Algorithm for automatically extracting and identifying photovoltaic abnormal faults

Abstract

The invention discloses an algorithm for automatically extracting and identifying a photovoltaic abnormal fault, and the algorithm comprises the steps: S1, importing photovoltaic power station data, collecting and classifying power station design information data, power station historical working data and power station historical meteorological data in a database, so as to obtain first data; s2, photovoltaic power station data preprocessing is carried out, integrity and rationality checking is carried out on first data, and type conversion and normalization processing are carried out on the checked first data to obtain actual photovoltaic system operation condition data; and S3, identifying the fault of each link of the photovoltaic system. According to the algorithm for automatically extracting and identifying the photovoltaic abnormal faults, an abnormal fault sequence is rapidly and accurately extracted through data acquisition and analysis and intelligent fault prediction and diagnosis, and digital, intelligent, remote and unmanned operation and maintenance of a photovoltaic power station are achieved through accurate positioning and calculation judgment.

Description

Technical Field

[0001] The invention belongs to the technical field of abnormal fault identification of photovoltaic power stations, and in particular relates to an algorithm for automatically extracting and identifying abnormal photovoltaic faults. Background Art

[0002] Due to the late start of the distributed photovoltaic industry, its informatization construction has not yet been fully developed. In terms of operation management and maintenance and repair, power stations often rely on manual data monitoring, data analysis and statistics, reporting of analysis results, and formulation of corresponding operation and maintenance plans. However, with the rapid increase in the output of distributed photovoltaic power stations, if the above work is still completed manually, due to the huge workload and lack of personnel experience, there will inevitably be omissions and errors in analysis results, which will not only reduce the power generation efficiency of the power station, but may even cause greater failures or accidents. Therefore, how to efficiently and timely discover problems in the operation and management of power stations, quickly and accurately determine the causes, and solve problems in a standardized and effective manner is a problem that most photovoltaic operating companies are currently facing. Summary of the invention

[0003] The main purpose of the present invention is to provide an algorithm for automatically extracting and identifying abnormal photovoltaic faults, using data collection and analysis, intelligent fault prediction and diagnosis to quickly and accurately extract abnormal fault sequences, perform precise positioning and calculation and judgment to realize the digitalization, intelligence, remoteness and unmanned operation and maintenance of photovoltaic power stations.

[0004] To achieve the above objectives, the present invention provides an algorithm for automatically extracting and identifying photovoltaic abnormal faults, comprising the following steps:

[0005] Step S1: importing photovoltaic power station data, collecting and classifying power station design information data, power station historical working data and power station historical meteorological data in the database to obtain first data;

[0006] Step S2: Preprocess the photovoltaic power station data (due to the different units and large differences in magnitude of the data of each power station, there is a certain range limit on the output, and the data needs to be preprocessed), check the integrity and rationality of the first data, and perform type conversion on the first data after the inspection (different data acquisition systems are different, and each system stores different data types. When the evaluation platform reads the data, it needs to be converted before it can be used by the system) and normalization processing to obtain the actual photovoltaic system operation condition data (here, the data of each photovoltaic module is classified, including strings, inverters and grid connection points, etc., for subsequent comparative analysis);

[0007] Step S3: After the actual PV system operating condition data is imported into the platform, the actual power generation is compared with the designed predicted power generation (before each PV power station is installed, the power generation of the PV power station will be roughly calculated based on the scale of each component of the PV power station and the local light radiation, etc.) from the three dimensions of structure, link and time, as well as PV power stations of different structural types (including cement roofs and color steel tile roofs). The differences between the PV system operating condition data before and after and the differences with the predicted power generation are compared, so as to identify faults in various links of the PV system (strings, inverters, and grid connection).

[0008] As a further preferred technical solution of the above technical solution, in step S1:

[0009] The power station design information data includes: the name of the photovoltaic power station, the longitude and latitude of the area where it is located, the length of time it has been put into operation, the grid-connected voltage level, the number of photovoltaic modules, the rated power of the photovoltaic modules, the actual capacity and rated capacity, the module installation method (angle installation or flat laying), the inverter configuration, the confluence method and the transformer nameplate information;

[0010] The historical working data of the photovoltaic power station includes: online power generation (obtained by the power supply bureau), inverter output data, inverter input data and photovoltaic string data;

[0011] The historical meteorological data of the power station include: light intensity, temperature and wind speed (the requirements for each type of meteorological data are also different);

[0012] (The platform database has a large amount of data, so the input data needs to be classified) The classification criteria are:

[0013] Classify according to the region where the power station is located (the meteorological conditions in the same region are similar, and the working status of the power station is mainly affected by the design and operation and maintenance of the power station);

[0014] Classification by structural type (there are differences in the design of power stations with cement roofs and color steel tile roofs);

[0015] Classify according to the grid connection level of the power station (the losses in the grid connection link are different between the low-voltage grid connection method and the high-voltage grid connection method, and the transformer loss needs to be calculated for the high-voltage grid connection);

[0016] Classify according to the specifications of electrical equipment (electrical equipment with the same specifications has similar brand parameters and similar working performance).

