A photovoltaic string anomaly risk assessment method and system

Abstract

The application discloses a photovoltaic string abnormal risk assessment method and system, and the method comprises the following steps: based on the output voltage and the output current of each photovoltaic string branch in each evaluation period, the unit capacity output power, the reference unit capacity output power and the output loss rate of each photovoltaic string branch are calculated; according to the output loss rate of the photovoltaic string branch in the historical reference window and the recent judgment window, the stage loss difference and the loss change in the recent window of the photovoltaic string branch are calculated; the number of times of repeated output loss of the photovoltaic string branch in the same fixed time period of continuous multiple effective power generation days is counted; the abnormal risk pre-judgment index of the photovoltaic string branch is constructed, and the abnormal risk grade and the risk tendency of the photovoltaic string are judged. The application comprehensively analyzes the stage change, the recent change and the fixed time period repetition characteristics of the output loss of the string, realizes the pre-judgment of the abnormal risk degree and the abnormal risk tendency of the string, and provides a reference for operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power generation technology, and in particular to a method and system for assessing abnormal risks in photovoltaic strings. Background Technology

[0002] Small-scale photovoltaic (PV) power plants typically include multiple PV string branches, intelligent combiner boxes, inverters, and monitoring equipment. They are widely used in rooftop distributed PV, PV carports, and industrial park distributed PV applications. Due to factors such as partial shading, dust accumulation, bird droppings, module degradation, loose wiring, and string open circuits, some PV strings may experience a decrease in output power, necessitating monitoring of the PV strings in the PV power plant.

[0003] Existing photovoltaic power plant monitoring methods mostly rely on inverter output power, system power generation, or grid-connected parameters as the basis for judgment. While these methods can reflect the overall operating status of the power plant, their ability to identify early anomalies in individual string branches is limited. Some existing methods determine whether a string is abnormal by measuring string voltage, current, or output power, but these primarily focus on whether the current output is below a preset threshold. They struggle to simultaneously reflect changes in string output loss relative to historical conditions, recent growth trends, and repetitive characteristics within the same time period. When string output has not yet reached a significant fault level, but output loss continues to increase or recurs repeatedly within a fixed time period, timely risk assessment may not be possible. Summary of the Invention

[0004] Purpose of the invention: The present invention aims to provide a photovoltaic string anomaly risk assessment method and system that can comprehensively analyze the stage changes, recent changes and repetitive characteristics of string output loss, and make a preliminary judgment on the degree and tendency of string anomaly risk, so as to provide a reference for operation and maintenance.

[0005] Technical solution: The present invention provides a method for assessing the abnormal risk of photovoltaic strings, comprising:

[0006] The power generation period of the photovoltaic power station is divided into continuous evaluation periods, and the output voltage and output current of each photovoltaic string branch are collected in each evaluation period.

[0007] Based on the output voltage, output current and installed capacity of each photovoltaic string branch, calculate the unit capacity output power, reference unit capacity output power and output loss rate of each photovoltaic string branch in the corresponding evaluation period.

[0008] Starting from the most recent valid power generation day corresponding to the assessment time, consecutive selections are made backwards. The effective power generation days are the recent judgment window, and the calculation is based on the effective power generation days preceding the start date of the recent judgment window, taking backwards. Using the effective power generation day as the historical benchmark window, the stage loss difference and recent loss change of the photovoltaic string branch are calculated based on the output loss rate of the photovoltaic string branch within the two windows.

[0009] The daily power generation period is divided into multiple fixed time periods. The number of times the output loss of the photovoltaic string branch occurs repeatedly in the same fixed time period on multiple consecutive effective power generation days is counted to determine the number of times the output loss is repeated in the same time period.

[0010] An abnormal risk prediction index for photovoltaic string branches is constructed based on the stage loss difference, recent loss change within the window, and the number of repetitions within the same time period to determine the abnormal risk level and abnormal risk tendency of the photovoltaic string.

[0011] Preferably, the unit capacity output power of the photovoltaic string branch in each evaluation period is the ratio of the product of the output voltage and output current of the photovoltaic string branch in the corresponding evaluation period to the installed capacity of the photovoltaic string branch.

