Method and apparatus for low efficiency detection of photovoltaic devices
By constructing a photovoltaic equipment inefficiency detection model based on SCADA data, the problem of inefficient identification of inverters, combiner boxes and strings in photovoltaic power generation systems has been solved, achieving accurate equipment detection and improving operation and maintenance efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- BEIJING JINFENG HUINENG TECH CO LTD
- Filing Date
- 2024-12-20
- Publication Date
- 2026-06-23
Smart Images
Figure CN122268273A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of photovoltaic power generation technology, and more specifically, to a method and apparatus for detecting inefficiencies in photovoltaic equipment. Background Technology
[0002] Solar energy, as a renewable new energy source, generates electricity through photovoltaics (PV). Sunlight shines directly onto the solar cells, generating direct current (DC) in the photovoltaic array via the photoelectric effect. This DC is then converted into usable alternating current (AC) by an inverter, and subsequently fed into the grid or directly supplied to loads. PV power plants are often installed in harsh environments, and various environmental factors lead to frequent malfunctions, reducing the power generation efficiency of the system and hindering stable, continuous, and efficient power generation. On one hand, inefficiency in PV module operation results in reduced power generation and economic losses; on the other hand, factors such as hot spots, shading, and dust accumulation in the PV string cause module inefficiency, affecting the energy conversion efficiency of the PV power generation system. Monitoring the operating status of PV modules helps to promptly identify inefficient modules for repair or replacement, provides data analysis and guidance support for on-site maintenance personnel, and offers theoretical support for unmanned operation modes at PV sites. Summary of the Invention
[0003] To address the aforementioned issues, this disclosure proposes a method and apparatus for detecting inefficiencies in photovoltaic equipment, a computing system, and a computer-readable storage medium.
[0004] According to one aspect of this disclosure, a method for detecting the inefficiency of a photovoltaic (PV) device is provided. The method includes: acquiring operating data of the PV device; determining, based on the acquired operating data, whether the inverter of the PV device is power-limited; in response to the inverter not being power-limited, determining, using an inverter inefficiency prediction model, whether the inverter exhibits inefficiency characteristics; and in response to determining that the inverter does not exhibit inefficiency characteristics, determining whether the combiner box or PV string connected to the inverter exhibits inefficiency characteristics.
[0005] Optionally, the operating data includes the actual active power of the inverter within a predetermined time period, the automatic power generation control limit flag of the photovoltaic power station, and the planned active power of the inverter.
[0006] Optionally, the step of determining whether the inverter of the photovoltaic equipment is power-limited based on the acquired operating data includes: determining that the inverter is power-limited in response to the difference between the planned active power and the actual active power of the inverter being less than a first predetermined threshold and the difference between the actual active power and the load capacity of the inverter being greater than a second predetermined threshold; and / or determining that the inverter is power-limited in response to the automatic generation control power limitation flag indicating that all inverters included in the photovoltaic power station are in a power-limited state; and / or determining that the inverter is power-limited in response to the load factor of the inverter subarray including the inverter being less than a third predetermined threshold and the power of all inverters included in the inverter subarray being greater than a fourth predetermined threshold.
[0007] Optionally, the first predetermined threshold and the second predetermined threshold are products of a predetermined percentage and the load capacity of the inverter, and the third predetermined threshold is a product of another predetermined percentage and the maximum load rate among all inverter subarrays.
[0008] Optionally, the load factor of the inverter subarray is the average of the ratio of the active power to the load capacity of each inverter in the inverter subarray, or the ratio of the sum of the active power of all inverters in the inverter subarray to the sum of the load capacity of all inverters in the inverter subarray.
[0009] Optionally, in response to determining that the inverter does not exhibit inefficiency characteristics, the step of determining whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics includes: in response to determining that the inverter does not exhibit inefficiency characteristics, using a combiner box inefficiency prediction model to determine whether the combiner box connected to the inverter exhibits inefficiency characteristics; and in response to determining that the combiner box does not exhibit inefficiency characteristics, using a string inefficiency prediction model to determine whether the photovoltaic string connected to the combiner box that does not exhibit inefficiency characteristics exhibits inefficiency characteristics.
[0010] Optionally, in response to determining that the inverter does not exhibit inefficiency characteristics, the step of determining whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics includes: in response to determining that the inverter does not exhibit inefficiency characteristics, using a string inefficiency prediction model to determine whether the photovoltaic string connected to the inverter exhibits inefficiency characteristics.
[0011] According to another aspect of this disclosure, an inefficiency detection device for photovoltaic equipment is provided. The inefficiency detection device for photovoltaic equipment includes: an operating data acquisition unit configured to acquire operating data of the photovoltaic equipment; a power limitation determination unit configured to determine whether the inverter of the photovoltaic equipment is power-limited based on the acquired operating data; an inverter inefficiency prediction unit configured to determine whether the inverter exhibits inefficiency characteristics using an inverter inefficiency prediction model in response to the inverter not being power-limited; and a downstream component inefficiency prediction unit configured to determine whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics in response to the determination that the inverter does not exhibit inefficiency characteristics.
[0012] Optionally, the operating data includes the actual active power of the inverter within a predetermined time period, the automatic power generation control limit flag of the photovoltaic power station, and the planned active power of the inverter.
