A photovoltaic module monitoring system based on power generation analysis

Abstract

The application discloses a photovoltaic module monitoring system based on power generation analysis, and relates to the technical field of photovoltaic module monitoring.The system comprises a data acquisition module, which is used for acquiring operation data of multiple photovoltaic modules in a photovoltaic array; a data processing module, which is used for identifying a candidate module set with a power generation attenuation trend based on the operation data; and a sequence generation module, which is used for sorting the modules in the candidate module set to generate a monitoring sequence; wherein the basis for sorting at least includes the attenuation suspiciousness of each module; the application identifies initial faults through hidden crack attenuation exclusive rules, combines dynamic scheduling of hierarchical retrieval and resource adaptation of associated modules, realizes high-risk module priority detection, and reduces the false alarm and missed alarm rates through short-period tracking and verification after maintenance.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic module monitoring technology, specifically a photovoltaic module monitoring system based on power generation analysis. Background Technology

[0002] As the core component of solar power generation systems, photovoltaic (PV) modules offer advantages such as cleanliness, environmental friendliness, and sustainability, and are widely used in various power generation scenarios. However, with the expansion of PV power plant scale, modules are susceptible to environmental factors and installation processes during long-term outdoor operation, leading to faults such as microcracks. This not only causes performance degradation but may also pose safety risks. Therefore, PV module monitoring systems have become crucial equipment for power plant operation and maintenance.

[0003] Existing photovoltaic module monitoring systems primarily rely on power generation analysis, comparing actual power generation with standard power generation to determine if modules exhibit abnormalities. These systems typically employ a uniform degradation assessment standard, with some incorporating adjustments based on factors such as ambient light and temperature. However, degradation caused by microcracks exhibits a unique pattern: initially slow, then accelerating, fundamentally different from the linear degradation caused by dust accumulation or slight shading. Therefore, existing systems have failed to effectively identify microcrack faults.

[0004] The characteristics of microcrack faults lead to false alarms and missed alarms in existing monitoring systems. Mistaking normal losses for microcracks can result in unnecessary maintenance work, while missing the initial attenuation of microcracks can cause the fault to expand, affecting the overall efficiency of the power plant and increasing safety hazards.

[0005] Therefore, it is necessary to develop specialized monitoring logic for the characteristics of microcracks in order to improve the ability to identify and judge them in the early stage, reduce operation and maintenance costs, and ensure the safe and efficient operation of the power plant. Summary of the Invention

[0006] The purpose of this invention is to provide a photovoltaic module monitoring system based on power generation analysis to solve the problems mentioned in the background art.

[0007] A photovoltaic module monitoring system based on power generation analysis includes:

[0008] The data acquisition module is used to acquire the operating data of multiple photovoltaic modules in the photovoltaic array;

[0009] The data processing module is used to identify a set of candidate components whose starting power shows a decreasing trend based on the running data;

[0010] It should be noted that the core function of the current module is to filter out components that may have hidden cracks or other faults from massive amounts of operational data, thereby reducing ineffective maintenance workload, improving fault diagnosis efficiency, avoiding maintenance personnel from blindly testing all components, and saving manpower and time costs.

[0011] The sequence generation module is used to sort the components in the candidate component set to generate a monitoring sequence; wherein, the sorting is based on at least the attenuation suspicion of each component; the attenuation suspicion is determined by analyzing whether there are persistent and nonlinear attenuation characteristics related to hidden crack faults in the historical power generation data of the components.

[0012] This is done to clarify the priority of on-site verification, so that operation and maintenance resources can be prioritized for components with higher suspicion of failure, thereby improving the efficiency of operation and maintenance response, avoiding delays in handling high-risk components due to prioritizing low-susceptibility components, and reducing the loss of power generation caused by failures.

[0013] The scheduling management module is used to assign differentiated first on-site verification time points to components in the candidate component set based on the monitoring sequence; the earlier the component is ranked in the monitoring sequence, the earlier the first on-site verification time point is assigned.

[0014] The core objective is to rationally arrange review tasks based on the status of operation and maintenance resources, ensuring that highly suspicious components are processed first, improving resource utilization efficiency, avoiding idle or overcrowded operation and maintenance resources, and ensuring that review tasks proceed in an orderly manner.

[0015] The task execution module is used to generate on-site verification instructions for performing on-site verification of hidden cracks on components in the candidate component set according to the allocated verification time points.

[0016] Specifically, this module mainly provides clear operation guidance for operations and maintenance personnel, while realizing closed-loop management of review results, so that the operation and maintenance process forms a complete link from instruction generation to result feedback, which is convenient for traceability and optimization.

[0017] Preferably, the sequence generation module determines the degree of doubt about the attenuation by the functional relationship between the total power generation attenuation rate and the attenuation duration of the component within a preset historical period;

[0018] The higher the total power generation attenuation rate and the longer the attenuation duration, the higher the degree of suspicion of attenuation.

[0019] Preferably, the scheduling management module performs the following operations:

[0020] The first N components in the monitoring sequence are included in the first round of review; the next M components are included in the second round of review; where N and M are determined based on the total amount of monitoring resources currently available.

[0021] A first time window is allocated to the first round of review batches, and a second time window is allocated to the second round of review batches, wherein the end time of the first time window is earlier than the start time of the second time window;

[0022] Based on the component's order within its batch, specify the exact first on-site review time point within the corresponding review time window for that batch.

[0023] Preferably, the data processing module analyzes the attenuation characteristics in the following ways:

[0024] Obtain power generation data of the component over multiple consecutive equal-length periods;

[0025] Calculate the rate of change of power generation in each time period relative to the previous period;

[0026] If there are K or more consecutive time periods in which the rate of change of power generation is negative, and the absolute value of the rate of change in the last period is greater than the absolute value of the rate of change in the first period, then it is determined that there is a continuous and nonlinear decay characteristic.