[0012] Preferably, the reference unit capacity output power is the average value of the unit capacity output power of each photovoltaic string branch within the same evaluation period after removing the minimum and maximum values.

[0013] Preferably, the calculation process for the output loss rate of the photovoltaic string branch in each evaluation period includes:

[0014]

[0015] in, Let be the output loss rate of the i-th photovoltaic string branch during the t-th evaluation period.

[0016] Preferably, the calculation process for the stage loss difference of the photovoltaic string branch includes:

[0017]

[0018] In the formula, Let be the stage loss difference of the i-th photovoltaic string branch; This represents the average output loss rate within the recent judgment window. The average output loss rate within the historical benchmark window;

[0019] The average output loss rate within the recent judgment window and the average output loss rate within the historical benchmark window are both calculated based on the daily average output loss rate of the photovoltaic string. The daily average output loss rate of the i-th photovoltaic string branch on the d-th effective power generation day is calculated according to the following formula:

[0020]

[0021] In the formula, For the first The average daily output loss rate of a photovoltaic string branch during the dth effective power generation day; Let be the output loss rate of the i-th photovoltaic string branch during the t-th effective evaluation period on the d-th effective power generation day; This represents the number of valid assessment cycles within the d-th valid power generation day.

[0022] Preferably, the change in loss within the recent window is the difference between the average daily output loss rate of the most recent effective power generation day in the recent judgment window and the average daily output loss rate of the first effective power generation day in the recent judgment window. When the difference is less than 0, the change in loss within the recent window is taken as 0.

[0023] Preferably, the average output loss rate of the i-th photovoltaic string branch during the j-th fixed time period on the d-th effective power generation day is calculated and compared with a preset loss judgment threshold; when the preset loss judgment threshold is exceeded, it is determined that the photovoltaic string branch has an output loss during the corresponding fixed time period on the corresponding effective power generation day.

[0024] Preferably, the abnormal risk prediction index is calculated according to the following formula:

[0025]

[0026] In the formula, The abnormal risk prediction index is the index for the i-th photovoltaic string branch. Let be the stage loss difference of the i-th photovoltaic string branch; This serves as a reference threshold for the difference in stage losses. Let represent the change in loss within the recent window for the i-th photovoltaic string branch; This serves as a reference threshold for recent changes in losses within the window. The number of times the output loss of the i-th photovoltaic string branch is repeated. Output a reference threshold for the number of repeated losses within the same time period; , and These are the weighting coefficients corresponding to the stage loss difference, the recent change in loss within the window, and the number of times the loss is repeated within the same time period.

[0027] Preferably, the determination of the abnormal risk level and abnormal risk tendency of the photovoltaic string includes:

[0028] When the abnormal risk prediction index is lower than the first prediction threshold, low abnormal risk is output.

[0029] When the abnormal risk prediction index is not lower than the first prediction threshold and is lower than the second prediction threshold, the abnormal risk is output.

[0030] When the abnormal risk prediction index is not lower than the second prediction threshold, a high abnormal risk is output.

[0031] When the number of repetitions in the same time period reaches the reference threshold for the number of repetitions of output loss in the same time period, the risk tendency of fixed time period occlusion is output.

[0032] When the stage loss difference reaches the stage loss difference reference threshold, and the loss change in the recent window is lower than the loss change in the recent window reference threshold, output the risk tendency of dust accumulation or component decay.

[0033] When the change in loss within the recent window reaches the reference threshold for the change in loss within the recent window, the string sudden abnormal risk tendency is output.

[0034] The photovoltaic string anomaly risk assessment system of the present invention includes:

[0035] The data acquisition unit is used to collect the output voltage and output current of each photovoltaic string branch according to a preset evaluation cycle;

[0036] The data processing unit is used to calculate the unit capacity output power, reference unit capacity output power and output loss rate of each photovoltaic string branch within the corresponding evaluation period based on the output voltage, output current and installed capacity of each photovoltaic string branch.