[0013] Optionally, the power limiting determination unit is configured to: determine that the inverter is power-limited in response to the difference between the planned active power and the actual active power of the inverter being less than a first predetermined threshold and the difference between the actual active power and the load capacity of the inverter being greater than a second predetermined threshold; and / or determine that the inverter is power-limited in response to the photovoltaic power plant automatic generation control power limiting flag indicating that all inverters included in the photovoltaic power plant are in a power-limited state; and / or determine that the inverter is power-limited in response to the load factor of the inverter subarray including the inverter being less than a third predetermined threshold and the power of all inverters included in the inverter subarray being greater than a fourth predetermined threshold.
[0014] Optionally, the first predetermined threshold and the second predetermined threshold are products of a predetermined percentage and the load capacity of the inverter, and the third predetermined threshold is a product of another predetermined percentage and the maximum load rate among all inverter subarrays.
[0015] Optionally, the load factor of the inverter subarray is the average of the ratio of the active power to the load capacity of each inverter in the inverter subarray, or the ratio of the sum of the active power of all inverters in the inverter subarray to the sum of the load capacity of all inverters in the inverter subarray.
[0016] Optionally, the lower-level component inefficiency prediction unit is configured to: in response to determining that the inverter does not exhibit inefficiency characteristics, use a combiner box inefficiency prediction model to determine whether the combiner box connected to the inverter exhibits inefficiency characteristics; and in response to determining that the combiner box does not exhibit inefficiency characteristics, use a string inefficiency prediction model to determine whether the photovoltaic string connected to the combiner box that does not exhibit inefficiency characteristics exhibits inefficiency characteristics.
[0017] Optionally, the lower-level component inefficiency prediction unit is configured to: in response to determining that the inverter does not exhibit inefficiency characteristics, use a string inefficiency prediction model to determine whether the photovoltaic string connected to the inverter exhibits inefficiency characteristics.
[0018] According to another aspect of this disclosure, a computing system is provided that includes at least one computing device and at least one storage device of storage instructions, wherein, when the instructions are executed by the at least one computing device, they cause the at least one computing device to perform the inefficient detection method for photovoltaic devices as described above.
[0019] According to another aspect of this disclosure, a computer-readable storage medium for storing instructions is provided, wherein when the instructions are executed by at least one computing device, the at least one computing device causes the at least one computing device to perform the inefficient detection method for photovoltaic devices as described above.
[0020] By adopting this disclosure, inverters exhibiting inefficiency characteristics can be identified more accurately, and considering the hierarchical relationship between inverters, combiner boxes, and strings in a photovoltaic power plant, inefficient equipment can be precisely located and identified. Attached Figure Description
[0021] The above and / or other objects and advantages of this disclosure will become clearer from the following description of embodiments in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart illustrating an inefficiency detection method for a photovoltaic device according to an exemplary embodiment of the present disclosure; Figure 2 This is a diagram illustrating the composition and hardware architecture of an inefficient detection function module of a photovoltaic device according to an exemplary embodiment of the present disclosure; Figures 3A to 3C This is a flowchart illustrating an example of an inefficient method for a photovoltaic device according to the present disclosure; Figure 4 This is a flowchart illustrating a power limiting determination method according to an embodiment of the present disclosure; Figure 5 This is a diagram illustrating the characteristics of inverter inefficiency; Figure 6 This is a diagram illustrating the characteristics of string inefficiency in string inverters; Figure 7 This is a diagram illustrating the inefficiencies of the combiner box string branches; Figure 8 This is a diagram illustrating the inefficiencies of the combiner box; Figure 9 This is a block diagram illustrating an inefficiency detection device for a photovoltaic device according to an exemplary embodiment of the present disclosure; Figure 10This is a block diagram illustrating a computing system including at least one computing device and at least one storage device of storage instructions according to an exemplary embodiment of the present disclosure. Detailed Implementation
[0022] The following description, in conjunction with the accompanying drawings, provides specific embodiments to aid the reader in gaining a comprehensive understanding of the methods, apparatus, and / or systems described herein. However, upon understanding this disclosure, various changes, modifications, and equivalents of the methods, apparatus, and / or systems described herein will become apparent. For example, the order of operations described herein is merely illustrative and is not limited to those orders set forth herein, but may be altered as will become clear upon understanding this disclosure, except for operations that must occur in a specific order. Furthermore, for clarity and conciseness, descriptions of features known in the art may be omitted.
[0023] A photovoltaic (PV) string refers to the smallest unit in a photovoltaic (PV) power generation system formed by connecting multiple PV modules in series to achieve the required DC output voltage. Some existing solutions use tilt angles and image data to identify inefficient PV strings, but this data is often difficult to obtain or requires additional equipment. Other solutions use SCADA data analysis to identify string inefficiencies, which can identify a few strings with low-circuit inefficiencies, but cannot identify inefficiencies caused by combiner box or inverter problems. Alternatively, existing combiner box inefficiency identification relies on input / output current and temperature data, without considering the acquisition of actual SCADA data or the hierarchical relationship between the inverter, combiner box, and PV string.