[0017] As a further preferred technical solution of the above technical solution, in step S2:

[0018] For integrity inspection, according to the sampling time and time span requirements, check whether the amount of data collected is consistent with the set planned amount, whether the collected data is arranged in a certain order, and whether the collection time point is the same as the plan;

[0019] For rationality testing, determine whether the measured data is within a reasonable preset range and is consistent with the current situation;

[0020] For normalization processing, (in order to improve the accuracy of evaluation) the historical working data of the power station and various historical meteorological data of the power station are screened to remove singular data. When conducting grid-connected fault diagnosis, inverter fault diagnosis and photovoltaic component fault diagnosis for the photovoltaic power station, normalization processing is performed to ensure the validity of the comparison results.

[0021] As a further preferred technical solution of the above technical solution, in step S3:

[0022] Analyze from the time dimension. Different time dimensions correspond to different operation and maintenance priorities.

[0023] From the perspective of link dimensions, different link dimensions use different system efficiency coefficients K when calculating the predicted power generation of power plant design;

[0024] From the perspective of structural dimension analysis, different structural dimensions use different design information data when calculating the predicted power generation of power plant operation, and the calculated total radiation exposure (HTX) is also different.

[0025] As a further preferred technical solution of the above technical solution, in step S3, comparison (due to the large amount of power plant data information, a variety of data comparison methods are used when comparing power plant data information) is performed, including month-on-month (comparison between months of a year, comparison between days of a month, etc.), year-on-year (comparison between the same month in two consecutive years, comparison between time periods with similar environmental conditions, etc.) and comparison between actual working parameters and predicted values.

[0026] As a further preferred technical solution of the above technical solution, in step S3, the operation data of the photovoltaic power station is analyzed from the time dimension, and the types of photovoltaic power station faults are divided into three categories: emergency stop faults, equipment inefficient operation and equipment performance degradation, among which:

[0027] For emergency stop faults, the characteristics of emergency stop faults are that when the fault occurs, it directly affects the operation or communication of the equipment. The operation and maintenance priority of this type of fault is the highest. It is necessary to analyze the power station operation data in a short period of time to discover the fault in time, which includes:

[0028] String open circuit fault

[0029] By comparing the input voltage and current data of each string before and after the preset time (5 minutes), the current string voltage and current values ​​suddenly change to 0, and the voltage values ​​of other strings under the same inverter are greater than the inverter startup voltage, and the current values ​​are greater than 0 (month-on-month);

[0030] Inverter communication failure

[0031] Through the preset time (5 minutes) before and after each inverter input, output power and grid-connected point power, when the inverter communication failure occurs, the inverter works normally, the inverter input and output power suddenly drops to 0, and the total grid-connected power of the inverter is greater than the sum of the normal working inverter power;

[0032] Abnormal shutdown of the inverter

[0033] Through the preset time (5 minutes) before and after each inverter, the input and output power and the grid-connected point power are calculated. When the inverter communication failure occurs, the inverter works normally, the input and output power of the inverter suddenly change to 0, and the total grid-connected power of the grid-connected point where this inverter is located is equal to the sum of the normal working inverter power (month-on-month);

[0034] Grid connection point open circuit fault

[0035] By comparing the input power of the grid connection point before and after the preset time (5 minutes), when an open circuit fault occurs at the grid connection point, the input power of the grid connection point suddenly changes to 0, and the grid connection power of the grid connection point in the same area or the same power station is greater than 0 (month-on-month);

[0036] For inefficient equipment operation, the characteristics of inefficient equipment operation failure are that when the failure occurs, it will not affect the operation of the photovoltaic power station, but will reduce the operating efficiency of the photovoltaic power station. The operation and maintenance priority of this type of failure is higher, and it is necessary to analyze the power station operation data within a week or even a month to diagnose and troubleshoot the failure, which includes;