[0037] The pre-judgment analysis unit is used to divide the recent judgment window and the historical benchmark window. Based on the output loss rate of the photovoltaic string branch in the two windows, it calculates the stage loss difference of the photovoltaic string branch, the loss change in the recent window, and the number of times the output loss of the photovoltaic string branch repeatedly occurs in the same fixed time period of multiple consecutive effective power generation days, and constructs the abnormal risk pre-judgment index.

[0038] The result output unit is used to output the abnormal risk level and abnormal risk tendency of the photovoltaic string based on the abnormal risk pre-judgment index, combined with the stage loss difference of the photovoltaic string branches, the loss change in the recent window, and the number of repetitions in the same time period.

[0039] Beneficial Effects: Compared with the prior art, the present invention has the following significant advantages: 1. The present invention constructs reference values ​​based on the unit capacity output power of multiple photovoltaic string branches within the same evaluation period, which can reduce the impact of string installed capacity differences and overall weather changes on the judgment results, and is conducive to identifying the output loss of a single string relative to other strings; 2. The present invention constructs the stage loss difference and the loss change within the window through historical benchmark windows and recent judgment windows, and combines the number of repetitions of output loss within a fixed time period to comprehensively analyze the stage rise, recent abrupt change and fixed-time repetition characteristics of string output loss, thereby realizing string anomaly risk. 3. This invention normalizes and weights the stage loss difference, the loss change within the window, and the number of repetitions in the same time period to obtain an abnormal risk pre-judgment index, which can output the abnormal risk level and further provide risk tendencies such as fixed-time shading, dust accumulation or component attenuation, and sudden string anomalies, providing a reference for operation and maintenance; 4. This invention can be implemented directly using the string output voltage and output current data collected by the intelligent combiner box, without the need for additional image recognition equipment, infrared detection equipment, or complex environmental sensors, making it easy to apply in photovoltaic power stations with multiple string branches, such as rooftop distributed photovoltaic, photovoltaic carports, and park distributed photovoltaic. Attached Figure Description

[0040] Figure 1 This is a flowchart of the method of the present invention;

[0041] Figure 2 This is a schematic diagram of the system structure of the present invention;

[0042] Figure 3 This is a graph showing the variation of the daily average output loss rate under five operating conditions in this embodiment of the invention.

[0043] Figure 4 for Figure 3 A magnified view of a portion of the image;

[0044] Figure 5 This is a comparison chart of the abnormal risk prediction index under five operating conditions in this embodiment of the invention. Detailed Implementation

[0045] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0046] like Figure 1 As shown, the photovoltaic string anomaly risk assessment method of the present invention includes the following steps:

[0047] (1) Data acquisition and preprocessing

[0048] The power generation period of the photovoltaic power station is divided into continuous evaluation cycles according to a preset duration. Output voltage and current data for each photovoltaic string branch are acquired within each evaluation cycle, and the installed capacity of each photovoltaic string branch is also obtained. The evaluation cycle has a preset duration. When multiple sets of output voltage and current data for a photovoltaic string branch are acquired within an evaluation cycle, the average value of these multiple sets of data is calculated and used as the output voltage and output current for that photovoltaic string branch in the corresponding evaluation cycle. When only one set of output voltage and current data for a photovoltaic string branch is acquired within an evaluation cycle, this set of data is used as the output voltage and output current for that photovoltaic string branch in the corresponding evaluation cycle.

[0049] (2) Based on the output voltage, output current, and installed capacity of each photovoltaic string branch, calculate the unit capacity output power and reference unit capacity output power of each photovoltaic string branch within the corresponding evaluation period. The unit capacity output power is calculated according to the following formula:

[0050]

[0051] in, Let be the unit capacity output power of the i-th photovoltaic string branch in the t-th evaluation period; Let be the installed capacity of the i-th photovoltaic string branch; These are the output voltage and output current of the i-th photovoltaic string branch during the t-th evaluation period, respectively.

[0052] The unit capacity output power of each photovoltaic string branch within the same evaluation period Sort by power output per unit capacity from smallest to largest, remove the smallest and largest unit capacity output power, and use the average value of the remaining unit capacity output power as the reference unit capacity output power. .