[0027] Preferably, the value of K is set according to the length of the time period: when the time period is days, K≥7; when the time period is weeks, K≥3.

[0028] Preferably, it also includes a dynamic optimization module, used to perform the following operations when a component is confirmed to have a hidden crack fault based on an on-site verification instruction:

[0029] Based on the attribute information of the confirmed microcracked component, related components with at least one common hidden danger attribute are automatically retrieved in the photovoltaic array; the common hidden danger attribute includes: belonging to the same production batch, having the same installation time, or being located on the same installation plane;

[0030] The associated components are added to the first batch of reviews, and the scheduling management module is triggered to adjust the time window of the batch forward. At the same time, the earliest review time point within the adjusted time window is assigned to the added associated components.

[0031] Preferably, the dynamic optimization module is also used for:

[0032] When the number of components in the first review batch exceeds the threshold N due to the addition of associated components, one or more of the last components in the batch are moved to the head of the second review batch, and the scheduling management module is triggered to adjust the second time window of the second review batch accordingly.

[0033] The innovation of this invention lies in the following aspects. First, the mechanism innovatively sets a quantity threshold N for resource adaptation. Through rigid control, it ensures that the number of components in the first batch is always accurately matched with the operation and maintenance resources, avoiding problems such as excessive workload per person and equipment usage conflicts. This directly solves the problem of efficiency decline caused by resource imbalance, while ensuring sufficient testing resources for high-risk components and eliminating the risk of missed detection caused by simplifying the process to meet deadlines.

[0034] Secondly, the cross-batch transmission design adopted in this step ensures that low-priority components removed in the first round can still maintain their high ranking in the second round. This design does not crowd out the testing resources of high-risk components in the first round, and avoids delayed testing of removed components due to batch adjustments, effectively solving the priority gap problem that is prone to occur in traditional removal operations.

[0035] Finally, the linkage adjustment of the time windows of the two batches was realized. Because the window of the associated component is moved forward in the first round, the window is adjusted according to the number of components moved in in the second round, so that the scheduling of the two core review batches can form a synergistic effect, which solves the problem of low overall efficiency caused by isolated batch scheduling in the traditional mode.

[0036] Preferably, the dynamic optimization module retrieves related components in the following ways:

[0037] If there are other components from the same production batch as the confirmed microcrack component, they will all be classified as first-priority associated components.

[0038] Otherwise, other components that have the same installation time as the confirmed cracked component will be classified as second-priority associated components;

[0039] Among them, when the first priority associated component is added to the first round of review batch, the review order assigned to it takes precedence over the second priority associated component.

[0040] Preferably, the dynamic optimization module is also used for:

[0041] After adding the associated components to the first review batch, the newly added associated components are merged with the original components in the first review batch.

[0042] Based on the degree of attenuation suspicion of the merged components, all components in the first review batch were re-sorted;

[0043] Based on the reordered sequence, the scheduling management module is triggered to reallocate specific review time points for all components in the first review batch within the first time window of the forward adjustment.

[0044] The core innovation of this invention lies in prioritizing high-risk components by sorting them in descending order of attenuation suspicion level, and differentiating priorities within the same suspicion level to prevent high-risk components from being delayed, thereby improving the accuracy of risk control in the first round of review. By dividing the time window into equal periods and allocating them according to the order, testing resources are precisely matched to risk levels, while also facilitating route planning for maintenance personnel and improving overall testing efficiency. This mechanism also forms a closed loop with associated component retrieval and quantity threshold control. After associated components are merged and sorted, threshold verification is performed to ensure the synergy between sorting accuracy and quantity control, achieving a refined and efficient first-round review process, and comprehensively ensuring the early and accurate detection of hidden crack faults in photovoltaic power plants.

[0045] Preferably, it also includes a closed-loop management module, which is used to initiate a continuous tracking and monitoring process after performing maintenance or replacement operations on components with confirmed microcracks and related components with confirmed microcracks through on-site verification.

[0046] Set a specific tracking monitoring cycle for the component that is shorter than the regular monitoring cycle, and continuously collect its power generation data;

[0047] If the power generation data of the component remains within the normal fluctuation range based on its rated power within a preset T consecutive tracking and monitoring cycles, it is determined that its state has stabilized, the hidden danger has been eliminated, and the continuous tracking and monitoring will end, restoring it to the normal monitoring mode.

[0048] The core innovation of this invention lies in ensuring the elimination of potential hazards and reducing the occurrence of secondary failures through continuous follow-up after maintenance. The module employs a judgment logic that combines short-cycle high-frequency monitoring with multi-cycle verification. By shortening the monitoring cycle and setting continuous verification rules, it effectively identifies anomalies and eliminates environmental interference, thereby improving the accuracy of hazard judgment.

[0049] This module also works in conjunction with the dynamic optimization module to ensure that abnormal components are promptly returned to the first round of review, enabling a rapid connection between verification and troubleshooting, preventing the spread of faults, while stable components automatically return to routine monitoring to optimize resource allocation.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] The innovation of this invention lies in addressing the problems of false alarms and missed alarms, and the inability to distinguish between faults and normal losses caused by the lack of dedicated logic for detecting microcracks in existing systems. This system designs dedicated identification rules by analyzing the progressive attenuation characteristics of microcracks, accurately capturing the initial weak attenuation and distinguishing microcracks from normal losses such as dust accumulation. Combined with priority ranking and graded handling, it reduces ineffective operation and maintenance caused by false alarms and avoids performance degradation and safety risks caused by missed alarms. The whole process management helps power plants reduce costs and increase efficiency while ensuring safe operation. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the system framework structure of the present invention. Detailed Implementation

[0053] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Please see Figure 1This application provides a photovoltaic module monitoring system based on power generation analysis, comprising:

[0055] The data acquisition module is used to acquire the operating data of multiple photovoltaic modules in the photovoltaic array;

[0056] It should be noted that the data collected by this module covers the real-time power generation of each component, the cumulative value of power generation per hour, the daily power generation curve, as well as the irradiance and ambient temperature of the environment in which the component is located.