[0024] This disclosure provides an inefficient detection method based on SCADA data. It can not only effectively accommodate situations where data such as voltage, power, and temperature are unavailable without the installation of additional devices, but also take into account the hierarchical relationship of equipment in photovoltaic power plants. It can accurately detect detailed information on abnormal conditions of photovoltaic inverters, combiner boxes, and strings. Furthermore, through subsequent data display, it can provide effective theoretical support and judgment basis for on-site operation and maintenance personnel, thereby improving on-site operation and maintenance efficiency.
[0025] Figure 1 This is a flowchart illustrating an inefficiency detection method for a photovoltaic device according to an exemplary embodiment of the present disclosure.
[0026] like Figure 1 As shown, in step S101, the operating data of the photovoltaic equipment is acquired. In this example, the acquired operating data of the photovoltaic equipment includes the actual active power of the inverter within a predetermined time period, the automatic power generation control power limit flag of the photovoltaic power station, and the planned active power of the inverter.
[0027] In step S102, based on the acquired operating data, it is determined whether the inverter of the photovoltaic equipment is power-limited. In the example, the inverter is determined to be power-limited in response to the difference between the planned active power and the actual active power of the inverter being less than a first predetermined threshold and the difference between the actual active power and the load capacity of the inverter being greater than a second predetermined threshold; and / or in response to the photovoltaic power plant automatic generation control power limiting flag indicating that all inverters included in the photovoltaic power plant are in a power-limited state, the inverter is determined to be power-limited; and / or in response to the load rate of the inverter subarray including the inverter being less than a third predetermined threshold and the power of all inverters included in the inverter subarray being greater than a fourth predetermined threshold, the inverter is determined to be power-limited. For example, the first and second predetermined thresholds are the product of a predetermined percentage and the load capacity of the inverter, and the third predetermined threshold is the product of another predetermined percentage and the maximum value of the load rates of all inverter subarrays. In the example, the load factor of the inverter subarray is the average ratio of the active power to the load capacity of each inverter in the inverter subarray, or the ratio of the sum of the active power of all inverters in the inverter subarray to the sum of the load capacity of all inverters in the inverter subarray.
[0028] In step S103, in response to the inverter not being power-limited, an inverter inefficiency prediction model is used to determine whether the inverter exhibits inefficiency characteristics. In step S104, in response to determining that the inverter does not exhibit inefficiency characteristics, it is determined whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics. In this example, in response to determining that the inverter does not exhibit inefficiency characteristics, a combiner box inefficiency prediction model is used to determine whether the combiner box connected to the inverter exhibits inefficiency characteristics; in response to determining that the combiner box does not exhibit inefficiency characteristics, a string inefficiency prediction model is used to determine whether the photovoltaic string connected to the combiner box that does not exhibit inefficiency characteristics exhibits inefficiency characteristics. In another example, in response to determining that the inverter does not exhibit inefficiency characteristics, a string inefficiency prediction model is used to determine whether the photovoltaic string connected to the inverter exhibits inefficiency characteristics.
[0029] The inefficiency detection method for photovoltaic equipment according to the exemplary embodiments of this disclosure can utilize existing SCADA data to detect inefficiencies in photovoltaic inverters, combiner boxes, and strings, providing strong support for on-site operation and maintenance. Furthermore, considering the hierarchical relationship between inverters, combiner boxes, and strings in a photovoltaic power station, it can more accurately locate and identify inefficient equipment.
[0030] Figure 2 This is a diagram illustrating the composition and hardware architecture of an inefficiency detection function module of a photovoltaic device according to an exemplary embodiment of the present disclosure.
[0031] String photovoltaic (PV) power generation systems include inverters and PV strings, while centralized or distributed PV power generation systems include inverters, combiner boxes, and PV strings. Accurately detecting inefficient inverters, combiner boxes, and PV strings allows for timely identification of abnormal equipment and planned maintenance, improving the operational efficiency of PV power plants. Therefore, this disclosure proposes a method and apparatus for detecting inefficient PV equipment based on SCADA data. This method uses data on inverter active power, load capacity, time, and optional irradiance intensity to perform data cleaning and preprocessing, eliminating inverters under power-limited conditions, and constructing an inverter power prediction model. Considering the hierarchy and interrelationships among the equipment in a photovoltaic power plant, for centralized or distributed inverter stations, a machine learning-based combiner box inefficiency detection model can be constructed based on inverter inefficiency identification results, using inverter power, combiner box load capacity, time, etc., and then a string inefficiency monitoring model can be constructed based on the composition relationship of the combiner boxes; for string stations, a string inefficiency detection model can be constructed based on the inverter inefficiency identification results and the composition relationship of the inverters.
[0032] like Figure 2 As shown, the inefficiency detection module of the photovoltaic equipment according to the exemplary embodiments of this disclosure mainly includes a data acquisition module, a parameter input module, a data processing module (including data cleaning, screening, and filtering), a feature extraction module, an inverter inefficiency model construction and detection module, an inverter inefficiency data display module, a string inefficiency model construction and detection module, and a string inefficiency data display module. Furthermore, for centralized or distributed photovoltaic power stations, it also includes a DC combiner box inefficiency model construction and detection module and a DC combiner box inefficiency data display module.