[0053] (3) Calculate the output loss rate of each photovoltaic string branch in each evaluation period based on the unit capacity output power of each photovoltaic string branch and the reference unit capacity output power. The output loss rate is calculated according to the following formula:

[0054]

[0055] in, Let be the output loss rate of the i-th photovoltaic string branch in the t-th evaluation period. When the unit capacity output power of the photovoltaic string branch is not lower than the reference unit capacity output power, the output loss rate of the corresponding effective evaluation period is recorded as 0.

[0056] (4) Set a historical benchmark window and a recent judgment window from the effective power generation days, and determine the stage loss difference and the loss change in the recent window based on the output loss rate of the photovoltaic string branch in the two windows.

[0057] A valid power generation day is defined as a day in which the photovoltaic power station is operating during the specified power generation period, all photovoltaic string branch operation data are collected normally, and the number of cycles meeting the evaluation requirements reaches the pre-set standard. If the photovoltaic power station is shut down, data acquisition communication is interrupted, or string operation data is missing, resulting in the number of valid evaluation cycles falling below the preset number, that date is not considered a valid power generation day. The recent judgment window refers to selecting consecutively backwards from the most recent valid power generation day corresponding to this evaluation. The recent judgment window must contain at least two valid power generation days; and the historical benchmark window refers to the window that takes the period backward from the previous valid power generation day of the recent judgment window. Effective power generation days.

[0058] The calculation process for the stage loss difference and the change in loss within the window includes:

[0059] The average daily output loss rate of the i-th photovoltaic string branch during the d-th effective power generation day is calculated according to the following formula:

[0060]

[0061] In the formula, For the first The average daily output loss rate of a photovoltaic string branch during the dth effective power generation day; Let be the output loss rate of the i-th photovoltaic string branch during the t-th effective evaluation period on the d-th effective power generation day; This represents the number of valid assessment cycles within the d-th valid power generation day.

[0062] The average output loss rate of the i-th photovoltaic string branch within the historical benchmark window Calculate according to the following formula:

[0063]

[0064] The average output loss rate of the i-th photovoltaic string branch within the recent judgment window Calculate according to the following formula:

[0065]

[0066] In the formula, Let be the average daily output loss rate of the i-th photovoltaic string branch on the d-th effective power generation day within the historical benchmark window; Let be the average daily output loss rate of the i-th photovoltaic string branch on the d-th effective power generation day within the recent judgment window.

[0067] The difference in losses during the aforementioned stages is calculated using the following formula:

[0068]

[0069] In the formula, Let be the stage loss difference of the i-th photovoltaic string branch.

[0070] The average daily output loss rate of the photovoltaic string branch on the last effective power generation day of the recent judgment window is compared with the average daily output loss rate on the first effective power generation day of the recent judgment window to obtain the change in loss within the recent window; the change in loss within the recent window is calculated according to the following formula:

[0071]

[0072] In the formula, Let represent the change in loss within the recent window for the i-th photovoltaic string branch; The average daily output loss rate is the first effective power generation day within the recent judgment window. The average daily output loss rate is the last valid power generation day of the recent judgment window.

[0073] (5) Divide the daily preset power generation period into multiple fixed time periods, count the number of times the output loss of the photovoltaic string branch occurs repeatedly in the same fixed time period of multiple consecutive effective power generation days, and determine the number of times the output loss is repeated in the same time period.

[0074] The fixed time period consists of multiple consecutive evaluation cycles. The daily preset power generation period is divided into s fixed time periods. The average output loss rate of the i-th photovoltaic string branch within the j-th fixed time period on the d-th effective power generation day is calculated and compared with a preset loss judgment threshold. When the preset loss judgment threshold is exceeded, it is determined that the target photovoltaic string branch has an output loss within the corresponding fixed time period on the corresponding effective power generation day, and is marked according to the following formula:

[0075]

[0076] In the formula, Let be the average output loss rate of the i-th photovoltaic string branch during the j-th fixed time period on the d-th effective power generation day; The preset loss judgment threshold is set.

