A digital integrated operation and maintenance management system of a distributed photovoltaic power station

Abstract

The application discloses a kind of digital integration operation and maintenance management systems of distributed photovoltaic power station.It relates to photovoltaic power station operation and maintenance technical field, system includes: diagnosis excitation module is configured to apply diagnostic electric pulse to target photovoltaic group string to actively induce defect characteristics;Synchronous capture module is configured to carry out high-precision time sequence synchronization with electric pulse to capture thermal response signal;Causal decision module is configured to make decision on the authenticity of defect based on spatiotemporal homology verification;Fingerprint management module is configured to create and track the full life cycle digital fingerprint of confirmed defect.The application can filter out environmental interference, realize high-confidence diagnosis of early latent fault, and based on defect evolution trend, predictive maintenance is carried out, which improves the operation and maintenance efficiency and asset health management level of photovoltaic power station.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant operation and maintenance technology, and in particular to a digital integrated operation and maintenance management system for distributed photovoltaic power plants. Background Technology

[0002] Due to their large number and geographically dispersed nature, distributed photovoltaic (PV) power stations place extremely high demands on efficiency and accuracy in their operation and maintenance management. Currently, the mainstream operation and maintenance monitoring systems in the industry, regardless of the specific analytical algorithms they employ, are all based on the passive observation of data from PV power stations under natural operating conditions.

[0003] However, this technological paradigm, which relies entirely on passive observation, suffers from a fundamental technical bottleneck: the diagnostic signals it depends on are inherently uncontrollable, have low signal-to-noise ratios, and are random. Specifically, the manifestation of fault characteristics (whether subtle changes in electrical parameters or anomalies in infrared temperature) is heavily dependent on uncontrollable external environmental conditions such as high light levels and low wind speeds, leading to a high degree of randomness and lag in fault detection. More critically, these weak fault signals are easily obscured or confused by instantaneous environmental noise such as clouds, bird droppings, or leaves, resulting in extremely low signal-to-noise ratios.

[0004] This fundamental defect in diagnostic signals directly leads to a technical contradiction that existing solutions struggle to overcome in engineering practice: to improve diagnostic sensitivity, one must accept lower accuracy (introducing more noise); while to improve accuracy, one must reduce sensitivity and wait for longer observation times. Summary of the Invention

[0005] This invention provides a digital integrated operation and maintenance management system for distributed photovoltaic power plants to solve the fundamental technical problem in existing technologies that rely on passive observation and cannot obtain high-fidelity, controllable, and deterministic diagnostic signals.

[0006] In view of the above problems, the present invention provides a digital integrated operation and maintenance management system for distributed photovoltaic power plants, comprising:

[0007] The diagnostic excitation module is configured to apply a diagnostic electrical pulse to a target photovoltaic string in a photovoltaic power plant;

[0008] A synchronization capture module is configured to perform high-precision time synchronization with the application of the diagnostic electrical pulse in order to capture the thermal response signal excited by the diagnostic electrical pulse on the target photovoltaic string;

[0009] The causal determination module is configured to determine whether there are real physical defects in the target photovoltaic string based on the spatiotemporal consistency between the application information of the diagnostic electrical pulse and the thermal response signal.

[0010] The fingerprint management module is configured to create or update the full lifecycle digital fingerprint of defect points that are determined by the causal adjudication module to have real physical defects.

[0011] The technical solution provided in this application has at least the following technical effects or advantages:

[0012] This invention, through the method of actively applying diagnostic electrical pulses and synchronously capturing them, can identify and file physical defects at their nascent stage, solving the technical problem that existing technologies, which passively rely on data from photovoltaic modules under natural operating conditions, cannot detect early latent faults.

[0013] Since the diagnostic signal is actively generated by the system, its characteristics are greatly reduced in relation to external environmental conditions, thus maintaining a high degree of consistency in diagnostic capabilities under different operating conditions. This solves the technical problem that existing technologies are highly dependent on ideal environmental conditions, leading to the randomness and lag in diagnostic timing.