[0037] Hot spot phenomenon of photovoltaic modules

[0038] The detection of hot spot phenomenon of photovoltaic modules requires comparative analysis of meteorological conditions and input voltage and current values ​​of the strings. The irradiation intensity data obtained by the meteorological instrument is used to determine that as the irradiance increases, the output voltage of the photovoltaic strings decreases. The voltage and current of each string are compared, and the voltage and current discrete rates of each string are calculated. The calculation formula is as follows:

[0039]

[0040] Where: μ is the average value of each string voltage (current);

[0041] σ is the standard deviation of the voltage (current) of each string;

[0042]

[0043]

[0044] x i is the voltage (current) of the i-th string;

[0045] If the voltage and current discrete rate of each string is less than 5%, the string is working stably and well. If the voltage and current discrete rate of each string is greater than 5%, the similarity with the irradiance analysis curve at the corresponding moment is further analyzed. By analyzing the power station operation data of each day within a month, if the similarity decreases with the increase of irradiation intensity, there are hot spots in the photovoltaic components in the string (it should be included in the monthly inspection plan. The above is a month-on-month comparison. It can also be compared with the prediction. The actual power generation of the photovoltaic string is lower than the predicted power generation; the stronger the irradiance, the greater the difference between the output voltage of the photovoltaic string and the average string output voltage);

[0046] Photovoltaic panels are shadowed

[0047] (During the design and survey stage of the PV power station, shadows of tall buildings and roof shielding objects around the power station will be measured. During the operation of the power station, shadows will be blocked in the power station due to building renovation, weeds, and trees. There is a regular time for this shadow blocking. The bypass diode of the PV module blocked by the shadow will be turned on, causing the output voltage of the string to drop.) By calculating the discrete rate of voltage and current of each string, and comparing the discrete rate of string voltage in the preset time period (every 15 minutes from 9 am to 3 pm) within a week, if the voltage discrete rate is greater than 5% and the time period is consistent every day, it is judged that the PV module is shadowed in this time period (the above is a month-on-month comparison, and it can also be compared with the prediction. The actual power generation is lower than the predicted power generation. At sunrise or sunset, the difference between the output voltage of the PV string and the average output voltage of the string changes);

[0048] Photovoltaic module dust pollution

[0049] (The operating environment of photovoltaic power stations is harsh. The surrounding pollution sources will cause dust pollution to photovoltaic modules, which will directly affect the working efficiency of photovoltaic modules. The output current of photovoltaic modules polluted by dust will decrease, and the dust will appear on a large area in the power station, resulting in a decrease in the power generation of the entire power station.) The actual power generation of photovoltaic strings is lower than the predicted power generation, and the difference is increasing. The output current of the strings is lower than the output current of the strings under similar historical environments;

[0050] Equipment performance degradation includes:

[0051] Inverter inefficient operation

[0052] According to the inverter input data and inverter output data, the actual efficiency of the inverters of different power stations in the same area is calculated. According to the power station design information and historical meteorological data, the normal efficiency thresholds of different power stations in the local area are calculated (different time, different meteorological conditions, different power stations, the calculated thresholds are different). The inverter loss of the photovoltaic power station is evaluated through data comparison and analysis. The actual efficiency calculation formula of the inverter is:

[0053]

[0054] Where: E pni Input power to the inverter;

[0055] E pno Input power to the inverter;

[0056] Calculate the working efficiency of the inverter, and analyze the working efficiency after averaging, η n The normal range is 95% to 98.5% (the actual output power of the inverter is lower than the predicted power generation; the difference between the input-efficiency curve and the output detection curve is large);

[0057] Grid connection loss is too high

[0058] According to the online power generation (obtained by the power supply bureau) and the inverter output data, the actual grid-connected loss of the power station is calculated, and according to the power station design information, the normal grid-connected loss threshold of the power station is calculated. Through data comparison and analysis, the grid-connected loss of the photovoltaic power station is evaluated to determine whether there is a fault in the grid-connected link of the power station, and the power station with high grid-connected loss is warned (written into the monthly or quarterly inspection plan);

[0059]

[0060] Where: η b The grid connection loss of the power station;

[0061] E pb The amount of electricity generated by the power station connected to the grid;

[0062] E pno It is the power generation on the output side of the power station inverter (the actual output power generation at the grid-connected point is lower than the predicted power generation, and lower than the grid-connected power under the same historical input conditions).