[0077] The number of times the output loss is repeated within the j-th fixed time period is calculated according to the following formula:

[0078]

[0079] The number of repetitions within the same time period is determined according to the following formula:

[0080] In the formula, D represents the number of consecutive valid power generation days included in the count of repeated occurrences; The number of times the output loss is repeated within the j-th fixed time period; Output the number of times the loss is repeated in the same time period for the i-th photovoltaic string branch.

[0081] (6) Normalize the stage loss difference, the recent loss change in the window and the number of times the output loss is repeated in the same time period respectively, and calculate the abnormal risk prediction index of the photovoltaic string branch according to the corresponding weight coefficient.

[0082] The abnormal risk prediction index is calculated according to the following formula:

[0083]

[0084] In the formula, The abnormal risk prediction index is the index for the i-th photovoltaic string branch. Let be the stage loss difference of the i-th photovoltaic string branch; This serves as a reference threshold for the difference in stage losses. Let represent the change in loss within the recent window for the i-th photovoltaic string branch; This serves as a reference threshold for recent changes in losses within the window. The number of times the output loss of the i-th photovoltaic string branch is repeated. Output a reference threshold for the number of repeated losses within the same time period; , and The weighting coefficients are respectively for the stage loss difference, the recent loss change within the window, and the number of times the loss is repeated within the same time period, and satisfy the following:

[0085] .

[0086] The abnormal risk level of the target photovoltaic string branch is output based on the abnormal risk prediction index. Based on the characteristics of the stage loss difference, the loss change in the recent window, and the number of times the loss is repeated in the same time period, the abnormal risk level and abnormal risk tendency of the target photovoltaic string branch are output.

[0087] When the abnormal risk prediction index is lower than the first prediction threshold, low abnormal risk is output.

[0088] When the abnormal risk prediction index is not lower than the first prediction threshold and is lower than the second prediction threshold, the abnormal risk is output.

[0089] When the abnormal risk prediction index is not lower than the second prediction threshold, a high abnormal risk is output.

[0090] When the number of repetitions in the same time period reaches the reference threshold for the number of repetitions of loss in the same time period, i.e., the preset repetition threshold, the fixed time period occlusion risk tendency is output.

[0091] When the stage loss difference reaches the stage loss difference reference threshold, i.e. the preset stage loss threshold, and the loss change in the recent window is lower than the loss change in the recent window reference threshold, i.e. the preset mutation threshold, the risk tendency of dust accumulation or component decay is output.

[0092] When the change in loss within the recent window reaches the reference threshold for the change in loss within the recent window, i.e., the preset mutation threshold, the string sudden abnormal risk tendency is output.

[0093] When the target photovoltaic string branch simultaneously meets the judgment conditions of two or more abnormal risk tendencies, the corresponding two or more abnormal risk tendencies are output.

[0094] Based on the same inventive concept, the photovoltaic string anomaly risk assessment system of the present invention includes:

[0095] The data acquisition unit includes a data acquisition gateway and a string information configuration unit. The data acquisition gateway is communicatively connected to the smart combiner box of the photovoltaic power station and is used to read the output voltage and output current data of the photovoltaic string branches corresponding to each input channel of the smart combiner box according to a preset evaluation cycle. The string information configuration unit is used to store the string number, installed capacity, and correspondence between each photovoltaic string branch and the input channel of the smart combiner box.

[0096] The data processing unit is connected to the data acquisition gateway and the string information configuration unit, respectively, and is used to calculate the unit capacity output power, reference unit capacity output power and output loss rate of each photovoltaic string branch;

[0097] The pre-judgment analysis unit, connected to the data processing unit, is used to determine the effective power generation day, set the historical benchmark window and the recent judgment window, calculate the stage loss difference of the photovoltaic string branch, the loss change within the recent window and the number of times the loss output is repeated in the same time period, and construct the abnormal risk pre-judgment index.

[0098] The result output unit, connected to the pre-judgment analysis unit, is used to output the abnormal risk level of the target photovoltaic string branch according to the abnormal risk pre-judgment index, and to output the abnormal risk tendency according to the stage loss difference, the loss change in the recent window, and the number of times the loss output is repeated in the same time period.