[0033] Figures 3A to 3C This is a flowchart illustrating an example of an inefficient method for a photovoltaic device according to the present disclosure.
[0034] The inefficiency method for photovoltaic equipment disclosed herein can utilize existing SCADA data to eliminate inefficiencies caused by inverter power limiting. Considering the impact of different seasons and months on power, a power prediction model is constructed based on inverter power, load capacity, month, time point, and irradiance. Thresholds are given based on the training model results to identify inverter inefficiencies. In addition, for inverters with relatively low inefficiency, models for combiner box and string inefficiencies are constructed. For combiner box devices with suspected inefficiencies but without diagnosed inefficiencies, a string inefficiency identification model is constructed.
[0035] Specifically, Figure 3A The inverter inefficiency detection process according to this disclosure is shown. Figure 3B This demonstrates an inefficient testing process for centralized or distributed photovoltaic power generation systems. Figure 3C The procedure for detecting inefficiencies in string photovoltaic power generation systems is shown.
[0036] like Figure 3A As shown, in step S301, data (e.g., inverter active power data) is read. In this example, if the inverter inefficiency detection process is being run for the first time, a longer period of inverter active power data (e.g., at least 2 years) needs to be read. If it is not the first time the inverter inefficiency detection process is being run, a shorter period of inverter active power data (e.g., 3-7 days) can be read. In step S302, the data is cleaned, for example, removing data with fluctuating inverter active power or low inverter load rates. Considering the impact of different seasons, power curtailment, and rainy weather, effective data cleaning enables accurate identification. In step S303, data from devices identified as having power curtailment is removed. The method for determining whether a device is being power curtailed can be as follows (refer to the following). Figure 4 The described method is as follows: In step S304, an inverter power prediction model is constructed. Model input may include the inverter's installed capacity, month, and time (hour). If irradiance data exists, it needs to be input into the model. The model output is the inverter's active power, stored as one model per site. The prediction method can be XGBoost, SVR, etc. In step S305, it is determined whether the inverter power prediction model is being run for the first time. If it is determined in step S305 that the inverter power prediction model is being run for the first time, model training is performed in step S306; otherwise, the trained inverter power prediction model is imported in step S307. Next, in step S308, it is determined whether the inverter inefficiency judgment logic is met. For example, based on the deviation between the predicted value and the actual value of the trained model, combined with the inverter inefficiency characteristics, a specific inverter inefficiency judgment range is designed. For example, if the inverter's predicted power is 20kW, but the actual power is only 10kW, the inverter is judged to be inefficient. The deviation can be proportional or absolute. When the cumulative or average effect of the deviations from multiple time points reaches a set limit, it is judged to be inefficient. If step S309 determines that the inverter inefficiency judgment logic is met, then step S309 determines that the inverter exhibits inefficiency characteristics. If step S309 determines that the inverter inefficiency judgment logic is not met, then step S310 is proceeded to determine whether the inefficiency detection conditions of the subordinate level equipment are met.
[0037] like Figure 3BAs shown, if step S310 determines that the inefficiency detection conditions of the subordinate-level equipment are met, then step S311 determines whether the photovoltaic power generation system being tested is a string photovoltaic power generation system, a centralized photovoltaic power generation system, or a distributed photovoltaic power generation system. If step S311 determines that the photovoltaic power generation system being tested is a centralized photovoltaic power generation system or a distributed photovoltaic power generation system, then the process proceeds to step S312 to construct the combiner box power prediction model. For example, the model input includes the active power of the inverter, the load capacity of the combiner box, the month, and the time (hour). If irradiance data exists, the irradiance needs to be input into the model. The model output is the combiner box power, and the power prediction of all combiner boxes for each site is stored as a model. The prediction method can be XGBoost, SVR, etc. In the example, if the inverter inefficiency detection process is being run for the first time, it is necessary to read inverter active power, combiner box voltage, string current, combiner box power, and combiner box current data for a relatively long period (e.g., at least 2 years). If it is not the first time the inverter inefficiency detection process is being run, data for a shorter period (e.g., 3-7 days) can be read. In step S313, data cleaning is performed, for example, removing data on inverter active power fluctuations, low inverter load rates, combiner box power fluctuations, anomalies, and dead values. Considering the impact of different seasons, power curtailment, and rainy weather, effective data cleaning enables accurate identification. In step S314, power-limited data is removed, that is, data indicating that the inverter is classified as power-limited is removed. In step S315, power calculation is performed. In the example, if combiner box power data exists, it is used directly; if combiner box voltage and current data exist, the combiner box power equals the product of the combiner box voltage and current; if combiner box voltage and string branch current data exist, the combiner box power equals the product of the combiner box voltage and the sum of all string branch currents. In step S316, it is determined whether the combiner box power prediction model is being run for the first time. If it is determined in step S316 that the combiner box power prediction model is being run for the first time, model training is performed in step S317; otherwise, the trained combiner box power prediction model is imported in step S318. Next, in step S319, it is determined whether the combiner box inefficiency judgment logic is met. For example, based on the deviation between the predicted and actual values of the trained model, combined with the combiner box inefficiency characteristics, a specific combiner box inefficiency judgment range is designed. If the difference between the predicted and actual values is within the combiner box inefficiency judgment range, the combiner box is judged to be inefficient. If it is determined in step S319 that the inefficiency judgment logic of the combiner box is met, then in step S320 it is determined that the combiner box exhibits inefficiency characteristics.