[0063] The beneficial effects of the present invention are:

[0064] By inventing this algorithm to conduct preliminary fault diagnosis and operation status evaluation of photovoltaic power stations, it provides strong support for operation and maintenance decisions, strengthens the standardization and intelligence of photovoltaic power station management, and ensures efficient and stable operation of photovoltaic power stations. Based on the information of photovoltaic power station power generation efficiency, grid-connected loss, grid-connected power generation, inverter output data, inverter input data, photovoltaic string data, historical meteorological data, power station design information, power station equipment information, etc., the photovoltaic power station is evaluated and analyzed in terms of power generation prediction, grid-connected fault diagnosis, inverter fault diagnosis, and photovoltaic component fault diagnosis, to find out the design defects of photovoltaic power stations, and provide data support for future photovoltaic power station design optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a flow chart of an algorithm for automatically extracting and identifying abnormal photovoltaic faults according to the present invention.

[0066] Figure 2 It is a comparative analysis diagram of an algorithm for automatically extracting and identifying photovoltaic abnormal faults according to the present invention. DETAILED DESCRIPTION

[0067] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be applied to other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0068] In the preferred embodiment of the present invention, those skilled in the art should note that the photovoltaic power station and the design and prediction of power generation involved in the present invention can be regarded as prior art.

[0069] Preferred embodiments.

[0070] like Figure 1-2 As shown, the present invention discloses an algorithm for automatically extracting and identifying photovoltaic abnormal faults, comprising the following steps:

[0071] Step S1: importing photovoltaic power station data, collecting and classifying power station design information data, power station historical working data and power station historical meteorological data in the database to obtain first data;

[0072] Step S2: Preprocess the photovoltaic power station data (due to the different units and large differences in magnitude of the data of each power station, there is a certain range limit on the output, and the data needs to be preprocessed), check the integrity and rationality of the first data, and perform type conversion on the first data after the inspection (different data acquisition systems are different, and each system stores different data types. When the evaluation platform reads the data, it needs to be converted before it can be used by the system) and normalization processing to obtain the actual photovoltaic system operation condition data (here, the data of each photovoltaic module is classified, including strings, inverters and grid connection points, etc., for subsequent comparative analysis);

[0073] Step S3: After the actual PV system operating condition data is imported into the platform, the actual power generation is compared with the designed predicted power generation (according to the scale of each component of the PV power station and the local light radiation, etc., the power generation of the PV power station can be roughly calculated) from the three dimensions of structure, link and time, as well as PV power stations of different structural types (including cement roofs and color steel tile roofs). The differences between the PV system operating condition data before and after and the differences with the predicted power generation are compared, so as to identify the faults of each link (string, inverter, grid connection) of the PV system (including multiple PV power stations).

[0074] Specifically, in step S1:

[0075] The power station design information data includes: the name of the photovoltaic power station, the longitude and latitude of the area where it is located, the length of time it has been put into operation, the grid-connected voltage level, the number of photovoltaic modules, the rated power of the photovoltaic modules, the actual capacity and rated capacity, the module installation method (angle installation or flat laying), the inverter configuration, the confluence method and the transformer nameplate information;

[0076] The historical working data of the photovoltaic power station includes: online power generation (obtained by the power supply bureau), inverter output data, inverter input data and photovoltaic string data;

[0077] The historical meteorological data of the power station include: light intensity, temperature and wind speed (the requirements for each type of meteorological data are also different);

[0078] (The platform database has a large amount of data, so the input data needs to be classified) The classification criteria are:

[0079] Classify according to the region where the power station is located (the meteorological conditions in the same region are similar, and the working status of the power station is mainly affected by the design and operation and maintenance of the power station);

[0080] Classification by structural type (there are differences in the design of power stations with cement roofs and color steel tile roofs);

[0081] Classify according to the grid connection level of the power station (the losses in the grid connection link are different between the low-voltage grid connection method and the high-voltage grid connection method, and the transformer loss needs to be calculated for the high-voltage grid connection);

[0082] Classify according to the specifications of electrical equipment (electrical equipment with the same specifications has similar brand parameters and similar working performance).

[0083] More specifically, in step S2:

[0084] For integrity inspection, according to the sampling time and time span requirements, check whether the amount of data collected is consistent with the set planned amount, whether the collected data is arranged in a certain order, and whether the collection time point is the same as the plan;

[0085] For rationality testing, determine whether the measured data is within a reasonable preset range and is consistent with the current situation;

[0086] For normalization processing, (in order to improve the accuracy of evaluation) the historical working data of the power station and various historical meteorological data of the power station are screened to remove singular data. When conducting grid-connected fault diagnosis, inverter fault diagnosis and photovoltaic component fault diagnosis for the photovoltaic power station, normalization processing is performed to ensure the validity of the comparison results.