[0057] In practice, the module establishes stable communication with the combiner box and inverter via industrial Ethernet, interacts with environmental monitoring sensors via RS485 bus, and completes data acquisition according to a preset acquisition cycle. The communication uses an encrypted transmission protocol to prevent tampering, and the error of the environmental monitoring sensors is controlled within ±5% after calibration, which meets the requirements for acquisition accuracy in the photovoltaic power plant operation and maintenance management guidelines. The communication bus and encrypted transmission protocol are existing mature technologies and will not be elaborated on here.

[0058] In this embodiment, the initial data acquisition cycle is set to 15 minutes. The system monitors irradiance fluctuations in real time. Once the minute-level fluctuation exceeds 200W / ㎡, it automatically shortens to 5 minutes. This minimum cycle also meets the requirement in the photovoltaic power plant consumption monitoring and statistical management regulations that the data acquisition cycle for power generation should generally not exceed 5 minutes. This design of dynamically adjusting the acquisition cycle reduces unnecessary resource consumption when the data is stable and ensures the integrity of key data when data fluctuations are large, laying the foundation for subsequent accurate analysis.

[0059] The data processing module is used to identify a set of candidate components that show a decreasing trend in power consumption based on the running data.

[0060] Specifically, to achieve this, the operational data must first be preprocessed, including three steps: removing outlier data, filling in missing data, and data standardization. Data reliability is a prerequisite for accurate subsequent analysis, and preprocessing can effectively reduce the risk of misjudgment or omission due to data problems.

[0061] In this embodiment, abnormal data is removed by determining whether it falls within a reasonable range of 0 to 1.2 times the nominal power. This range refers to the normal fluctuation range of the nominal power of the component.

[0062] When data is missing for three or more consecutive time periods, the weighted average interpolation method of the data of adjacent strings of modules in the same period is used to fill the missing data. The weights can be set to 0.6, 0.3, or 0.1 according to the distance. This is because the closer the physical distance within the string of modules, the higher the correlation of power generation characteristics. This is a consensus in the industry and will not be elaborated on further.

[0063] Data standardization converts actual power generation into equivalent power generation under standard irradiance conditions (1000W / ㎡) and standard temperature (25℃), which meets the standard operating conditions requirements of national photovoltaic module degradation testing and ensures that the power generation of modules under different environments is comparable.

[0064] Furthermore, after preprocessing, the current module analyzes attenuation characteristics in a specific way to filter candidate components, specifically:

[0065] The power generation data of the component is obtained over several consecutive equal time periods, and the rate of change of power generation in each time period relative to the previous period is calculated. If there are K or more consecutive time periods in which the rate of change of power generation is negative, and the absolute value of the rate of change in the last period is greater than the absolute value of the rate of change in the first period, then it is determined that there is a continuous and non-linear decay characteristic, and such components will be included in the candidate component set.

[0066] It should be understood that the value of K is set according to the length of the time period: when the time period is days, K≥7; when the time period is weeks, K≥3. This setting is based on the gradual aggravation law of microcrack faults and refers to the periodic verification requirements for microcrack fault monitoring in the "Technical Specification for Fault Diagnosis of Photovoltaic Modules". In short periods (days), the power generation of the module is greatly affected by the diurnal temperature difference and instantaneous irradiance fluctuations. A longer continuous period (≥7 days) is needed to eliminate accidental factors to verify the degradation trend. In long periods (weeks), the data has smoothed out short-term fluctuations, and the degradation characteristics can be accurately captured by fewer consecutive times (≥3 weeks).

[0067] The theoretical power generation is calculated according to the industry standard formula: Theoretical power generation = nominal power of the module × actual irradiance ÷ standard irradiance × [1 - temperature coefficient × (actual ambient temperature - 25℃)], where the standard irradiance is 1000W / ㎡ and the temperature coefficient is preset to 0.0042 according to the module model, which can help verify the rationality of the attenuation characteristics.

[0068] Meanwhile, it is clarified that the normal fluctuation range is ±10% of the power generation of modules of the same batch, model, and installation environment during the same period. This range is set with reference to the consistency requirements of module power generation in the "Evaluation Specification for Operation Performance of Photovoltaic Power Stations". Taking a nominal 300W module as an example, if the power generation change rate for 7 consecutive days is -1%, -1.2%, -1.5%, -1.8%, -2%, -2.3%, and -2.6% respectively, all of which are negative, and the absolute value of the change rate on the last day (2.6%) is greater than that on the first day (1%), then it is determined that there is a continuous and non-linear decay characteristic, and it is included in the candidate set.

[0069] If the rate of change is negative for 6 consecutive days, or if it is negative for 7 consecutive days but the absolute value of the rate of change on the last day does not exceed that of the first day, or if the power generation fluctuation is within ±10% of the same batch, it will not be included. This judgment method can accurately distinguish between hidden crack faults and normal attenuation.

[0070] The sequence generation module is used to sort the components in the candidate component set and generate a monitoring sequence. It is important to understand that the sorting is based on the decay suspicion as the core indicator, combined with the installation age and crossbit importance as auxiliary indicators. The decay suspicion has the highest priority, followed by the installation age, and crossbit importance is used as a supplement. This aligns with the actual need in operation and maintenance to prioritize monitoring faults with higher correlation.

[0071] In this embodiment, the attenuation suspicion is determined by the functional relationship between the total power generation attenuation rate and the attenuation duration of the module within a preset historical period. The higher the total power generation attenuation rate and the longer the attenuation duration, the higher the attenuation suspicion is determined. This functional relationship is constructed based on the gradual attenuation characteristics of the microcrack fault and refers to the microcrack attenuation data model in the "On-site Attenuation Test Method for Crystalline Silicon Photovoltaic Modules", which can accurately match the fault development law.