[0038] If step S319 determines that the combiner box inefficiency judgment logic is not met, then proceed to step S321 to determine whether the inefficiency detection conditions of the subordinate level equipment are met. If step S321 determines that the inefficiency detection conditions of the subordinate level equipment are met, then in step S322, a string current prediction model is constructed. In the example, the inputs to the string current prediction model include the combiner box power, the connected capacity of the string branch, the month, and the time (hour). If irradiance data exists, the irradiance needs to be input into the model. The model output is the string power, and all string current predictions under each site are stored as one model. The prediction method can be XGBoost, SVR, etc. In the example, if the string inefficiency detection process is run for the first time, it is necessary to read inverter active power, string current, and string voltage data for a longer period (e.g., at least 2 years). If it is not the first time the inverter inefficiency detection process is run, it is possible to read inverter active power, string current, and string voltage data for a shorter period (e.g., 3-7 days). In step S323, data cleaning is performed, for example, removing data related to inverter active power fluctuations, low inverter load rates, abnormal string branch currents, fluctuations, and dead values. In step S324, it is determined whether the string current prediction model is being run for the first time. If it is determined in step S324 that the string current prediction model is being run for the first time, model training is performed in step S325; otherwise, the trained string current prediction model is imported in step S326. Next, in step S327, it is determined whether the string inefficiency judgment logic is met. For example, based on the deviation between the predicted and actual values of the trained model, combined with string inefficiency characteristics, a specific string inefficiency judgment range is designed. If the difference between the predicted and actual values is within the string inefficiency judgment range, the string is judged to be inefficient. If it is determined in step S327 that the string inefficiency judgment logic is met, then in step S328, the string is judged to exhibit inefficiency characteristics.
[0039] like Figure 3CAs shown, if it is determined in step S311 that the photovoltaic power generation system being tested is a string photovoltaic power generation system, then proceed to step S329. In step S329, a string power prediction model is constructed. In this example, the inputs to the string power prediction model include the inverter's active power, the load capacity of the string branches, the current month, and the current time (hour). If irradiance data exists, it needs to be input into the model. The model output is the string power, and the prediction of all string power under each power station is stored as one model. The prediction method can be XGBoost, SVR, etc. In this example, if it is the first time running the string inefficiency detection process, it is necessary to read inverter active power, string current, and string voltage data for a longer period (e.g., at least 2 years). If it is not the first time running the inverter inefficiency detection process, it is possible to read inverter active power, string current, and string voltage data for a shorter period (e.g., 3-7 days). In step S330, data cleaning is performed, for example, removing data showing inverter active power fluctuations, low inverter load rates, abnormal or fluctuating string branch currents, and dead values. Considering the impact of different seasons, power curtailment, and rainy weather, effective data cleaning enables accurate identification. In step S331, power-limited data is removed, i.e., data indicating the inverter is classified as power-limited is removed. In step S332, power calculation is performed. In this example, string power equals the product of string current and string voltage. In step S333, it is determined whether the string power prediction model is running for the first time. If it is determined in step S333 that the string power prediction model is running for the first time, model training is performed in step S334; otherwise, the trained string power prediction model is imported in step S335. Next, in step S336, it is determined whether the string inefficiency judgment logic is met. For example, based on the deviation between the predicted and actual values of the trained model, and combined with the string inefficiency characteristics, a specific string inefficiency judgment range is designed. If the difference between the predicted and actual values falls within the string inefficiency judgment range, the string is judged to be inefficient. If the string inefficiency judgment logic is satisfied in step S336, then the string is judged to exhibit inefficiency characteristics in step S337.
[0040] Considering the hierarchical relationship of equipment in a photovoltaic power plant, and based on the operating mechanism that inefficiency in a few strings does not necessarily lead to inefficiency in the upstream combiner box or string inverter, and inefficiency in a few combiner boxes does not necessarily lead to inefficiency in the upstream inverter, the inefficiency method for photovoltaic equipment disclosed in this invention can accurately detect abnormalities in photovoltaic inverters, combiner boxes, and strings, providing on-site maintenance personnel with an efficient detection method and improving on-site maintenance efficiency. Furthermore, this disclosure considers the influence of seasons, irradiance, and weather, and cleans and filters the data, using time (month and hour) as input to the inverter, combiner box, and string power or current prediction model. Simultaneously, the model input depends on the identification results of the upstream equipment and the power change characteristics, effectively improving the accuracy of the model's inefficiency identification.
[0041] Figure 4 This is a flowchart illustrating a power limiting determination method according to an embodiment of the present disclosure.
[0042] Photovoltaic power plants may implement power limiting controls on some equipment due to power generation conditions. This can cause inverter equipment to exhibit inefficiency. Therefore, to improve the accuracy of inverter inefficiency detection, it is necessary to determine the power-limited equipment at the power plant. The specific logic of the power limiting determination method according to embodiments of this disclosure is described below.