[0087] It is worth mentioning that for normalization processing, for example: the online power generation is the power generation calculated by the grid-connected voltage and grid-connected current measured by the two-way meter installed by the local power supply bureau after the grid-connected point. In order to facilitate calculation, the daily start and end active power meter codes are generally obtained. The historical data can be obtained by querying the local power supply bureau, and the actual daily online power generation of the photovoltaic power station is calculated based on the voltage sensor and current sensor multipliers. The calculation formula is as follows:

[0088] E p3 =(W2-W1)K CT K PT

[0089] Where: E p3 ——On-grid electricity generation;

[0090] W2——meter active power stop code (must be greater than W1);

[0091] W1——The initial code of active power of the meter;

[0092] K CT ——Current transformer ratio;

[0093] K PT ——Voltage transformer ratio.

[0094] Furthermore, in step S3:

[0095] Analyze from the time dimension. Different time dimensions correspond to different operation and maintenance priorities.

[0096] From the perspective of link dimensions, different link dimensions use different system efficiency coefficients K when calculating the predicted power generation of power plant design;

[0097] From the perspective of structural dimension analysis, different structural dimensions use different design information data when calculating the predicted power generation of power plant operation, and the calculated total radiation exposure (HTX) is also different.

[0098] (For any photovoltaic power station, the design prediction of power generation has multi-dimensional and diversified characteristics. For example, for a distributed photovoltaic power station on a cement roof, when calculating the design prediction of power generation, the daily prediction of power generation, the monthly prediction of power generation, and the annual prediction of power generation can be calculated according to the time dimension, and the string prediction of power generation, the inverter prediction of power generation, and the grid-connected prediction of power generation can be calculated according to the link dimension. When conducting a comparative analysis of power generation, the predicted power generation will be positioned in three dimensions, such as: cement roof (structure) string (link) daily (time) prediction of power generation)

[0099] Furthermore, in step S3, comparison (due to the large amount of power plant data information, a variety of data comparison methods are used when comparing power plant data information) is performed, including month-on-month (comparison between months of a year, comparison between days of a month, etc.), year-on-year (comparison between the same month in two consecutive years, comparison between time periods with similar environmental conditions, etc.) and comparison between actual working parameters and predicted values.

[0100] Preferably, in step S3, the photovoltaic power station operation data is analyzed from the time dimension, and the photovoltaic power station fault types are divided into three categories: emergency stop fault, equipment inefficient operation and equipment performance degradation, where:

[0101] For emergency stop faults, the characteristics of emergency stop faults are that when the fault occurs, it directly affects the operation or communication of the equipment. The operation and maintenance priority of this type of fault is the highest. It is necessary to analyze the power station operation data in a short period of time to discover the fault in time, which includes:

[0102] String open circuit fault

[0103] By comparing the input voltage and current data of each string before and after the preset time (5 minutes), the current string voltage and current values ​​suddenly change to 0, and the voltage values ​​of other strings under the same inverter are greater than the inverter startup voltage, and the current values ​​are greater than 0 (month-on-month);

[0104] Inverter communication failure

[0105] Through the preset time (5 minutes) before and after each inverter input, output power and grid-connected point power, when the inverter communication failure occurs, the inverter works normally, the inverter input and output power suddenly drops to 0, and the total grid-connected power of the inverter is greater than the sum of the normal working inverter power;

[0106] Abnormal shutdown of the inverter

[0107] Through the preset time (5 minutes) before and after each inverter, the input and output power and the grid-connected point power are calculated. When the inverter communication failure occurs, the inverter works normally, the input and output power of the inverter suddenly change to 0, and the total grid-connected power of the grid-connected point where this inverter is located is equal to the sum of the normal working inverter power (month-on-month);

[0108] Grid connection point open circuit fault

[0109] By comparing the input power of the grid connection point before and after the preset time (5 minutes), when an open circuit fault occurs at the grid connection point, the input power of the grid connection point suddenly changes to 0, and the grid connection power of the grid connection point in the same area or the same power station is greater than 0 (month-on-month);

[0110] For inefficient equipment operation, the characteristics of inefficient equipment operation failure are that when the failure occurs, it will not affect the operation of the photovoltaic power station, but will reduce the operating efficiency of the photovoltaic power station. The operation and maintenance priority of this type of failure is higher, and it is necessary to analyze the power station operation data within a week or even a month to diagnose and troubleshoot the failure, which includes;