[0072] Specifically, the preset historical period is set to 90 days. The total power generation attenuation rate is calculated as (average power generation in the first 30 days of the period - average power generation in the last 30 days of the period) ÷ average power generation in the first 30 days of the period × 100%. The attenuation duration refers to the number of consecutive days from when the power generation first appears to be below the normal fluctuation range (±10% of the power generation of components in the same batch, model, and installation environment during the same period) to the present.

[0073] When the total power generation attenuation rate exceeds 8% and the attenuation lasts for more than 30 days, it can be judged as highly suspicious; when the attenuation rate is between 5% and 8% and the duration is 20 to 30 days, it is considered moderately suspicious; when the attenuation rate is between 3% and 5% and the duration is 10 to 20 days, it is considered lowly suspicious. The above threshold ranges are set in combination with the measured attenuation data of photovoltaic module microcrack faults and the industry fault judgment standards in the "Technical Specification for Fault Diagnosis of Photovoltaic Modules".

[0074] In addition, the longer the installation period, the higher the probability of microcracks in the modules due to material aging and encapsulation layer degradation, and the higher the priority of sorting will be. This is consistent with the rule that the failure rate of photovoltaic modules increases with the service life.

[0075] The importance of the components in the photovoltaic string is set according to their position. The first component affects the stability of the power supply of the entire string and is therefore more important. It can be ranked earlier. This design can prioritize the normal operation of components in key positions and reduce the impact on the overall power generation of the array.

[0076] After comprehensively evaluating these indicators, a monitoring sequence is generated from high to low based on the suspicion of faults, and synchronized to the scheduling and management module. The sequence sorting algorithm is a mature existing technology, which will not be elaborated on here. For example, in the candidate set, components A and B have different stats. Component A has a 90-day total power generation decay rate of 12% and a decay duration of 45 days. It has been installed for 5 years and is the first component in the series. Component B has a 90-day total power generation decay rate of 6% and a decay duration of 25 days. It has been installed for 3 years and is a middle component in the series. Considering all factors, A is ranked higher than B.

[0077] The scheduling management module is used to assign differentiated initial on-site verification time points to each component in the candidate component set based on the monitoring sequence. In this embodiment, in specific operations, the module first completes batch division and time window allocation, and then refines it to the specific verification time point for each component. The entire process conforms to the actual needs of batch scheduling of operation and maintenance tasks.

[0078] First, there's the batch division. The module will include the top N components in the monitoring sequence in the first round of review batches, and the remaining M components in the second round of review batches. The specific values ​​of N and M are determined entirely based on the total amount of currently schedulable monitoring resources. It should be understood that the total amount of monitoring resources includes core resources such as the number of maintenance personnel, the number of available EL detectors, and the maintenance vehicle scheduling plan. The system's built-in resource status monitoring unit will collect and update this resource information in real time. This unit is a mature existing technology and will not be elaborated on here.

[0079] Furthermore, to enhance the objectivity of N and M value settings, the calculation logic for daily review volume is clarified: Daily review volume = number of maintenance personnel × average daily inspection volume per person (refer to the "Skill Specifications for Photovoltaic Power Station Maintenance Personnel", with the average daily inspection volume per person set at 2 modules / person). At the same time, the minimum value of the three factors is taken as the maximum daily review volume, considering the number of available EL testing instruments (1 instrument supports the testing of 10 modules per day) and the dispatching capacity of maintenance vehicles (1 vehicle supports the transportation of 30 modules per day).

[0080] For example, if there are currently 15 maintenance personnel (15×2=30 blocks), 10 EL testing machines (10×10=100 blocks), and 3 maintenance vehicles (3×30=90 blocks), then the maximum daily review volume is the minimum value of 30 blocks. If it is expected that there will be no additional resource usage in the next 3 days, then N=60 (the first round is completed in 2 working days, 30×2=60 blocks) and M=90 (the second round is completed in 3 working days, 30×3=90 blocks) can be set to ensure that resources can fully cover the review needs of each batch.

[0081] Next is the allocation of time windows. The first time window is allocated to the first batch of review, and the second time window is allocated to the second batch of review. It is clear that the end time of the first time window is earlier than the start time of the second time window. This can effectively avoid the overlap of review tasks between the two batches, reduce resource scheduling conflicts, and allow maintenance personnel to concentrate on completing the tasks of the current batch.

[0082] In this method, if the first round of review batches N=60, combined with the daily review capacity of 30 blocks, the first time window is set to 1-2 working days after the instruction is generated;

[0083] The second round of review batch M=90, corresponding to the second time window being set as the 3rd to 5th working days, which ensures that the first round is prioritized while allowing sufficient time for resource allocation in the second round.

[0084] Finally, after completing the batch and time window division, based on the component's ranking within its batch, a specific initial on-site review time point is assigned to each component within the corresponding batch's review time window. Specifically, the earlier the component is ranked within the batch, the earlier its specific time point within the time window. This further refines priorities within the batch, ensuring that highly suspicious components within the same batch still receive priority attention. For example, components 1-15 in the first round of review are scheduled for the morning of the first working day of the first time window; components 16-30 are scheduled for the afternoon of the first working day; components 31-45 are scheduled for the morning of the second working day; components 46-60 are scheduled for the afternoon of the second working day; components 1-30 in the second round of review (i.e., components 61-90 in the monitoring sequence) are scheduled for the third working day, and so on.

[0085] If a temporary shortage of resources occurs in a certain batch, such as a sudden failure of one EL tester during the first round of review, the number of available instruments will be reduced to 9 (9 x 10 = 90 units, still higher than the daily review volume of 30 units, and will not affect the daily review capacity); if the maintenance personnel are temporarily reduced by 5 people, leaving 10 people (10 x 2 = 20 units), the daily review volume will be reduced to 20 units. The specific time point for the 41st to 60th components in the first round can be postponed to the morning of the 3rd working day. At the same time, the time window for the second round will be moved back by 1 working day, and a notification containing the reason for the adjustment and the new time arrangement will be sent to the maintenance management terminal to ensure that the maintenance personnel are informed in a timely manner.