[0014] This invention establishes a strict logical connection between the "cause" of active excitation and the "effect" of physical response through a spatiotemporal causal adjudication mechanism. This effectively distinguishes between the response of real physical defects and random thermal noise caused by environmental factors, solving the technical problem of low diagnostic accuracy and rampant alarms in existing technologies due to their inability to effectively distinguish between signals and noise.

[0015] This invention constructs a digital fingerprint containing active excitation response parameters and combines it with a defect evolution prediction model to achieve quantitative prediction of the future development trend of each confirmed defect. This solves the technical problem that existing technologies can only perform ex-post, static fault diagnosis and lack the ability to perform forward-looking and dynamic management of the health status of power plant assets. Attached Figure Description

[0016] Figure 1 A schematic diagram of the overall architecture of a digital integrated operation and maintenance management system for distributed photovoltaic power plants provided by the present invention;

[0017] Figure 2 This is a schematic diagram of the workflow of the proactive predictive intelligent diagnostic engine provided by the present invention. Detailed Implementation

[0018] The above technical solutions will now be described in detail with reference to the accompanying drawings and specific embodiments to provide a better understanding of them. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. It should be understood that the present invention is not limited to the exemplary embodiments used only to explain the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention. Furthermore, it should be noted that, for ease of description, only the parts related to the present invention are shown in the drawings, not all of them.

[0019] See attached document Figure 1 This figure is a schematic diagram of the overall system architecture according to an embodiment of the present invention. This embodiment provides a digital integrated operation and maintenance management system for distributed photovoltaic power plants, including:

[0020] The diagnostic excitation module is configured to apply a diagnostic electrical pulse to a target photovoltaic string in a photovoltaic power plant;

[0021] A synchronization capture module is configured to perform high-precision time synchronization with the application of the diagnostic electrical pulse in order to capture the thermal response signal excited by the diagnostic electrical pulse on the target photovoltaic string;

[0022] The causal determination module is configured to determine whether there are real physical defects in the target photovoltaic string based on the spatiotemporal consistency between the application information of the diagnostic electrical pulse and the thermal response signal.

[0023] The fingerprint management module is configured to create or update the full lifecycle digital fingerprint of defect points that are determined by the causal adjudication module to have real physical defects.

[0024] Logically, the system can be divided into a data acquisition layer, a data processing and storage layer, and an application analysis layer.

[0025] This invention implements a proactive predictive intelligent diagnostic engine in the application analysis layer. This diagnostic engine is responsible for scheduling resources in the data acquisition and processing layer and executing all the diagnostic and predictive processes described in this invention. Internally, the diagnostic engine includes: a diagnostic excitation module for applying diagnostic electrical pulses; a synchronization capture module for synchronously capturing thermal response signals; a causal determination module for performing spatiotemporal homology verification; and a fingerprint management module for tracking the entire lifecycle of defects. Its functionality relies on the data and infrastructure support provided by the data acquisition layer and the data processing and storage layer.

[0026] To implement the diagnostic process of this invention, the system needs to be configured and a basic database needs to be built during the deployment phase.

[0027] The data acquisition layer includes a SCADA system and a drone inspection system. The SCADA system is used to collect real-time electrical data from each string inverter in the photovoltaic power station. To implement the control flow of this invention, the SCADA system has the function of receiving and executing control commands from the application analysis layer. The drone inspection system includes a drone equipped with a high-precision real-time dynamic positioning module and an infrared thermal imaging camera, and is configured with an interface that can receive external synchronous trigger signals.

[0028] The data processing and storage layer is used for persistent storage and management of the raw data acquired by the data acquisition layer. To support the diagnostic algorithm of this invention, this layer pre-constructs two databases:

[0029] One is a physical electrical mapping database. This database pre-stores the geographic coordinates of each photovoltaic module within the power plant, as well as the static mapping relationship between these geographic coordinates and the unique identifier of the module in the power plant's electrical topology. In one embodiment, this database is generated by combining the electrical topology information from the power plant design drawings with mapping data from real-time dynamic positioning equipment on-site.

[0030] Secondly, there is a component-level fault fingerprint archive, which is used to structurally store the full lifecycle status data generated in subsequent diagnostic processes and associated with each confirmed physical defect. This archive is established during system initialization and is continuously populated and updated during subsequent operation.