[0099] To further verify the method of this invention, this embodiment uses MATLAB to construct a photovoltaic array operation data simulation model. The simulation object includes 10 photovoltaic string branches, with the installed capacity of each string preset and varying slightly within the range of 5.30 to 5.50 kW. The third photovoltaic string branch is used as the target string, and the remaining strings are used together to calculate the reference unit capacity output power.

[0100] The power generation period is set from 08:00 to 17:00 daily, and continuous evaluation cycles are divided into 5-minute intervals. For each evaluation cycle, the unit capacity output power of each photovoltaic string branch is calculated. The unit capacity output power of each string within the same evaluation cycle is sorted from smallest to largest. After removing the maximum and minimum values, the average of the remaining values ​​is calculated to obtain the reference unit capacity output power. Based on the difference between the target string's unit capacity output power and the reference unit capacity output power, the output loss rate of the target string in each evaluation cycle is calculated.

[0101] Five operating conditions are set: normal operation, overall weather change, fixed period of shading, slow dust accumulation or component degradation, and sudden string abnormality.

[0102] Under normal operating conditions, each photovoltaic string branch operates according to a similar intraday power curve, with slight string differences and random measurement fluctuations superimposed.

[0103] Under overall weather change conditions, the power reduction caused by cloud changes is synchronously added to each photovoltaic string branch to simulate the common change in irradiance.

[0104] Under fixed-period shading conditions, the unit capacity output power of the target string will decrease by 18% to 22% during the period from 10:00 to 11:00 each day in the last 8 consecutive power generation days.

[0105] Under conditions of slow dust accumulation or component degradation, starting from the 13th power generation day, the output power per unit capacity of the target string will decrease daily, with a maximum decrease of approximately 16%.

[0106] In the event of a sudden abnormal operating condition of the string, the output power per unit capacity of the target string suddenly drops by 30% to 38% on the last day of power generation.

[0107] For each operating condition, 30 sets of random samples were generated, resulting in a total of 150 sets of simulation samples. The overall daytime irradiance level, inherent output differences of the string, measurement noise, and anomaly degree were randomly varied among the different samples.

[0108] In this invention, a grid search method based on MATLAB simulation is used to traverse the candidate combinations of historical baseline window length, recent judgment window length, various reference thresholds, weight coefficients and risk level thresholds. The average recognition rate of low, medium and high risk levels, the detection rate of abnormal working conditions and the accuracy rate of normal working conditions are used as evaluation indicators to select the parameter combination with the best comprehensive evaluation result.

[0109] This embodiment obtains the following results by iterating and comparing the historical benchmark window length, recent judgment window length, fixed time period loss judgment threshold, reference thresholds for each indicator, weighting coefficients, and risk level thresholds:

[0110]

[0111]

[0112] The weighting coefficients for the stage loss difference, the change in loss within the window, and the number of times the loss output is repeated within the same time period are as follows:

[0113]

[0114] Therefore, the abnormal risk prediction index in this embodiment is:

[0115]

[0116] The first and second pre-judgment thresholds are set as follows:

[0117]

[0118] when When, output low anomaly risk; when At that time, there is an abnormal risk in the output; when At that time, the output shows a high risk of anomalies.

[0119] when When, output the risk tendency of occlusion during a fixed period; when and At that time, the output is prone to dust accumulation or component degradation risk; when When the condition is met, output the sudden abnormal risk tendency of the string. If two or more judgment conditions are met simultaneously, output the corresponding risk tendency at the same time.

[0120] The simulation results for the five operating conditions are shown in the table below.

[0121]

[0122] like Figure 3-4 As shown, under normal operating conditions, the daily average output loss rate of the target string remains close to 0; under overall weather change conditions, the daily average output loss rate of the target string fluctuates smoothly, with small differences in losses during periods and small changes in losses within recent windows; under fixed-period shading conditions, the daily average output loss rate increases slightly after an anomaly occurs and remains stable; under slow dust accumulation or module degradation conditions, the daily average output loss rate gradually increases with the number of effective power generation days; under sudden string anomaly conditions, the daily average output loss rate increases significantly on the last effective power generation day.