[0043] like Figure 4 As shown, in step S401, data is read. For example, the inverter active power, the power grid AGC power limit flag, and the inverter planned active power for one hour are read, and data with large fluctuations are removed. In step S402, data cleaning is performed. For example, time data where the power grid load factor (power grid load factor = active power of all inverters in the power grid / installed capacity of all inverters in the power grid) exceeds 40% are filtered out, and the amount of cleaned data is required to exceed the amount of original data by more than 30%. Subsequently, in step S403, it is determined whether the data meets the requirements. If it is determined in step S403 that the data meets the requirements, then in step S404, it is determined whether inverter planned power data exists. If it is determined in step S404 that inverter planned power data exists, then proceed to step S405; otherwise, proceed to step S407.
[0044] In step S405, it is determined whether the difference between the inverter's planned active power and actual active power is less than a first predetermined threshold and the difference between the inverter's actual active power and its installed capacity is greater than a second predetermined threshold, wherein the first and second predetermined thresholds are the product of a predetermined percentage (e.g., 5%) and the inverter's installed capacity. If in step S405 it is determined that the difference between the inverter's planned active power and actual active power is less than the first predetermined threshold and the difference between the inverter's actual active power and its installed capacity is greater than the second predetermined threshold, then in step S406 it is determined that the inverter's power is limited.
[0045] In step S407, it is determined whether the automatic generation control (AGC) power limiting flag exists. If the AGC power limiting flag is found to exist in step S407, proceed to step S408; otherwise, proceed to step S410. In step S408, it is determined whether the AGC power limiting flag is 1. If the AGC power limiting flag is found to be 1 in step S408, then in step S409, it is determined that all inverters under the power station are power limited.
[0046] In step S410, the load rates among the inverter subarrays (subarrays being the parent nodes of the inverters) are compared and analyzed. In step S411, the load rate of each subarray and the maximum load rate among all subarrays are calculated. The load rate of an inverter subarray is the average ratio of the active power to the installed capacity of each inverter in the subarray, or the ratio of the sum of the active power of all inverters in the subarray to the sum of the installed capacity of all inverters in the subarray. In step S412, it is determined whether the load rate of the inverter subarray is less than a third predetermined threshold and the power of all inverters in the subarray is greater than a fourth predetermined threshold. The third predetermined threshold is the product of another predetermined percentage (e.g., 80%) and the maximum load rate among all inverter subarrays. The fourth predetermined threshold is, for example, 5kW. If in step S412 it is determined that the load factor of the inverter subarray is less than the third predetermined threshold and the power of all inverters included in the inverter subarray is greater than the fourth predetermined threshold, then in step S413 it is determined that all inverters in the subarray are power limited.
[0047] By employing the power limitation discrimination method according to embodiments of the present disclosure, and utilizing data from different scenarios to design different power limitation judgment logics, inverter devices with power limitations can be effectively identified, improving model accuracy and thus ensuring the accuracy of photovoltaic equipment inefficiency detection.
[0048] Figure 5 This is a diagram illustrating the characteristics of inverter inefficiency. Figure 6 This is a diagram illustrating the characteristics of string inefficiency in string inverters. Figure 7 This diagram illustrates the inefficiencies of the combiner box string branches. Figure 8 This is a diagram illustrating the inefficiencies of the combiner box.
[0049] Reference Figure 5 The bar chart shown indicates that one inverter has a significantly lower equivalent generating hours than the others. (Refer to...) Figure 6 It can be seen that the power of string #6 is significantly lower than that of the other branches, indicating that string #6 is an inefficient string. Through... Figure 7 It can be seen that the current of string group 6 and string group 7 is significantly lower than the average current of other strings (horizontal line), indicating that string groups 6 and 7 exhibit inefficiency. Furthermore, in Figure 8 In the graph, the horizontal axis represents time (displayed as a numerical value, which can be understood as the nth data point in the time period from 00:00:00 to 23:59:59 on a certain day), and the vertical axis represents power. By comparing the power of an inefficient combiner box (the curve on the left) with the power of a normal combiner box (the curve on the right), it can be found that the power of an abnormal combiner box is unstable.
[0050] Figure 9This is a block diagram illustrating an inefficiency detection device for a photovoltaic device according to an exemplary embodiment of the present disclosure.
[0051] like Figure 9 As shown, an inefficiency detection device 900 for a photovoltaic device according to an exemplary embodiment of the present disclosure includes: an operating data acquisition unit 901 configured to acquire operating data of the photovoltaic device; a power limitation determination unit 902 configured to determine whether the inverter of the photovoltaic device is power-limited based on the acquired operating data; an inverter inefficiency prediction unit 903 configured to determine whether the inverter exhibits inefficiency characteristics using an inverter inefficiency prediction model in response to the inverter not being power-limited; and a downstream component inefficiency prediction unit 904 configured to determine whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics in response to the determination that the inverter does not exhibit inefficiency characteristics.
[0052] In the example, the acquired operational data includes the actual active power of the inverter within a predetermined time period, the automatic power generation control power limit flag of the photovoltaic power station, and the planned active power of the inverter.