[0031] The following will be combined with the appendix Figure 2 This paper elaborates on the internal structure and specific workflow of the proactive predictive intelligent diagnostic engine.

[0032] Collaborative planning of diagnostic tasks: The proactive predictive intelligent diagnostic engine (hereinafter referred to as the diagnostic engine) executes a complete diagnosis, which begins with the triggering and collaborative planning of a diagnostic task.

[0033] Diagnostic tasks can be initiated manually by maintenance personnel through the user interface or automatically by the system according to a preset maintenance plan (e.g., periodic inspections of the power plant). When a task is triggered, the target area for this diagnosis needs to be clearly defined. Specifically, maintenance personnel or the system plan can specify the entire power plant, one or more specific photovoltaic arrays, or even all strings connected to a single inverter as the diagnostic object for this task. This target area information serves as the initial input to the diagnostic process and is received by the diagnostic engine.

[0034] After receiving and parsing the target area information, the diagnostic engine's planning function begins operation. It generates a precisely synchronized collaborative work plan for the drone inspection system and the grid-connected inverters within the target area. This plan generation process includes the following steps: First, the diagnostic engine retrieves the precise geographic coordinates and electrical topology of all photovoltaic modules within the target area from the physical electrical mapping database. Based on this information, the diagnostic engine calculates and generates an optimal drone flight path that fully covers all target modules. This path not only plans the drone's trajectory in three-dimensional space but also, by setting the flight speed, marks each key node on the path (especially the position where it flies over each string of photovoltaic cells) with a high-precision estimated arrival time stamp.

[0035] Simultaneously, the diagnostic engine generates a corresponding excitation sequence for the grid-connected inverters within the target area based on the UAV's flight time plan. This sequence is a structured list of instructions, each item of which defines in detail the precise time at which a diagnostic electrical pulse should be applied for a specific target string, as well as the specific parameter set of that pulse (including pulse type, width, voltage and current limits, etc.). It is important to note that the application timestamp of each pulse in the excitation sequence is strictly aligned with the expected arrival timestamp of the UAV flying over the corresponding string in its flight plan, thus laying a precise time synchronization foundation at the planning level for subsequent synchronization capture operations. This complete collaborative work plan, containing the flight path and excitation sequence, is ultimately distributed by the diagnostic engine to the UAV ground control station and the corresponding grid-connected inverters for execution.

[0036] Online implementation of diagnostic stimuli: After the collaborative work plan is issued, the diagnostic engine begins to schedule the system's hardware resources to execute diagnostic stimuli. This process is uniformly coordinated by the diagnostic engine and is achieved through deep communication with the grid-connected inverter to implement online stimuli.

[0037] The diagnostic engine first generates a structured excitation command data packet based on the collaborative work plan. This packet is not a simple trigger signal, but a data frame containing a complete execution context. Its fields include at least: the logical address of the target string, the excitation mode identifier, the pulse parameter set (including pulse width, voltage limit, and a critical dynamic current limit), and a high-precision execution timestamp synchronized with the system master clock. Subsequently, the diagnostic engine sends this command data packet to the target grid-connected inverter via industrial Ethernet using standard protocols such as Modbus TCP / IP. Upon receiving the command, the inverter's firmware verifies its content and loads it into a dedicated diagnostic task execution queue, awaiting the arrival of the execution timestamp in the command.

[0038] When the real-time clock inside the inverter precisely matches the execution timestamp in the excitation command, the inverter's power control unit takes over. At the core of this unit is a digital signal processor (DSP) that executes a dedicated diagnostic pulse generation algorithm to temporarily modify the control logic of its internal boost circuit to achieve reverse bias. Specifically, this algorithm applies energy from the inverter's DC bus to the target photovoltaic string in reverse by precisely controlling the pulse width modulation (PWM) of the power semiconductor switches in the boost circuit at a frequency of several kilohertz. To ensure that the generated pulse waveform strictly conforms to the command requirements, this process is carried out under high-speed closed-loop control. The power control unit monitors the actual reverse voltage and current applied to the string in real time with a microsecond-level sampling period and continuously compares them with the upper voltage and current limits set in the command. A PID controller dynamically and in real time adjusts the duty cycle of the PWM signal based on the error signal, thereby ensuring that the width and amplitude of the generated pulse waveform (e.g., a square wave) are precisely controlled.