[0111] Hot spot phenomenon of photovoltaic modules

[0112] (Based on a large amount of data accumulation and information research, authoritative testing agencies have shown that when the irradiance is greater than 800W / m2, the temperature difference between the highest temperature of the hot spot and the average temperature of the component is greater than 10 degrees, and the photovoltaic component is judged to have a hot spot. Usually, photovoltaic components have 2-3 parallel diodes. When the light intensity is high and the photovoltaic component has a hot spot effect, the diode is turned on to protect the photovoltaic component.) The detection of the hot spot phenomenon of photovoltaic components requires the use of meteorological conditions and the comparative analysis of the input voltage and current values ​​of the string. The irradiation intensity data obtained by the meteorological instrument is used to determine that as the irradiance increases, the output voltage of the photovoltaic string decreases. The voltage and current of each string are compared, and the voltage and current discrete rates of each string are calculated respectively. The calculation formula is as follows:

[0113]

[0114] Where: μ is the average value of each string voltage (current);

[0115] σ is the standard deviation of the voltage (current) of each string;

[0116]

[0117]

[0118] x i is the voltage (current) of the i-th string;

[0119] If the voltage and current discrete rate of each string is less than 5%, the string is working stably and well. If the voltage and current discrete rate of each string is greater than 5%, the similarity with the irradiance analysis curve at the corresponding moment is further analyzed. By analyzing the power station operation data of each day within a month, if the similarity decreases with the increase of irradiation intensity, there are hot spots in the photovoltaic components in the string (it should be included in the monthly inspection plan. The above is a month-on-month comparison. It can also be compared with the prediction. The actual power generation of the photovoltaic string is lower than the predicted power generation; the stronger the irradiance, the greater the difference between the output voltage of the photovoltaic string and the average string output voltage);

[0120] Photovoltaic panels are shaded

[0121] (During the design and survey stage of the PV power station, shadows of tall buildings and roof shielding objects around the power station will be measured. During the operation of the power station, shadows will be blocked in the power station due to building renovation, weeds, and trees. There is a regular time for this shadow blocking. The bypass diode of the PV module blocked by the shadow will be turned on, causing the output voltage of the string to drop.) By calculating the discrete rate of voltage and current of each string, and comparing the discrete rate of string voltage in the preset time period (every 15 minutes from 9 am to 3 pm) within a week, if the voltage discrete rate is greater than 5% and the time period is consistent every day, it is judged that the PV module is shadowed in this time period (the above is a month-on-month comparison, and it can also be compared with the prediction. The actual power generation is lower than the predicted power generation. At sunrise or sunset, the difference between the output voltage of the PV string and the average output voltage of the string changes);

[0122] Photovoltaic module dust pollution

[0123] (The operating environment of photovoltaic power stations is harsh. The surrounding pollution sources will cause dust pollution to photovoltaic modules, which will directly affect the working efficiency of photovoltaic modules. The output current of photovoltaic modules polluted by dust will decrease, and the dust will appear on a large area in the power station, resulting in a decrease in the power generation of the entire power station.) The actual power generation of photovoltaic strings is lower than the predicted power generation, and the difference is increasing. The output current of the strings is lower than the output current of the strings under similar historical environments;

[0124] Equipment performance degradation includes:

[0125] Inverter inefficient operation

[0126] According to the inverter input data and inverter output data, the actual efficiency of the inverters of different power stations in the same area is calculated. According to the power station design information and historical meteorological data, the normal efficiency thresholds of different power stations in the local area are calculated (different time, different meteorological conditions, different power stations, the calculated thresholds are different). The inverter loss of the photovoltaic power station is evaluated through data comparison and analysis. The actual efficiency calculation formula of the inverter is:

[0127]

[0128] Where: E pni Input power to the inverter;

[0129] E pno Input power to the inverter;

[0130] Calculate the working efficiency of the inverter, and analyze the working efficiency after averaging, η n The normal range is 95% to 98.5% (the actual output power of the inverter is lower than the predicted power generation; the difference between the input-efficiency curve and the output detection curve is large);

[0131] Grid connection loss is too high

[0132] According to the online power generation (obtained by the power supply bureau) and the inverter output data, the actual grid-connected loss of the power station is calculated, and according to the power station design information, the normal grid-connected loss threshold of the power station is calculated. Through data comparison and analysis, the grid-connected loss of the photovoltaic power station is evaluated to determine whether there is a fault in the grid-connected link of the power station, and the power station with high grid-connected loss is warned (written into the monthly or quarterly inspection plan);

[0133]

[0134] Where: η b The grid connection loss of the power station;

[0135] E pb The amount of electricity generated by the power station connected to the grid;

[0136] E pno It is the power generation on the output side of the power station inverter (the actual output power generation at the grid-connected point is lower than the predicted power generation, and lower than the grid-connected power under the same historical input conditions).