[0086] Taking a power plant as an example, the currently available resources are 15 maintenance personnel, 10 EL detectors, and 3 maintenance vehicles. The maximum daily verification volume is 30 components, and the monitoring sequence has a total of 150 components. N=60 (first round) and M=90 (second round) are set. The first time window is 1-2 working days after the instruction is generated, and the second time window is the 3rd-5th working days. In the first round, components 1-30 are arranged on the 1st working day, and components 31-60 are arranged on the 2nd working day. In the second round, components 1-30 (sequence 61-90) are arranged on the 3rd working day, components 31-60 (sequence 91-120) are arranged on the 4th working day, and components 61-90 (sequence 121-150) are arranged on the 5th working day. The entire scheduling plan matches resource capacity and ensures priority order.

[0087] The task execution module generates on-site verification instructions for performing microcrack on-site verification on components in the candidate component set, based on the initial on-site verification time. Specifically, a complete verification instruction includes the unique code of the component to be verified, installation coordinates accurate to the bracket number, a verification time window accurate to the hour, a recommended EL (Elastic Optical Detector) model, and standard operating procedures for microcrack detection, such as avoiding detection during periods of strong sunlight and ensuring the shooting angle is perpendicular to the component surface. These requirements comply with industry inspection standards for photovoltaic module EL testing, ensuring that maintenance personnel can immediately use the instruction, reducing operational errors and improving detection accuracy.

[0088] In addition, the module can receive and verify the review results. The results must include the detection photos and the fault description. The system will automatically compare the photos with historical data and generate different levels of fault work orders according to the maintenance principle that the larger the scope of the fault impact, the higher the priority of handling. This hierarchical handling method can enable serious faults to be responded to quickly and reduce the risk of fault expansion.

[0089] If no hidden cracks are found during the review, the system will send the complete component data and review report back to the data processing module to reassess the cause of the degradation and update the health record, providing data support for subsequent operation and maintenance.

[0090] In this embodiment, a Level 1 fault work order is generated when the area of ​​microcracks exceeds 5%, requiring replacement and simultaneous procurement of the module within 48 hours; a Level 2 work order is generated when the area is between 1% and 5%, requiring repair and development of a follow-up monitoring plan within 7 working days. For example, the review instruction for a candidate component clearly specifies the code PV202305001, the installation coordinate 3 bracket, the 5th piece in the 2nd row, the review time is 9:00 to 11:00 the next day, the EL-800 detector is recommended, and a special reminder is given to avoid strong light from 10:00 to 14:00. After the maintenance personnel provide the inspection photos and description, the system determines that the microcrack area is 6.2%, and immediately generates a Level 1 review work order, requiring replacement and simultaneous procurement of the module within 48 hours.

[0091] The innovation of this invention lies in addressing the problems of false alarms and missed alarms, and the inability to distinguish between faults and normal losses caused by the lack of dedicated logic for detecting microcracks in existing systems. This system designs dedicated identification rules by analyzing the progressive attenuation characteristics of microcracks, accurately capturing the initial weak attenuation and distinguishing microcracks from normal losses such as dust accumulation. Combined with priority ranking and graded handling, it reduces ineffective operation and maintenance caused by false alarms and avoids performance degradation and safety risks caused by missed alarms. The whole process management helps power plants reduce costs and increase efficiency while ensuring safe operation.

[0092] Specifically, microcrack failures in photovoltaic modules are not entirely isolated events. Modules from the same production batch may have the same structural defects due to inherent factors such as fluctuations in raw material quality and deviations in manufacturing processes, resulting in a much higher microcrack incidence rate than other batches. Modules installed at the same time may experience synchronous aging or damage due to consistent installation techniques (such as excessive installation torque or uneven stress) or environmental stresses experienced concurrently (such as extreme temperature differences or strong wind loads). Modules on the same installation plane may exhibit similar degradation and microcrack development trends due to acquired factors such as consistent lighting conditions, similar shading, and similar degrees of environmental corrosion. Existing systems do not perform correlation analysis on these common potential hazards, resulting in the inability to promptly identify related modules of confirmed faulty modules. This may lead to individual handling followed by batch recurrence, increasing subsequent operation and maintenance costs and causing even larger-scale power generation losses due to the continued development of potential faults.

[0093] As one embodiment of the present invention, a dynamic optimization module is also included, which is used to perform the following operations when a component is confirmed to have a hidden crack fault based on an on-site verification instruction:

[0094] Based on the attribute information of the confirmed microcracked module, related modules are automatically searched and classified in the photovoltaic array. The specific search rules are as follows: other modules belonging to the same production batch as the confirmed microcracked module are searched first. If such modules exist, they are all classified as first-priority related modules. If other modules in the same production batch do not exist, other modules with the same installation time as the confirmed microcracked module are searched and classified as second-priority related modules.

[0095] In this embodiment, the attribute information of the confirmed hidden crack component is retrieved in real time through the system database. The database pre-stores all attribute data collected by the data acquisition module when all components are connected to the grid, including production batch number, installation timestamp, and installation coordinates. The production batch number is in the format of manufacturer code plus production year plus month plus batch number, such as JCD20230503. The installation timestamp is accurate to YYYY-MM-DD, such as 2023-06-15. The installation coordinates are in the format of plane number plus bracket number plus row number plus block number, such as P3-Z5-2-8, where P3 represents installation plane number 3.

[0096] In this embodiment, during retrieval, the system matches attribute fields using SQL statements, first matching the production batch number, and if they are completely identical, they are determined to be the first priority associated components;

[0097] If no match is found, the installation timestamp is matched. If they match exactly, the component is classified as a second-priority associated component. For example, if the faulty component number is confirmed to be PV20230615089 and its production batch is JCD20230503, the system will retrieve all components with production batch JCD20230503 and classify them all as first-priority associated components.