[0053] In the example, the power limiting determination unit 902 is configured to: determine that the inverter is power-limited in response to the difference between the planned active power and the actual active power of the inverter being less than a first predetermined threshold and the difference between the actual active power and the load capacity of the inverter being greater than a second predetermined threshold; and / or determine that the inverter is power-limited in response to the photovoltaic power plant automatic generation control power limiting flag indicating that all inverters included in the photovoltaic power plant are in a power-limited state; and / or determine that the inverter is power-limited in response to the load rate of the inverter subarray including the inverter being less than a third predetermined threshold and the power of all inverters included in the inverter subarray being greater than a fourth predetermined threshold. For example, the first and second predetermined thresholds are the product of a predetermined percentage and the load capacity of the inverter, and the third predetermined threshold is the product of another predetermined percentage and the maximum value of the load rates of all inverter subarrays. For example, the load factor of an inverter subarray is the average ratio of the active power to the load capacity of each inverter in the inverter subarray, or the ratio of the sum of the active power of all inverters in the inverter subarray to the sum of the load capacity of all inverters in the inverter subarray.
[0054] In the example, the lower-level component inefficiency prediction unit 904 is configured to: in response to determining that the inverter does not exhibit inefficiency characteristics, use a combiner box inefficiency prediction model to determine whether the combiner box connected to the inverter exhibits inefficiency characteristics; and in response to determining that the combiner box does not exhibit inefficiency characteristics, use a string inefficiency prediction model to determine whether the photovoltaic string connected to the combiner box that does not exhibit inefficiency characteristics exhibits inefficiency characteristics.
[0055] In the example, the lower-level component inefficiency prediction unit 904 is configured to: in response to determining that the inverter does not exhibit inefficiency characteristics, use a string inefficiency prediction model to determine whether the photovoltaic string connected to the inverter exhibits inefficiency characteristics.
[0056] The above combination Figures 1 to 4 The specific operations shown are respectively by Figure 9 The corresponding unit in the photovoltaic equipment inefficiency detection device 900 shown performs the operation; specific operational details will not be elaborated here. Through the photovoltaic equipment inefficiency detection device according to the exemplary embodiments of this disclosure, inefficiency detection of photovoltaic inverters, combiner boxes, and strings can be achieved using existing SCADA data, providing strong support for on-site operation and maintenance. Furthermore, considering the hierarchical relationship between photovoltaic power plant inverters, combiner boxes, and strings, inefficiency equipment can be located and identified more accurately.
[0057] Figure 10 This is a block diagram illustrating a computing system including at least one computing device and at least one storage device of storage instructions according to an exemplary embodiment of the present disclosure.
[0058] like Figure 10 As shown, the computing system 1000 provided according to an exemplary embodiment of the present invention includes a computing device 1001 and a storage device 1002. The storage device 1002 stores computer-executable instructions. When the computer-executable instructions are executed by the computing device 1001, the inefficiency detection method of photovoltaic equipment described in any of the foregoing embodiments is executed.
[0059] The computing device 1001 can be deployed in a server or client, or on a node device in a distributed network environment. Furthermore, the computing device 1001 can be a PC, tablet, personal digital assistant, smartphone, web application, or other device capable of executing the aforementioned set of instructions. Here, the computing device is not necessarily a single computing device; it can be any collection of devices or circuits capable of executing the aforementioned instructions (or instruction sets) individually or in combination. The computing device can also be part of an integrated control system or system manager, or can be configured to interconnect locally or remotely (e.g., via wireless transmission) through an interface. In the computing device, the processor includes a central processing unit (CPU), a graphics processing unit (GPU), a programmable logic device, a dedicated processor system, a microcontroller, or a microprocessor. By way of example and not limitation, the processor also includes analog processors, digital processors, microprocessors, multi-core processors, processor arrays, network processors, etc.
[0060] According to another aspect of this disclosure, a computer-readable storage medium is provided that stores instructions, which, when executed by at least one computing device, cause the at least one computing device to perform the inefficiency detection method for photovoltaic devices described in any of the foregoing embodiments. The computer-readable storage medium includes magnetic media such as floppy disks and magnetic tapes, optical media (including optical disc (CD) ROMs and DVD ROMs), magneto-optical media such as floppy discs, hardware devices such as ROMs and RAMs designed for storing and executing program commands, and flash memory. The instructions may include language code executable by a computer using an interpreter and machine language code generated by a compiler.
[0061] By adopting this disclosure, inverters exhibiting inefficiency characteristics can be identified more accurately, and considering the hierarchical relationship between inverters, combiner boxes, and strings in a photovoltaic power plant, inefficient equipment can be precisely located and identified.
[0062] The processes, methods, or algorithms disclosed herein can be transmitted to, or implemented by, a processing device, controller, or computer, which may include any existing programmable electronic control unit or a dedicated electronic control unit. Similarly, the processes, methods, or algorithms can be stored in various forms as data and instructions executable by a controller or computer, including but not limited to information permanently stored on non-writable storage media (such as ROM devices) and information variablely stored on writable storage media (such as floppy disks, magnetic tapes, CDs, RAM devices, and other magnetic and optical media). The processes, methods, or algorithms can also be implemented in a software executable object. Optionally, the processes, methods, or algorithms can be implemented wholly or partially using suitable hardware components (such as ASICs, FPGAs, state machines, controllers, or other hardware components or devices) or a combination of hardware components, software components, and firmware components.