[0137] The existing calculation for the design forecast power generation is:

[0138] The environmentally predicted power generation is calculated based on the geographical location and local meteorological conditions, with the local horizontal surface's total annual solar radiation as the basis and the K1 empirical correction factor.

[0139]

[0140] Where: Es is the irradiance under standard conditions; E p1 is the annual environmental predicted power generation (kW·h); P AZ H is the installed capacity of the module (kWp); T is the average annual solar radiation on the horizontal surface (kW·h / m 2 ) provided by the Meteorological Bureau; K1 is the comprehensive efficiency coefficient. It can be obtained through experience of grid-connected power generation of completed power stations, which is about 0.8-0.85.

[0141] It is worth mentioning that the technical features such as photovoltaic power stations and designed predicted power generation involved in the patent application of this invention should be regarded as prior art. The specific structure, working principle and possible control method and spatial layout method of these technical features can be selected according to the conventional selection in the field, and should not be regarded as the inventive point of the patent of this invention. The patent of this invention will not be further elaborated.

[0142] For those skilled in the art, it is still possible to modify the technical solutions described in the aforementioned embodiments, or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An algorithm for automatically extracting and identifying photovoltaic abnormal faults, characterized in that: The following steps are involved: Step S1: importing photovoltaic power station data, collecting and classifying power station design information data, power station historical working data and power station historical meteorological data in the database to obtain first data; Step S2: preprocessing the photovoltaic power station data, checking the integrity and rationality of the first data, and performing type conversion and normalization processing on the checked first data to obtain actual photovoltaic system operating condition data; Step S3: After the actual PV system operating condition data is imported into the platform, the actual power generation is analyzed and compared with the designed predicted power generation from the three dimensions of structure, link and time, as well as PV power stations of different structural types, and the differences between the PV system operating condition data before and after and the predicted power generation are compared, so as to identify the faults in various links of the PV system.

2. The algorithm for automatically extracting and identifying photovoltaic abnormal faults according to claim 1 is characterized in that: In step S1: The power station design information data includes: the name of the photovoltaic power station, the longitude and latitude of the area where it is located, the length of time it has been put into operation, the grid-connected voltage level, the number of photovoltaic modules, the rated power of the photovoltaic modules, the actual capacity and rated capacity, the module installation method, the inverter configuration, the confluence method and the transformer nameplate information; The historical working data of the photovoltaic power station includes: online power generation, inverter output data, inverter input data and photovoltaic string data; The power station’s historical meteorological data include: light intensity, temperature and wind speed; The classification criteria are: Classification according to the region where the power station is located; Classification by structural type; Classification according to the grid connection level of power stations; Classify according to electrical equipment specifications.

3. The algorithm for automatically extracting and identifying photovoltaic abnormal faults according to claim 2 is characterized in that: In step S2: For integrity inspection, according to the sampling time and time span requirements, check whether the amount of data collected is consistent with the set planned amount, whether the collected data is arranged in a certain order, and whether the collection time point is the same as the plan; For rationality testing, determine whether the measured data is within a reasonable preset range and is consistent with the current situation; For normalization processing, the historical working data of the power station and various historical meteorological data of the power station are screened to remove singular data. When conducting grid-connected fault diagnosis, inverter fault diagnosis and photovoltaic component fault diagnosis for the photovoltaic power station, normalization processing is performed to ensure the validity of the comparison results.

4. The algorithm for automatically extracting and identifying photovoltaic abnormal faults according to claim 3 is characterized in that: In step S3: Analyze from the time dimension. Different time dimensions correspond to different operation and maintenance priorities. From the perspective of link dimensions, different link dimensions use different system efficiency coefficients K when calculating the predicted power generation of power plant design; From the perspective of structural dimension analysis, different structural dimensions use different design information data when calculating the predicted power generation of power plant operation, and the calculated total radiation amount is also different.

5. The algorithm for automatically extracting and identifying photovoltaic abnormal faults according to claim 4 is characterized in that: In step S3, the comparison includes comparison between month-on-month and year-on-year, and comparison between actual operating parameters and predicted values.