[0098] If no other components are found in the production batch JCD20230503, the module will search for the component installed on June 15, 2023, and classify it as a second-priority associated component. After the search is complete, the module will automatically generate a list of associated components, which includes the component's unique code, priority level, and a description of its association with the confirmed faulty components.

[0099] The associated components are added to the first review batch, and the scheduling management module is triggered to adjust the time window of the batch forward. At the same time, the review time points are allocated according to priority. Specifically, after the associated components are added to the first review batch, the sequence generation module automatically sorts them before the original components in the batch, and the first priority associated components are sorted before the second priority associated components. Associated components of the same priority are sorted together.

[0100] Additionally, the time window adjustment needs to first calculate the existing resource balance through the resource status monitoring unit. The calculation logic is the daily review volume of existing resources minus the number of unfinished components remaining in the first review batch. The daily review volume of existing resources is calculated according to the formula mentioned above.

[0101] If the number of associated components is less than or equal to the remaining resources, the start time of the time window will be advanced by one business day, while the end time will remain unchanged.

[0102] If the number of associated components is greater than the resource reserve, the start time will be advanced by one working day, and the end time will be postponed by the corresponding working day by rounding up the number of associated components minus the resource reserve divided by the daily review volume.

[0103] After the adjustment is completed, the scheduling management module assigns the earliest time point within the adjusted time window to the associated components, in order of first priority and second priority. Associated components of the same priority occupy consecutive detection time periods. For example, if the first round of time window after the adjustment is 9:00-17:00 one day earlier than the original start date, the 12 first-priority associated components are allocated to 9:00-12:00, the 8 second-priority associated components are allocated to 13:00-15:00, and the remaining time period is used for the review of the original components.

[0104] Through the above operations, this invention achieves coordinated operation and maintenance of fault component handling and related component troubleshooting, ensuring rapid and priority detection of related components and effectively making up for the shortcomings of existing systems that only handle individual fault components.

[0105] The specific beneficial effects and innovative features of this step are as follows: Firstly, this module does not judge related components equally. Instead, based on the core understanding that the risk of hidden crack failure due to inherent process defects is higher than the risk of subsequent installation stress, it establishes a retrieval order that prioritizes production batches over installation time. This achieves a precise match between risk level and retrieval priority, enabling the priority locking of components from the same production batch with higher risk. This solves the technical pain point of the existing system's lack of focus in related troubleshooting and greatly improves the pertinence of fault diagnosis.

[0106] Secondly, the current module directly includes related components in the first round of review and places them before the original components. It also clearly prioritizes first-priority related components over second-priority ones, adjusting the time window based on real-time resource availability to achieve flexible scheduling where higher-risk components are prioritized. This design addresses the shortcomings of existing system scheduling plans that cannot differentiate between the risks of related components, ensuring that limited operational resources are precisely allocated to high-risk components and improving resource utilization efficiency.

[0107] Thirdly, by analyzing the results of investigations of components with different priorities, the attenuation feature identification thresholds of the corresponding attribute components in the data processing module are optimized respectively, thereby improving the system's early identification capability for faults associated with different risk levels. This solves the problem that the existing system's operation and maintenance effect cannot be accurately iterated, reduces the probability of fault recurrence, and improves the long-term operational reliability of the power plant.

[0108] As one embodiment of the present invention, the dynamic optimization module is further used for:

[0109] When the number of components in the first review batch exceeds the number threshold N due to the addition of associated components, one or more of the last components in the batch are moved to the head of the second review batch, and the scheduling management module is triggered to adjust the second time window of the second review batch accordingly.

[0110] In this embodiment, the quantity threshold N needs to be determined in conjunction with the operation and maintenance resource allocation capability of the photovoltaic power station. N is equal to the maximum number of verification components that can be carried out by operation and maintenance on a single day. That is, N is equal to the minimum value among the number of operation and maintenance personnel multiplied by the maximum number of effective verification components per person per day, the number of testing equipment multiplied by the maximum number of testing components per equipment per day, and the number of transportation coverage components of vehicles that can be dispatched on a single day. N needs to be entered in advance through the system backend and can be manually adjusted according to the increase or decrease of resources.

[0111] Furthermore, the component removal rules must strictly follow the sorting priority. As mentioned earlier, the sorting order within the first round of review batches is first priority associated components, second priority associated components, and original components. Within the same priority, components are arranged in ascending order by component number. Therefore, when the number of components in the first round exceeds N after adding associated components, the removal objects must be selected from the original components that are last in the sorting. The number of components to be removed is the difference between the current total number of components in the first round and N. If the difference is 3, then the last 3 original components in the sorting are removed.

[0112] After the removal operation is completed, the components should be moved directly to the head of the next review batch, while maintaining their relative order in the first batch. For example, if the components removed from the first batch are ordered as A1, A2, and A3, they will remain A1, A2, and A3 after being moved into the second batch, and will be located before all the original components in the second batch.

[0113] Furthermore, the adjustment of the second time window for the next review batch needs to be combined with the number of components removed and the original resource configuration of the next round. The scheduling management module first calculates the existing resource balance of the next review batch, that is, the number of components that can be carried by daily operation and maintenance in the next round minus the number of original components in the next round. If the number of components removed is less than or equal to the resource balance, the start time of the second time window will be advanced by 1 working day, and the end time will remain unchanged.

[0114] If the number of components removed exceeds the remaining resources, the start time will be advanced by one business day, and the end time will be postponed by the corresponding business day based on the number of components removed minus the remaining resources divided by the daily maintenance capacity of the next round, rounded up. After the adjustment is completed, the scheduling management module must simultaneously push the first batch of component removal notifications, the second batch of component addition notifications, and the second time window adjustment notification to the maintenance management terminal to ensure that maintenance personnel receive batch change information in a timely manner.