[0123] like Figure 5 As shown, the abnormal risk prediction index for normal operation and overall weather change conditions is lower than the first prediction threshold. The abnormal risk prediction index for fixed-period shading and slow dust accumulation or component degradation conditions is in the medium-risk range. The abnormal risk prediction index for sudden string abnormal conditions reaches the second prediction threshold and is judged as high-risk.

[0124] Under fixed-period occlusion conditions, the number of repetitions within the same time period The change in loss within the window is relatively large. A smaller value indicates that the target string has experienced recurring losses over multiple consecutive days within the same time period. Under conditions of slow dust accumulation or component degradation, the difference in stage losses... The change in loss within the window is relatively large. The relatively small value indicates that the overall output loss level of the target string has increased significantly recently compared to the historical benchmark, but the change is relatively slow in the short term. The difference in stage loss under sudden string abnormal operating conditions. and the change in loss within the window All reached the corresponding reference threshold, and the number of repetitions within the same time period was relatively small, indicating that the anomaly was obviously sudden.

[0125] Within the 150 sets of simulation samples set in this embodiment, the identification rate for low-risk, medium-risk, and high-risk conditions was 100%. Normal and overall weather change conditions were all classified as low-risk, and all three types of abnormal conditions were detected. Slow dust accumulation or component degradation conditions may result in output loss for several consecutive days over multiple fixed time periods, and may also be accompanied by a risk of output shading during fixed time periods; this result represents multiple risk tendencies in parallel output and does not affect the risk level determination.

[0126] Simulation results show that the present invention can reflect the phased increase, recent abrupt change and fixed period repetition characteristics of the output loss of the target photovoltaic string by using the stage loss difference, the loss change within the recent window and the number of times the loss output is repeated in the same time period. By normalizing and weighting the three indicators, it can distinguish between normal operation, overall weather changes and abnormal risks of different strings.

Claims

1. A method for assessing the abnormal risk of photovoltaic strings, characterized in that, include: The power generation period of the photovoltaic power station is divided into continuous evaluation periods, and the output voltage and output current of each photovoltaic string branch are collected in each evaluation period. Based on the output voltage, output current and installed capacity of each photovoltaic string branch, calculate the unit capacity output power, reference unit capacity output power and output loss rate of each photovoltaic string branch in the corresponding evaluation period. Starting from the most recent effective power generation day corresponding to the assessment time, several consecutive effective power generation days are selected backward as the recent judgment window. From the effective power generation day before the start date of the recent judgment window, several effective power generation days are selected backward as the historical benchmark window. Based on the output loss rate of the photovoltaic string branch in the two windows, the stage loss difference of the photovoltaic string branch and the loss change in the recent window are calculated. The daily power generation period is divided into multiple fixed time periods. The number of times the output loss of the photovoltaic string branch occurs repeatedly in the same fixed time period on multiple consecutive effective power generation days is counted to determine the number of times the output loss is repeated in the same time period. An abnormal risk prediction index for photovoltaic string branches is constructed based on the stage loss difference, recent loss change within the window, and the number of repetitions within the same time period to determine the abnormal risk level and abnormal risk tendency of the photovoltaic string.

2. The photovoltaic string anomaly risk assessment method according to claim 1, characterized in that, The unit capacity output power of a photovoltaic string branch in each evaluation period is the ratio of the product of the output voltage and output current of the photovoltaic string branch in the corresponding evaluation period to the installed capacity of the photovoltaic string branch.

3. The photovoltaic string anomaly risk assessment method according to claim 2, characterized in that, The reference unit capacity output power is the average value of the unit capacity output power of each photovoltaic string branch within the same evaluation period after removing the minimum and maximum values.

4. The photovoltaic string anomaly risk assessment method according to claim 1, characterized in that, The calculation process for the output loss rate of the photovoltaic string branch in each evaluation period includes: in, Let be the output loss rate of the i-th photovoltaic string branch during the t-th evaluation period.

