Calculate the average and standard deviation of the daily power generation per unit capacity of photovoltaic users in the divided sub-areas;

[0013] 5) Determine whether the daily power generation per unit capacity of each photovoltaic user in the sub-region is an abnormal value based on the mean value and standard deviation;

[0014] 6) If it is an outlier, mark it as the first outlier;

[0015] 7) Determine whether there is a first abnormal value among all the values; if so, proceed to step 8); if not, proceed to step 10);

[0016] 8) The photovoltaic users corresponding to the first abnormal value are marked as abnormal users and removed; the photovoltaic users in the area are sorted out to obtain the screened photovoltaic users; and the process returns to step 4) to recalculate the average value and standard deviation;

[0017] 9) Output all first abnormal values ​​and their corresponding abnormal users;

[0018] 10) Obtain each PV user, and arrange the daily power generation per unit capacity of PV users in the divided sub-areas from small to large, and calculate their upper quartile and lower quartile;

[0019] 11) Calculate the interquartile range based on the quartiles and determine the upper and lower limits;

[0020] 12) judging whether the daily power generation per unit capacity of the photovoltaic users in the region is an abnormal value according to the upper limit and the lower limit, and if so, marking it as a second abnormal value;

[0021] 13) Output all second abnormal values ​​and their corresponding abnormal users;

[0022] 14) Taking the intersection of the abnormal users corresponding to the first abnormal value and the second abnormal value as the final abnormal identification output result; obtaining the final abnormal user;

[0023] 15) Send abnormal alarm information to the final abnormal photovoltaic user according to the abnormal identification output result; and repair the corresponding abnormal photovoltaic equipment.

[0024] This technical solution first clusters all photovoltaic users according to the geographical location information recorded in the installation archive, calculates the daily power generation per unit capacity of each photovoltaic user in each divided sub-area, identifies the value as an outlier based on the 3-sigma and box plot methods, and combines the results of the two methods to take the intersection as the final abnormal identification result; the calculation process of these two methods is simple, the required data volume and the amount of data calculated are small, and the data dimension obtained by relying on smart meters is also very limited. A single identification method may have the problem of misjudgment and missed judgment. The above two methods are used for comprehensive judgment, with high accuracy, reducing the resource loss caused by identification errors. And through the identification of abnormal users, the scope and calculation amount of abnormal diagnosis can be reduced, and the abnormal diagnosis time can be shortened, which is of great significance for the subsequent maintenance and repair of abnormal users; at the same time, the identification of abnormal users provides users with corresponding abnormal alarm information, which helps household photovoltaic users to find abnormalities in time and improve the income and equipment utilization rate.

[0025] As a preferred technical means: in step 1), the original information obtained also includes the reported capacity saved by household photovoltaic users during installation.

[0026] As a preferred technical means: in step 2), all photovoltaic users are clustered based on a large-scale distributed user photovoltaic spatial clustering method based on spatial correlation to obtain the optimal number of clusters and their corresponding sub-region divisions; within the divided sub-regions, the meteorological conditions and household photovoltaic output are consistent.

[0027] As a preferred technical means: in step 3), when calculating the daily power generation of photovoltaic users, corresponding conversion is made according to the different system capacities and service life, and it is converted into daily power generation per unit capacity. The conversion calculation formula is as follows:

[0028]

[0029] Among them, X is the converted daily power generation per unit capacity, X' is the daily power generation collected by the smart meter, ε is the linear property hedging, P set To declare the installed capacity.

[0030] As a preferred technical means: in step 4), the average daily power generation per unit capacity of photovoltaic users in the divided sub-areas is calculated. and standard deviation σ, the calculation formulas are shown in (2) and (3):

[0031]

[0032]

[0033] Where n is the total number of PV users in the sub-region.

[0034] As a preferred technical means: in step 5), when judging whether the daily power generation per unit capacity of each photovoltaic user is an abnormal value, if it meets formula (4), it is not a random error but a gross error, and can be identified as an abnormal value:

[0035]

[0036] It is determined whether any abnormal values ​​are identified. If so, the users corresponding to these abnormal values ​​are marked as abnormal users E1.