[0115] As one embodiment of the present invention, the dynamic optimization module is further used for:

[0116] After adding related components to the first review batch, the newly added related components are merged with the original components in the first review batch. Based on the decay suspicion of the merged components, all components in the first review batch are reordered. According to the reordered order, the scheduling management module is triggered to reallocate specific review time points for all components in the first review batch within the first time window of forward adjustment.

[0117] In this embodiment, the component merging operation needs to be automatically executed after all associated components are added to the first round of review batches. The system integrates the newly added first-priority associated components, second-priority associated components and the original components from the first round into a unified component set. At the same time, it retains the basic attribute information of each component (including production batch, installation time, attenuation data, etc.) and the attenuation suspicion calculation results. The attenuation suspicion data is pre-calculated and stored by the system data processing module. The calculation logic is to combine the three core indicators of the component's power attenuation rate, current fluctuation coefficient and voltage stability over the past 30 days, and calculate them by weighting them according to the weight ratio of 4:3:3. The value range is from 0 to 1.0. The higher the value, the higher the risk of hidden crack failure.

[0118] It should also be noted that the reordering should be based on the degree of decay suspicion as the core sorting criterion, and a descending order sorting rule should be adopted, that is, the component with the higher the degree of decay suspicion value should be sorted first.

[0119] If there are components with the same decay suspicion value, then refer to the previous priority rules for further ranking. The first priority component is ranked before the second priority component, and the second priority component is ranked before the original component.

[0120] If the decay suspicion is the same and the priority is consistent, then the components are sorted in ascending order by their component numbers. For example, in the merged component set, component A (first priority, decay suspicion 0.92), component B (original component, decay suspicion 0.92), component C (second priority, decay suspicion 0.85), and component D (first priority, decay suspicion 0.85) are reordered as component A, component D, component C, and component B. This ensures that higher suspicion is prioritized while also taking into account priority differences under the same suspicion.

[0121] Furthermore, time reallocation is performed by the scheduling management module based on the adjusted first time window after the reordering is completed. First, the adjusted time window is divided into several equal detection periods by hour. The duration of a single detection period is equal to the total duration of the time window divided by the total number of components after merging. If the division is not even, the remaining duration is allocated to the detection periods of the top-ranked components. For example, if the adjusted first time window is 9:00-17:00, with a total duration of 8 hours, and there are 16 components after merging, each detection period is 0.5 hours; if there are 15 components, the first component's detection period is 1 hour, and the remaining 14 components are each 0.5 hours. Then, according to the reordered component order, detection periods are allocated sequentially starting from the time window's start time. The first-ranked component is assigned to the first period, the second-ranked component to the second period, and so on. Simultaneously, the allocation results are synchronized to the mobile terminals of maintenance personnel, clearly marking the detection time, location information, and attenuation suspicion value of each component, facilitating advance planning of detection routes by maintenance personnel.

[0122] As one embodiment of the present invention, it also includes a closed-loop management module, used to initiate a continuous tracking and monitoring process after performing maintenance or replacement operations on components with confirmed microcracks and related components confirmed to have microcrack faults through on-site verification.

[0123] Set a specific tracking monitoring cycle for the component that is shorter than the regular monitoring cycle, and continuously collect its power generation data;

[0124] If the power generation data of the component remains within the normal fluctuation range based on its rated power for a continuous preset T tracking and monitoring cycle, it is determined that its status has stabilized, the hidden danger has been eliminated, and the continuous tracking and monitoring ends, restoring it to the regular monitoring mode. Specifically, when the maintenance personnel upload the confirmation information of the repair or replacement completion (including operation type, completion time, operator number, etc.) through the field terminal, the terminal automatically sends a trigger signal to the closed-loop management module. The module synchronously retrieves the basic information of the component (including component model, rated power, installation location, historical power generation data, etc.) from the system database and automatically starts the continuous tracking and monitoring process without manual intervention, ensuring the timeliness of the process connection.

[0125] More specifically, the difference in duration between the specific tracking and monitoring cycle and the regular monitoring cycle needs to be set based on the characteristics of component failure. The regular monitoring cycle is adapted to the monitoring efficiency of large-scale components in power plants and is usually set to 7 days, that is, power generation data is collected and analyzed every 7 days.

[0126] The specific tracking and monitoring cycle needs to balance monitoring accuracy and resource consumption, and is set to 1 day, that is, the cumulative power generation data of the component in the previous 24 hours is automatically collected at a fixed time every day (such as from 0:00 to 1:00 am).

[0127] It should be further explained that the core basis for this duration setting is that if a secondary fault (such as a loose connection at the welding point or a microcrack in the new component) occurs after the repair or replacement of a component with a microcrack, its power generation fluctuation will show obvious abnormalities within 24 hours. A one-day cycle can capture this abnormality in time, and the data collection once a day will not excessively occupy the system's storage and computing resources.

[0128] The normal fluctuation range and preset number of times T need to be set in conjunction with the rated power of the component and industry standards. The normal fluctuation range is based on the theoretical daily power generation corresponding to the rated power of the component. Considering the influence of environmental factors such as light intensity and temperature, the industry usually allows a fluctuation range of ±5%. Therefore, the module sets the normal fluctuation range as "theoretical daily power generation of rated power × (1 ± 5%)", where theoretical daily power generation = rated power × local standard sunshine hours on the day (obtained in real time by the meteorological data platform connected to the system).

[0129] The preset number of times T is set to 3, meaning that the power generation data for 3 consecutive days must be within the normal fluctuation range to determine that the status is stable. This setting is based on the fact that a single data point may be affected by extreme weather (such as short-term cloudy weather or gusts of wind) and may occasionally appear normal. Three consecutive data points can eliminate interference from accidental factors and ensure the reliability of the judgment result. At the same time, the total tracking time of 3 days will not excessively extend the special monitoring cycle, thus balancing verification accuracy and operation and maintenance efficiency.