[0039] The crucial dynamic current upper limit in the instruction data packet is calculated in real-time by the diagnostic engine using a dynamic safety domain constraint mechanism before issuing the instruction. This mechanism, implemented in the diagnostic engine software, is a pre-built multidimensional lookup table, essentially a digital description of the safe operating area of ​​the photovoltaic module. Before generating each excitation instruction, the diagnostic engine performs the following series of operations: First, it queries and obtains the current ambient temperature in real-time through the SCADA system; next, it sends a fast sampling instruction to the target inverter to obtain the current instantaneous open-circuit voltage of the target string; then, using these two real-time acquired ambient temperatures and open-circuit voltages as input coordinates, the diagnostic engine performs a fast interpolation calculation in the multidimensional lookup table to obtain a specific maximum permissible safe reverse excitation current value for the current instant and operating condition; finally, the diagnostic engine encapsulates this calculated current limit value into the "dynamic current upper limit" field of the excitation instruction data packet before issuing the complete instruction. Through this complete process of calculation and distribution executed by the diagnostic engine at the host computer level, it is ensured that the safety boundaries of every excitation operation performed by the inverter at the lower computer are dynamic, precise, and tailored to the current operating conditions.

[0040] Synchronous capture of diagnostic responses: While the diagnostic engine schedules the inverter to perform active excitation, the UAV inspection system must also be precisely controlled to capture the physical response signal at the same instant. This process mainly involves establishing a unified, high-precision time base covering the entire system and using this as a basis to achieve hardware-level synchronous triggering across devices.

[0041] To achieve this goal, the system first establishes a unified time reference for the entire system through a time synchronization unit deployed on a ground server. In one specific embodiment, this time synchronization unit acts as a master clock server, calibrating its own system clock to Coordinated Universal Time (UTC) via a GPS receiver. All key devices involved in collaborative diagnostics within the system, including the grid-connected inverters that execute excitation commands and the ground control station of the UAV inspection system, run client programs that implement the Precision Time Protocol (PTPv2, IEEE 1588v2 standard). These client programs dynamically calculate and compensate for network transmission latency by continuously exchanging synchronization messages with precise timestamps with the master clock server, thereby ensuring that the synchronization error between their respective internal real-time clocks and the master clock is continuously controlled within the sub-millisecond level.

[0042] After establishing a unified time base, synchronization commands are precisely transmitted to the infrared thermal imaging camera on the UAV, accomplished through a mechanism combining wireless communication and hardware triggering. Specifically, the acquisition command for the UAV in the collaborative work plan generated by the diagnostic engine also includes a high-precision execution timestamp that is completely synchronized with the excitation command. The internal clock of the UAV ground control station is synchronized with the system master clock through the aforementioned protocol, and it continuously monitors the current time. When the time reaches a preset lead time before the execution timestamp in the command, the ground control station sends a "ready to trigger" preparatory command through its telemetry data link with the UAV.

[0043] The drone is equipped with a dedicated synchronization control module, which enters standby mode upon receiving a preparatory command. At the precise execution timestamp, the ground control station sends a hardware trigger pulse signal via a dedicated, low-latency physical channel on its remote controller or data link. This pulse signal is received by the synchronization control module on the drone and directly transmitted a standard TTL level signal to the external trigger port of the infrared thermal imaging camera via a physical electrical connection. Since the infrared thermal imaging camera is pre-configured to "external trigger" mode, when its trigger port receives the rising edge of this precisely synchronized hardware trigger pulse signal from the ground, it immediately performs an image frame exposure and acquisition operation. This direct hardware-level triggering minimizes the uncertainties and delays introduced by the drone's onboard operating system and software stack, ensuring that the time error between the image capture moment and the ground master clock is kept within a very small range, meeting the engineering requirements for high-precision synchronous capture.