[0037] As a preferred technical means: in step 10), the daily power generation per unit capacity of the photovoltaic users in the divided sub-areas is arranged from small to large, and the upper quartile Q3 and the lower quartile Q1 are calculated; the upper quartile Q3 is the value at the 75% position after the sorting, and the lower quartile Q1 is the value at the 25% position after the sorting. The calculation formulas are shown in formulas (5) and (6):

[0038] Q3=(1+[(i+1) / 4]-(i+1) / 4)X [(i+1) / 4] +((i+1) / 4-[(i+1) / 4])X-(i+1) / 4]+1 (5)

[0039] Q1=(1+[3(i+1) / 4]-3(i+1) / 4)X [3(i+1) / 4] +(3(i+1) / 4-[3(i+1) / 4])X [3(i+1) / 4]+1 (6)

[0040] Where [x] is the rounding function, and the function value is the largest integer not exceeding the real number x; i is the total number of data X.

[0041] As a preferred technical means: in step 11), the interquartile range IQR is the difference between the upper quartile Q3 and the lower quartile Q1. The calculation formulas of the interquartile range IQR, the upper limit and the lower limit are shown in formulas (7), (8) and (9):

[0042] IQR=Q3-Q1 (7)

[0043] Upper limit=Q3+1.5IQR (8)

[0044] Lower limit = Q1-1.5IQR (9)

[0045] The upper and lower limits are the outlier cutoff points, and the area between the upper and lower limits is the inner limit; the data within the inner limit is the normal value, and the data outside the inner limit is the outlier; the users corresponding to these outliers are marked as abnormal users E2.

[0046] As a preferred technical means: in step 14), the intersection of the two abnormality recognition results is taken as the final abnormality recognition output result E:

[0047] E=E1∩E2.

[0048] Beneficial effects: This technical solution first clusters all photovoltaic users according to the geographical location information recorded in the installation archive, calculates the daily power generation per unit capacity of each photovoltaic user in each divided sub-area, identifies the value as an outlier based on the 3-sigma and box plot methods, and combines the results of the two methods to take the intersection as the final abnormal identification result. The calculation process of these two methods is simple, the required data volume and the amount of data calculated are small, and the data dimension obtained by relying on smart meters is also very limited. A single identification method may have the problem of misjudgment and missed judgment. The above two methods are used for comprehensive judgment, with high accuracy, reducing the resource loss caused by identification errors. And through the identification of abnormal users, the scope and calculation amount of abnormal diagnosis can be reduced, and the abnormal diagnosis time can be shortened, which is of great significance for the subsequent maintenance and repair of abnormal users; at the same time, the identification of abnormal users provides users with corresponding abnormal alarm information, which helps household photovoltaic users to find abnormalities in time and improve revenue and equipment utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings.

[0051] like Figure 1 As shown, the present invention comprises the following steps:

[0052] Step 1: Obtain the original data; this includes the data that household PV users leave on file with the power grid company during installation, generally including the reported capacity, the user’s geographic location, and the PV load and power generation recorded by the smart meter.

[0053] Step 2: Clustering based on geographic location; clustering all photovoltaic users based on the large-scale distributed user photovoltaic spatial clustering method based on spatial correlation. The algorithm can obtain the optimal number of clusters and their corresponding sub-region divisions, and it can be considered that within the divided sub-regions, the meteorological conditions and household photovoltaic output have good consistency.

[0054] Step 3: Calculate the daily power generation per unit capacity of each PV user; the daily power generation of PV users is converted accordingly according to the system capacity and service life, and converted into daily power generation per unit capacity. The conversion formula is as follows:

[0055]

[0056] Among them, X is the converted daily power generation per unit capacity, X' is the daily power generation collected by the smart meter, ε is the linear property hedging, P set To declare the installed capacity.

[0057] Step 4: Calculate the average and standard deviation of the daily power generation per unit capacity of PV users in the sub-regions after division; Calculate the average daily power generation per unit capacity of PV users in the sub-regions after division and standard deviation σ, the calculation formulas are shown in (4) and (5):

[0058]

[0059]

[0060] Where n is the total number of PV users in the sub-region.