[0130] The process termination and mode switching adopts automatic execution and manual notification. When the module monitors that the data of T consecutive cycles meet the standards, it automatically generates a "Hazard Elimination Confirmation Report", which includes information such as component number, tracking duration, daily power generation data, and fluctuation range. At the same time, the monitoring mode of the component is switched from "continuous tracking" to "routine monitoring", and the component status label in the system database is updated synchronously.

[0131] In addition, the module will push status switching notifications to the operation and maintenance management terminal, reminding operation and maintenance personnel to check and confirm the report, thus combining automatic execution with manual supervision. If any power generation data exceeds the normal fluctuation range within T consecutive cycles, the module will immediately trigger an early warning, push a notification to the terminal that a secondary investigation is required, and re-include the component in the first round of review batches, initiating a new round of related investigation and on-site review process.

[0132] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A photovoltaic module monitoring system based on power generation analysis, characterized in that, include: The data acquisition module is used to acquire the operating data of multiple photovoltaic modules in the photovoltaic array; The data processing module is used to identify a set of candidate components whose starting power shows a decreasing trend based on the running data; The sequence generation module is used to sort the components in the candidate component set to generate a monitoring sequence; wherein, the sorting is based on at least the attenuation suspicion of each component; the attenuation suspicion is determined by analyzing whether there are persistent and nonlinear attenuation characteristics related to hidden crack faults in the historical power generation data of the components. The scheduling management module is used to assign differentiated first on-site verification time points to components in the candidate component set based on the monitoring sequence; the earlier the component is ranked in the monitoring sequence, the earlier the first on-site verification time point is assigned. The task execution module is used to generate on-site verification instructions for performing on-site verification of hidden cracks on the components in the candidate component set according to the allocated verification time points; The scheduling management module performs the following operations: The first N components in the monitoring sequence are included in the first round of review; the next M components are included in the second round of review; where N and M are determined based on the total amount of monitoring resources currently available. A first time window is allocated to the first round of review batches, and a second time window is allocated to the second round of review batches, wherein the end time of the first time window is earlier than the start time of the second time window; Based on the component's order within its batch, specify the exact first on-site review time point within the corresponding review time window for that batch.

2. The photovoltaic module monitoring system based on power generation analysis according to claim 1, characterized in that, The sequence generation module determines the degree of doubt about the attenuation by the functional relationship between the total power generation attenuation rate and the attenuation duration of the component within a preset historical period; The higher the total power generation attenuation rate and the longer the attenuation duration, the higher the degree of suspicion of attenuation.

3. The photovoltaic module monitoring system based on power generation analysis according to claim 1, characterized in that, The data processing module analyzes the attenuation characteristics in the following ways: Obtain power generation data of the component over multiple consecutive equal-length periods; Calculate the rate of change of power generation in each time period relative to the previous period; If there are K or more consecutive time periods in which the rate of change of power generation is negative, and the absolute value of the rate of change in the last period is greater than the absolute value of the rate of change in the first period, then it is determined that there is a continuous and nonlinear decay characteristic.

4. The photovoltaic module monitoring system based on power generation analysis according to claim 3, characterized in that, The value of K is set according to the length of the time period: when the time period is days, K≥7; when the time period is weeks, K≥3.

5. A photovoltaic module monitoring system based on power generation analysis according to claim 1, characterized in that, It also includes a dynamic optimization module, which performs the following operations when a component is confirmed to have a hidden crack fault based on on-site verification instructions: Based on the attribute information of the confirmed microcracked component, related components with at least one common hidden danger attribute are automatically retrieved in the photovoltaic array; the common hidden danger attribute includes: belonging to the same production batch, having the same installation time, or being located on the same installation plane; The associated components are added to the first batch of reviews, and the scheduling management module is triggered to adjust the time window of the batch forward. At the same time, the earliest review time point within the adjusted time window is assigned to the added associated components.

6. A photovoltaic module monitoring system based on power generation analysis according to claim 5, characterized in that, The dynamic optimization module is also used for: When the number of components in the first review batch exceeds the threshold N due to the addition of associated components, one or more of the last components in the batch are moved to the head of the second review batch, and the scheduling management module is triggered to adjust the second time window of the second review batch accordingly.

7. A photovoltaic module monitoring system based on power generation analysis according to claim 5, characterized in that, The dynamic optimization module retrieves related components in the following ways: If there are other components from the same production batch as the confirmed microcrack component, they will all be classified as first-priority associated components. Otherwise, other components that have the same installation time as the confirmed cracked component will be classified as second-priority associated components; Among them, when the first priority associated component is added to the first round of review batch, the review order assigned to it takes precedence over the second priority associated component.

8. A photovoltaic module monitoring system based on power generation analysis according to claim 6, characterized in that, The dynamic optimization module is also used for: After adding the associated components to the first review batch, the newly added associated components are merged with the original components in the first review batch. Based on the degree of attenuation suspicion of the merged components, all components in the first review batch were re-sorted; Based on the reordered sequence, the scheduling management module is triggered to reallocate specific review time points for all components in the first review batch within the first time window of the forward adjustment.

9. A photovoltaic module monitoring system based on power generation analysis according to claim 5, characterized in that, It also includes a closed-loop management module, which is used to initiate a continuous tracking and monitoring process after performing maintenance or replacement operations on components with confirmed microcracks and related components with confirmed microcracks through on-site verification. Set a specific tracking monitoring cycle for the component that is shorter than the regular monitoring cycle, and continuously collect its power generation data; If the power generation data of the component remains within the normal fluctuation range based on its rated power within a preset T consecutive tracking and monitoring cycles, it is determined that its state has stabilized, the hidden danger has been eliminated, and the continuous tracking and monitoring will end, restoring it to the normal monitoring mode.

Images