[0044] Causal determination for diagnostic events: After simultaneously capturing infrared thermal response images with precise timestamps and geographic coordinate metadata, the diagnostic engine initiates a causal determination process. This process aims to identify, with high confidence, genuine physical defect responses directly caused by active stimuli from the complex raw image data, while excluding all irrelevant environmental interference.

[0045] The first step in this process is spatial anchoring. Its purpose is to accurately map pixel-based thermal anomaly regions in the image to specific photovoltaic modules in the physical world of the power plant, ultimately associating them with their unique electrical topology identifier. In one specific embodiment, this process first performs temperature calibration and noise filtering on the original infrared image. Then, image processing algorithms such as adaptive thresholding are used to identify and extract pixel regions in the image whose temperature is significantly higher than the background, forming a set of one or more "thermal anomaly regions." For each identified thermal anomaly region, the system calculates the pixel coordinates of its geometric center point in the image. Then, combining the camera attitude information recorded in the image metadata, the UAV's own 3D geographic coordinates, and the camera's intrinsic parameter model, the system uses a projection transformation algorithm from the camera coordinate system to the world coordinate system to accurately convert the pixel coordinates into the 3D geographic coordinates of the thermal anomaly region's center point on the Earth's surface. Finally, using these geographic coordinates as the query keyword, the system performs a nearest neighbor search in a pre-built physical-electrical mapping database to find the database entry that is spatially closest to these coordinates and retrieves the unique electrical topology identifier of the module from it. Through this series of processes, each original thermal anomaly region is given a precise electrical identity that can be used for subsequent logical judgments.

[0046] After spatial anchoring, the diagnostic engine performs a rigorous spatiotemporal consistency check on each thermal anomaly region assigned an electrical identity. This check is a logical AND operation; a thermal anomaly must pass both the temporal synchronization check and the spatial consistency check to be finally confirmed. The temporal synchronization check is implemented by the system extracting the precise capture timestamp from the metadata of the thermal response image and comparing it with the precise timestamp of the actively applied diagnostic electrical pulse recorded in the current diagnostic task. If the absolute value of the time difference between these two timestamps is less than a preset time tolerance threshold that matches the system's synchronization accuracy, the thermal anomaly is considered synchronous in the temporal dimension. The spatial consistency check is implemented by the system performing a string-by-string match between the electrical topology identifier obtained from the spatial anchoring step of the thermal anomaly and the electrical topology identifier of the target string of the applied diagnostic electrical pulse recorded in the current diagnostic task. If they match perfectly, the thermal anomaly is considered spatially consistent. Only thermal anomalies that simultaneously meet both conditions are ultimately determined by the diagnostic engine as a "high-confidence fault event." Any thermal anomaly that fails any verification will be classified as an "excluded interference event" caused by environmental factors or other random disturbances and will be archived.

[0047] To further enhance the system's diagnostic robustness under complex weather conditions, a signal preprocessing step can be added before executing the spatiotemporal source verification logic. In one optional embodiment, this preprocessing employs a multi-scale time-window energy integration algorithm. In this mode, the UAV does not capture only a single frame image, but rather performs a high-speed image sequence acquisition within a very short time window before and after the excitation pulse. The algorithm analyzes the temperature value of each pixel in this image sequence over time. Then, using a matched filter, the algorithm focuses only on pixel regions where the energy of the temperature change curve shows a significant peak at a time scale equal to the excitation pulse width, while remaining stable at other time scales. In this way, the algorithm can accurately filter out the rapid, localized thermal response signal caused by a brief, high-frequency electrical pulse excitation from the slow, wide-ranging temperature fluctuations caused by changes in background illumination, thus providing a cleaner thermal anomaly area data for final adjudication by the subsequent spatiotemporal source verification logic.

[0048] Defect Lifecycle Health Management: After the diagnostic engine confirms and outputs a "high-confidence failure event" through a causal adjudication process, the system initiates its health management function. This function aims to establish a unique, lifelong traceable digital profile for each confirmed physical defect, and through continuous analysis of the profile data, to achieve predictive maintenance of its health status evolution trend.