[0061] Step 5: Determine whether the daily power generation per unit capacity of each PV user in the sub-region is an abnormal value based on the mean value and standard deviation; determine whether the daily power generation per unit capacity of each PV user is an abnormal value. If it meets the following formula, it is not a random error but a gross error and can be identified as an abnormal value:

[0062]

[0063] This determines whether or not an outlier has been identified.

[0064] Step 6: When it is an abnormal value, mark it as the first abnormal value; mark the users corresponding to these abnormal values ​​as abnormal users E1.

[0065] Step 7: Determine whether there is a first abnormal value among all the values; if so, proceed to step 8; if not, proceed to step 10.

[0066] Step 8: Mark the photovoltaic users corresponding to the first abnormal value as abnormal users and remove them; sort out the screened photovoltaic users in the area; and return to step 4) to recalculate the average value and standard deviation.

[0067] Step 9: Output all first abnormal values ​​and their corresponding abnormal users E1.

[0068] Step 10: Obtain each PV user, and arrange the daily power generation per unit capacity of the PV users in the divided sub-areas from small to large, and calculate their upper quartile Q3 and lower quartile Q1. The upper quartile Q3 is the value at the 75% position after sorting, and the lower quartile Q1 is the value at the 25% position after sorting. The calculation formulas are shown in equations (5) and (6):

[0069] Q3=(1+[(i+1) / 4]-(i+1) / 4)X [(i+1) / 4] +((i+1) / 4-[(i+1) / 4])X [(i+1) / 4]+1 (5)

[0070] Q1=(1+[3(i+1) / 4]-3(i+1) / 4)X [3(i+1) / 4] +(3(i+1) / 4-[3(i+1) / 4])X [3(i+1) / 4]+1 (6)

[0071] Where [x] is the rounding function, and the function value is the largest integer not exceeding the real number x; i is the total number of data X.

[0072] Step 11: Calculate the interquartile range based on the quartiles and determine the upper and lower limits. The interquartile range IQR is the difference between the upper quartile Q3 and the lower quartile Q1. The calculation formulas for the interquartile range IQR, upper and lower limits are shown in equations (7), (8) and (9):

[0073] IQR=Q3-Q1 (7)

[0074] Upper limit=Q3+1.5IQR (8)

[0075] Lower limit = Q1-1.5IQR (9)

[0076] The upper and lower limits are the outlier cutoff points.

[0077] Step 12: Determine whether the daily power generation per unit capacity of the photovoltaic users in the area is an abnormal value based on the upper and lower limits, and the range between the upper and lower limits is the inner limit. The data within the inner limit is the normal value, and the data outside the inner limit is the second abnormal value; the users corresponding to these second abnormal values ​​are marked as abnormal users E2.

[0078] Step 13: Output all second abnormal values ​​and their corresponding abnormal users E2.

[0079] Step 14: Take the intersection of the abnormal users corresponding to the first abnormal value and the second abnormal value as the final abnormal identification output result; obtain the final abnormal user E:

[0080] E=E1∩E2.

[0081] Step 15: Send abnormal alarm information to the final abnormal photovoltaic user according to the abnormal identification output result; and repair the corresponding abnormal photovoltaic equipment.

[0082] above Figure 1 The household photovoltaic anomaly identification method based on smart meters and geographic information grouping is a specific embodiment of the present invention, which has embodied the substantial characteristics and progress of the present invention. According to actual use needs and under the guidance of the present invention, equivalent modifications in shape, structure, etc. can be made to it, which are all within the protection scope of this scheme.