5. The photovoltaic string anomaly risk assessment method according to claim 1, characterized in that, The calculation process for the stage loss difference of the photovoltaic string branch includes: In the formula, Let be the stage loss difference of the i-th photovoltaic string branch; This represents the average output loss rate within the recent judgment window. The average output loss rate within the historical benchmark window; The average output loss rate within the recent judgment window and the average output loss rate within the historical benchmark window are both calculated based on the daily average output loss rate of the photovoltaic string. The daily average output loss rate of the i-th photovoltaic string branch on the d-th effective power generation day is calculated according to the following formula: In the formula, For the first The average daily output loss rate of a photovoltaic string branch during the dth effective power generation day; Let be the output loss rate of the i-th photovoltaic string branch during the t-th effective evaluation period on the d-th effective power generation day; This represents the number of valid assessment cycles within the d-th valid power generation day.

6. The photovoltaic string anomaly risk assessment method according to claim 1, characterized in that, The change in loss within the recent window is the difference between the average daily output loss rate of the most recent effective power generation day in the recent judgment window and the average daily output loss rate of the first effective power generation day in the recent judgment window. When the difference is less than 0, the change in loss within the recent window is taken as 0.

7. The photovoltaic string anomaly risk assessment method according to claim 1, characterized in that, Calculate the average output loss rate of the i-th photovoltaic string branch during the j-th fixed time period on the d-th effective power generation day, and compare it with the preset loss judgment threshold; when it exceeds the preset loss judgment threshold, it is determined that the photovoltaic string branch has an output loss during the corresponding fixed time period on the corresponding effective power generation day.

8. The photovoltaic string anomaly risk assessment method according to claim 7, characterized in that, The abnormal risk prediction index is calculated according to the following formula: In the formula, The abnormal risk prediction index is the index for the i-th photovoltaic string branch. Let be the stage loss difference of the i-th photovoltaic string branch; This serves as a reference threshold for the difference in stage losses. Let represent the change in loss within the recent window for the i-th photovoltaic string branch; This serves as a reference threshold for recent changes in losses within the window. The number of times the output loss of the i-th photovoltaic string branch is repeated. Output a reference threshold for the number of repeated losses within the same time period; , and These are the weighting coefficients corresponding to the stage loss difference, the recent change in loss within the window, and the number of times the loss is repeated within the same time period.

9. The photovoltaic string anomaly risk assessment method according to claim 8, characterized in that, The determination of the abnormal risk level and abnormal risk tendency of photovoltaic strings includes: When the abnormal risk prediction index is lower than the first prediction threshold, low abnormal risk is output. When the abnormal risk prediction index is not lower than the first prediction threshold and is lower than the second prediction threshold, the abnormal risk is output. When the abnormal risk prediction index is not lower than the second prediction threshold, a high abnormal risk is output. When the number of repetitions in the same time period reaches the reference threshold for the number of repetitions of output loss in the same time period, the risk tendency of fixed time period occlusion is output. When the stage loss difference reaches the stage loss difference reference threshold, and the loss change in the recent window is lower than the loss change in the recent window reference threshold, output the risk tendency of dust accumulation or component decay. When the change in loss within the recent window reaches the reference threshold for the change in loss within the recent window, the string sudden abnormal risk tendency is output.

10. A photovoltaic string anomaly risk assessment system, characterized in that, include: The data acquisition unit is used to collect the output voltage and output current of each photovoltaic string branch according to a preset evaluation cycle; The data processing unit is used to calculate the unit capacity output power, reference unit capacity output power and output loss rate of each photovoltaic string branch within the corresponding evaluation period based on the output voltage, output current and installed capacity of each photovoltaic string branch. The pre-judgment analysis unit is used to divide the recent judgment window and the historical benchmark window. Based on the output loss rate of the photovoltaic string branch in the two windows, it calculates the stage loss difference of the photovoltaic string branch, the loss change in the recent window, and the number of times the output loss of the photovoltaic string branch repeatedly occurs in the same fixed time period of multiple consecutive effective power generation days, and constructs the abnormal risk pre-judgment index. The result output unit is used to output the abnormal risk level and abnormal risk tendency of the photovoltaic string based on the abnormal risk pre-judgment index, combined with the stage loss difference of the photovoltaic string branches, the loss change in the recent window, and the number of repetitions in the same time period.

Images