[0049] To achieve precise tracking and management of each physical defect, the system first defines a standardized "digital fingerprint" data structure. This is a multi-dimensional, scalable state object used to completely describe all the static and dynamic attributes of a defect. In a specific embodiment, each digital fingerprint in the component-level fault fingerprint archive is represented as a record containing the following key fields: a system-generated, globally unique fingerprint identifier; a set of static identity information, including the electrical topology identifier to which the defect belongs, high-precision three-dimensional geographic coordinates, and its precise two-dimensional relative position on the surface of the photovoltaic module to which it belongs; a set of initial diagnosis information, including the precise timestamp of the first confirmation as a "high-confidence fault" and a snapshot of the original infrared thermal response image at that time; and a dynamic diagnostic history set, which is a time-series array where each element represents a subsequent diagnostic record and includes a diagnostic timestamp, instantaneous operating parameters, and a core set of active excitation response parameters.

[0050] When the diagnostic engine generates a new "high-confidence fault event," it immediately performs a file creation or update operation. This process first extracts the fault event's unique electrical topology identifier and geographic coordinates, using these as a composite primary key to perform a precise search in the component-level fault fingerprint archive. If the search fails, indicating a newly discovered physical defect, the system performs a creation operation, generating a new unique fingerprint identifier. The static identity information of the fault, along with the initial diagnosis information, is completely filled into the new record, and the data from this diagnosis is stored as the first record in its "dynamic diagnostic history set." If the search succeeds, indicating an already archived physical defect, the system performs an update operation, directly appending the data from this diagnosis as a new element to the end of its "dynamic diagnostic history set" array. In this way, the system ensures that all diagnostic data for the same physical defect is completely and chronologically recorded under its unique fingerprint file.

[0051] When creating or updating fingerprints and needing to populate the "active excitation response parameter set," the system extracts core dynamic parameters from the original diagnostic data that profoundly reflect the physical characteristics of the defect. In a specific embodiment, this extraction process includes: First, the system analyzes the thermal response image sequence during and within a very short time after the excitation pulse, subtracting the background average temperature of the unaffected surrounding area at the same moment from the highest temperature of the thermal anomaly region to calculate the instantaneous peak temperature rise directly caused by the excitation. This parameter directly reflects the equivalent series resistance of the defect point. Second, the system treats the applied diagnostic electrical pulse waveform as the input excitation signal and the temperature change curve of the thermal anomaly region over time as the output response signal. By performing cross-correlation analysis or Fourier transform on these two time-series signals, the system calculates the phase delay of the output response signal relative to the input excitation signal. This phase difference reflects the characteristics of heat conduction and heat capacity within the defect point. Finally, the system will pay special attention to the process of the temperature in the thermal anomaly area naturally cooling down from the peak after the diagnostic electrical pulse ends. By extracting this cooling curve and fitting it with the least squares method using the exponential decay function, the thermal decay time constant of the defect point is calculated. This constant reflects the ability of the defect point to dissipate heat to the surroundings and is related to the type and depth of the defect.

[0052] After each fingerprint update, the diagnostic engine invokes its built-in defect evolution prediction model to reassess the future risk of the defect. This process first reads all historical diagnostic records from the target fingerprint's "dynamic diagnostic history set" and extracts one or more key active stimulus response parameters (e.g., peak temperature rise). These parameters are then combined with the corresponding diagnostic timestamps to form a time-series dataset. The system then uses this complete time-series dataset as input to a pre-trained defect evolution prediction model. In one specific embodiment, this model can be a Bayesian filtering-based algorithm. Such models are particularly suitable for handling noisy, non-linear time-series data. They iteratively update their estimate of the defect state based on new data points and predict its future trends. After the model finishes running, it outputs a set of quantified prediction results. These results may include one or more defect propagation rate parameters (e.g., a value representing the "average monthly increase rate of peak temperature rise"), predictions of the potential values ​​of the key parameters of the defect in the future, and ultimately, a comprehensive assessment of the updated risk level. These predictions will be stored back in the fingerprint archive and will serve as the most important basis for subsequent operation and maintenance decisions.

[0053] In summary, this embodiment provides a technical implementation path for a digitally integrated operation and maintenance management system for distributed photovoltaic power plants.