Claims

1. A household photovoltaic anomaly identification method based on smart meters and geographic information clustering, characterized by The following steps are involved: 1) Obtaining raw data; the principle data includes the user’s geographical location, the photovoltaic load and power generation recorded by the smart energy meter; 2) Clustering based on geographic location; 3) Calculate the daily power generation per unit capacity of each photovoltaic user; 4) Calculate the average and standard deviation of the daily power generation per unit capacity of photovoltaic users in the divided sub-areas; 5) Determine whether the daily power generation per unit capacity of each photovoltaic user in the sub-region is an abnormal value based on the mean value and standard deviation; 6) If it is an outlier, mark it as the first outlier; 7) Determine whether there is a first abnormal value among all the values; if so, proceed to step 8); if not, proceed to step 10); 8) The photovoltaic users corresponding to the first abnormal value are marked as abnormal users and removed; the photovoltaic users in the area are sorted out to obtain the screened photovoltaic users; and the process returns to step 4) to recalculate the average value and standard deviation; 9) Output all first abnormal values ​​and their corresponding abnormal users; 10) Obtain each PV user, and arrange the daily power generation per unit capacity of PV users in the divided sub-areas from small to large, and calculate their upper quartile and lower quartile; 11) Calculate the interquartile range based on the quartiles and determine the upper and lower limits; 12) judging whether the daily power generation per unit capacity of the photovoltaic users in the region is an abnormal value according to the upper limit and the lower limit, and if so, marking it as a second abnormal value; 13) Output all second abnormal values ​​and their corresponding abnormal users; 14) Taking the intersection of the abnormal users corresponding to the first abnormal value and the second abnormal value as the final abnormal identification output result; obtaining the final abnormal user; 15) Send abnormal alarm information to the final abnormal photovoltaic user according to the abnormal identification output result; and repair the corresponding abnormal photovoltaic equipment; In step 3), when calculating the daily power generation of photovoltaic users, the corresponding conversion is made according to the system capacity and service life, and it is converted into daily power generation per unit capacity. The conversion calculation formula is as follows: Among them, X is the converted daily power generation per unit capacity, X' is the daily power generation collected by the smart meter, ε is the linear property hedging, P set To declare the installed capacity; In step 4), the average daily power generation per unit capacity of photovoltaic users in the divided sub-areas is calculated. and standard deviation σ, the calculation formulas are shown in (2) and (3): Where n is the total number of PV users in the sub-region; In step 5), when judging whether the daily power generation per unit capacity of each photovoltaic user is an abnormal value, if it meets formula (4), it is not a random error but a gross error and can be identified as an abnormal value: Determine whether any abnormal values ​​are identified. If so, mark the users corresponding to these abnormal values ​​as abnormal users E1; In step 10), the daily power generation per unit capacity of the PV users in the divided sub-regions is arranged from small to large, and the upper quartile Q3 and lower quartile Q1 are calculated; the upper quartile Q3 is the value at the 75% position after sorting, and the lower quartile Q1 is the value at the 25% position after sorting. The calculation formulas are shown in equations (5) and (6): Q3=(1+[(i+1) / 4]-(i+1) / 4)X [(i+1) / 4] +((i+1) / 4-[(i+1) / 4])X [(i+1) / 4]+1 (5) Q1=(1+[3(i+1) / 4]-3(i+1) / 4)X [3(i+1) / 4] +(3(i+1) / 4-[3(i+1) / 4])X [3(i+1) / 4]+1 (6) Where [x] is the rounding function, and the function value is the largest integer that does not exceed the real number x; i is the total number of data X; In step 11), the interquartile range IQR is the difference between the upper quartile Q3 and the lower quartile Q1. The calculation formulas for the interquartile range IQR, the upper limit, and the lower limit are shown in equations (7), (8), and (9): IQR=Q3-Q1 (7) Upper limit=Q3+1.5IQR (8) Lower limit = Q1-1.5IQR (9) The upper and lower limits are the outlier cutoff points, and the area between the upper and lower limits is the inner limit; the data within the inner limit is the normal value, and the data outside the inner limit is the outlier; the users corresponding to these outliers are marked as abnormal users E2; In step 14), the intersection of the two abnormality recognition results is taken as the final abnormality recognition output result E: E=E1∩E2.

2. The household photovoltaic anomaly identification method based on smart meters and geographic information grouping according to claim 1 is characterized by: In step 1), the original information obtained also includes the reported capacity saved by household photovoltaic users during installation.

3. The household photovoltaic anomaly identification method based on smart meters and geographic information grouping according to claim 2 is characterized by: In step 2), all photovoltaic users are clustered based on a large-scale distributed user photovoltaic spatial clustering method based on spatial correlation to obtain the optimal number of clusters and their corresponding sub-region divisions; within the divided sub-regions, the meteorological conditions and household photovoltaic output are consistent.

Citation Information

Patent Citations

  • Distributed photovoltaic electricity-stealing supervising method based on multi-time scale output estimation

    CN105160595A

  • Robust-regression-based distributed photovoltaic generating electricity-stealing identification method

    CN105337574A