6. The algorithm for automatically extracting and identifying photovoltaic abnormal faults according to claim 5 is characterized in that: In step S3, the photovoltaic power station operation data is analyzed from the time dimension, and the photovoltaic power station fault types are divided into three categories: emergency stop fault, equipment inefficient operation and equipment performance degradation, among which: For emergency stop faults, the characteristics of emergency stop faults are that when the fault occurs, it directly affects the operation or communication of the equipment. The operation and maintenance priority of this type of fault is the highest. It is necessary to analyze the power station operation data in a short period of time to discover the fault in time, which includes: String open circuit fault By comparing the input voltage and current data of each string before and after the preset time, the current string voltage and current values ​​suddenly change to 0, and the voltage values ​​of other strings under the same inverter are greater than the inverter startup voltage, and the current values ​​are greater than 0; Inverter communication failure By setting the input and output power and grid-connected point power of each inverter before and after the preset time, when the inverter communication failure occurs, the inverter works normally, the input and output power of the inverter suddenly drops to 0, and the total grid-connected power of the grid-connected point where this inverter is located is greater than the sum of the inverter powers in normal operation; Abnormal shutdown of the inverter By setting the input and output power and grid-connected point power of each inverter before and after the preset time, when the inverter communication failure occurs, the inverter works normally, the input and output power of the inverter suddenly drops to 0, and the total grid-connected power of the grid-connected point where this inverter is located is equal to the sum of the power of the inverters working normally; Grid connection point open circuit fault By comparing the input power of the grid connection point before and after the preset time, when an open circuit fault occurs at the grid connection point, the input power of the grid connection point suddenly drops to 0, and the grid connection power of the grid connection point in the same area or the same power station is greater than 0; For inefficient equipment operation, the characteristics of inefficient equipment operation failure are that when the failure occurs, it will not affect the operation of the photovoltaic power station, but will reduce the operating efficiency of the photovoltaic power station. The operation and maintenance priority of this type of failure is higher, and it is necessary to analyze the power station operation data within a week or even a month to diagnose and troubleshoot the failure, which includes; Hot spot phenomenon of photovoltaic modules The detection of hot spot phenomenon of photovoltaic modules requires comparative analysis of meteorological conditions and input voltage and current values ​​of the strings. The irradiation intensity data obtained by the meteorological instrument is used to determine that as the irradiance increases, the output voltage of the photovoltaic strings decreases. The voltage and current of each string are compared, and the voltage and current discrete rates of each string are calculated. The calculation formula is as follows: Where: μ is the average value of each string voltage; σ is the standard deviation of each string voltage; x i is the voltage of the i-th string; If the voltage and current discrete rate of each string is less than 5%, the string is working stably and well. If the voltage and current discrete rate of each string is greater than 5%, the similarity with the irradiance analysis curve at the corresponding moment is further analyzed. By analyzing the power station operation data every day within a month, if the similarity decreases with the increase of irradiation intensity, there are hot spots in the photovoltaic components in the string; Photovoltaic panels are shaded By calculating the voltage and current discrete rates of each string, and comparing the voltage discrete rates of the strings in a preset time period within a week, if the voltage discrete rate is greater than 5% and occurs in the same time period every day, it is determined that the PV module is shaded during this time period; Photovoltaic module dust pollution The actual power generation of the photovoltaic strings is lower than the predicted power generation, and the difference is increasing. The output current of the strings is lower than the output current of the strings under similar historical circumstances. Equipment performance degradation includes: Inverter inefficient operation According to the inverter input data and inverter output data, the actual efficiency of the inverters of different power stations in the same area is calculated. According to the power station design information and historical meteorological data, the normal efficiency thresholds of different power stations in the local area are calculated. The inverter loss of the photovoltaic power station is evaluated through data comparison and analysis. The actual efficiency calculation formula of the inverter is: Where: E pni Input power to the inverter; E pno Input power to the inverter; Calculate the working efficiency of the inverter, and analyze the working efficiency after averaging, η n The normal range is 95% to 98.5%; Grid connection loss is too high Calculate the actual grid-connected loss of the power station based on the online power generation and inverter output data, and calculate the normal grid-connected loss threshold of the power station based on the power station design information. Evaluate the grid-connected loss of the photovoltaic power station through data comparison and analysis, determine whether there is a fault in the grid-connected link of the power station, and issue early warnings for power stations with high grid-connected losses; Where: η b The grid connection loss of the power station; E pb The amount of electricity generated by the power station connected to the grid; E pno It is the power generation on the output side of the power station inverter.

Images