[0054] This approach achieves proactive excitation of photovoltaic strings by planning a collaborative working sequence in the diagnostic engine and scheduling the grid-connected inverter to apply standardized diagnostic electrical pulses online.

[0055] This approach enables UAVs to synchronously capture excitation response signals by establishing a high-precision time reference at the system level and combining it with a hardware triggering mechanism.

[0056] This approach uses a causal adjudication process that includes spatial anchoring and spatiotemporal co-origin verification to logically correlate active excitation events with physical response signals, thereby identifying high-confidence fault events.

[0057] This approach also enables the management of physical defect lifecycle state data and the quantitative prediction of its future evolution trends by establishing digital fingerprints containing active excitation response parameters for confirmed fault events and combining them with defect evolution prediction models.

[0058] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details and should not be construed as limiting the invention to these specific implementations. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A digitally integrated operation and maintenance management system for distributed photovoltaic power stations, characterized in that, include: The diagnostic excitation module is configured to apply a diagnostic electrical pulse to a target photovoltaic string in a photovoltaic power plant; A synchronization capture module is configured to perform high-precision time synchronization with the application of the diagnostic electrical pulse in order to capture the thermal response signal excited by the diagnostic electrical pulse on the target photovoltaic string; The causal determination module is configured to determine whether there are real physical defects in the target photovoltaic string based on the spatiotemporal consistency between the application information of the diagnostic electrical pulse and the thermal response signal. The fingerprint management module is configured to create or update the full lifecycle digital fingerprint of defect points that are determined by the causal adjudication module to have real physical defects. The causal adjudication module includes a physical electrical mapping database that pre-stores the mapping relationship between the geographic coordinates of photovoltaic modules and their electrical topology identifiers. The causal adjudication module is configured to: first, use the physical electrical mapping database to anchor the geographic coordinates of the thermal response signal to its unique electrical topology identifier; then, determine whether the occurrence time of the thermal response signal and the application time of the diagnostic electrical pulse are synchronized within a preset time window, and determine whether the electrical topology identifier of the thermal response signal is consistent with the electrical topology identifier of the target of the diagnostic electrical pulse application, so as to adjudicate the real physical defect. The causal adjudication module is also configured to employ a multi-scale time window energy integration algorithm to verify synchronization within the preset time window, thereby suppressing environmental noise interference. The fingerprint management module also includes a defect evolution prediction model; the defect evolution prediction model is configured to predict the future evolution risk of the defect based on the time series data of the digital fingerprint.

2. The integrated digital operation and maintenance management system for distributed photovoltaic power stations as described in claim 1, characterized in that, The diagnostic excitation module applies the diagnostic electrical pulses online through a grid-connected inverter connected to the target photovoltaic string.

3. The integrated digital operation and maintenance management system for distributed photovoltaic power stations as described in claim 2, characterized in that, The diagnostic electrical pulse is a brief reverse bias diagnostic electrical pulse; and the diagnostic excitation module is configured to control the parameters of the reverse bias diagnostic electrical pulse within a preset safety range, so as to excite the thermal response signal without damaging the target photovoltaic string.

4. The integrated digital operation and maintenance management system for distributed photovoltaic power stations as described in claim 1, characterized in that, The synchronization capture module includes a drone equipped with an infrared thermal imaging camera; the system also includes a time synchronization unit configured to use a network time synchronization protocol and combine it with a hardware trigger signal to achieve high-precision time synchronization between the exposure of the infrared thermal imaging camera and the application of the diagnostic electrical pulse.

5. The integrated digital operation and maintenance management system for distributed photovoltaic power stations as described in claim 4, characterized in that, The causal determination module achieves high-precision spatial positioning of the thermal response signal at the sub-cell level through the physical electrical mapping database.

6. The integrated digital operation and maintenance management system for distributed photovoltaic power stations as described in claim 1, characterized in that, The digital fingerprint created or updated by the fingerprint management module includes at least one or more of the following active excitation response parameters: the peak temperature rise, phase difference, and thermal decay constant of the defect point in response to the diagnostic electrical pulse.

Images