[0063] Although this disclosure includes specific examples, it will be apparent to those skilled in the art that various changes in form and detail may be made to these examples without departing from the spirit and scope of the claims and their equivalents. The examples described herein are to be considered merely for descriptive purposes and not for limiting purposes. The description of features or aspects in each example is to be considered applicable to similar features or aspects in other examples. Suitable results may be obtained if the described techniques are performed in a different order, and / or if components in the described system, architecture, apparatus, or circuit are combined in a different manner and / or if components in the described system, architecture, apparatus, or circuit are replaced or supplemented with other components or their equivalents. Therefore, the scope of this disclosure is not limited by the specific embodiments but by the claims and their equivalents, and all variations within the scope of the claims and their equivalents shall be construed as included in this disclosure.
Claims
1. A method for detecting inefficiency in photovoltaic equipment, characterized in that, The method for detecting inefficiency of photovoltaic equipment includes: Obtain operational data from photovoltaic equipment; Based on the acquired operational data, determine whether the inverter of the photovoltaic equipment is being power limited; In response to the inverter not being power-limited, an inverter inefficiency prediction model is used to determine whether the inverter exhibits inefficiency characteristics. In response to determining that the inverter does not exhibit inefficiency characteristics, determine whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics.
2. The method for detecting inefficiency of photovoltaic equipment according to claim 1, characterized in that, The operating data includes the actual active power of the inverter within a predetermined time period, the automatic power generation control limit flag of the photovoltaic power station, and the planned active power of the inverter.
3. The method for detecting inefficiency of photovoltaic equipment according to claim 2, characterized in that, Based on the acquired operational data, the steps to determine whether the inverter of a photovoltaic device is being power-limited include: In response to the fact that the difference between the planned active power and the actual active power of the inverter is less than a first predetermined threshold and the difference between the actual active power and the installed capacity of the inverter is greater than a second predetermined threshold, the inverter is determined to be power-limited; and / or In response to the automatic power generation control power limiting flag indicating that all inverters included in the photovoltaic power plant are in a power limiting state, it is determined that the inverter is power limited; and / or In response to the fact that the load factor of the inverter subarray including the inverter is less than a third predetermined threshold and the power of all inverters included in the inverter subarray is greater than a fourth predetermined threshold, it is determined that the inverter is power limited.
4. The method for detecting inefficiency in photovoltaic equipment according to claim 3, characterized in that, The first predetermined threshold and the second predetermined threshold are products of a predetermined percentage and the load capacity of the inverter, and the third predetermined threshold is a product of another predetermined percentage and the maximum load rate among all inverter subarrays.
5. The method for detecting inefficiency in photovoltaic equipment according to claim 3, characterized in that, The load factor of an inverter subarray is the average ratio of the active power to the load capacity of each inverter in the inverter subarray, or the ratio of the sum of the active power of all inverters in the inverter subarray to the sum of the load capacity of all inverters in the inverter subarray.
6. The method for detecting inefficiency in photovoltaic equipment according to claim 1, characterized in that, In response to determining that the inverter does not exhibit inefficiency, the step of determining whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency includes: In response to determining that the inverter does not exhibit inefficiency characteristics, a combiner box inefficiency prediction model is used to determine whether the combiner box connected to the inverter exhibits inefficiency characteristics; In response to the determination that the combiner box does not exhibit inefficiency characteristics, a string inefficiency prediction model is used to determine whether the photovoltaic strings connected to the combiner box that does not exhibit inefficiency characteristics exhibit inefficiency characteristics.
7. The method for detecting inefficiency of photovoltaic equipment according to claim 1, characterized in that, In response to determining that the inverter does not exhibit inefficiency, the step of determining whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency includes: In response to determining that the inverter does not exhibit inefficiency characteristics, a string inefficiency prediction model is used to determine whether the photovoltaic string connected to the inverter exhibits inefficiency characteristics.
8. A device for detecting inefficiency in photovoltaic equipment, characterized in that, The inefficiency detection device for the photovoltaic equipment includes: The operation data acquisition unit is configured to acquire the operation data of the photovoltaic equipment; The power limiting determination unit is configured to determine whether the inverter of the photovoltaic equipment is subject to power limiting based on the acquired operating data. An inverter inefficiency prediction unit is configured to determine whether the inverter exhibits inefficiency characteristics using an inverter inefficiency prediction model in response to the inverter not being power-limited. The lower-level component inefficiency prediction unit is configured to determine whether the combiner box or photovoltaic string connected to the inverter exhibits inefficiency characteristics in response to determining that the inverter does not exhibit inefficiency characteristics.
9. A computing system comprising at least one computing device and at least one storage device for storing instructions, characterized in that, When the instruction is executed by the at least one computing device, it causes the at least one computing device to perform the inefficient detection method for photovoltaic equipment according to any one of claims 1-6.
10. A computer-readable storage medium for storing instructions, characterized in that, When the instruction is executed by at least one computing device, it causes the at least one computing device to perform the inefficient detection method for photovoltaic equipment according to any one of claims 1-6.