Distributed monitoring and control method for photovoltaic power station

Abstract

The invention discloses a distributed monitoring and control method for a photovoltaic power station, and belongs to the technical field of photovoltaic power station monitoring, and the method specifically comprises the following steps: S1, when any local monitoring unit monitors that parameters meet the abnormal fluctuation of a preset triggering condition, marking the local monitoring unit as an event initiation unit; s2, the event initiating unit takes other local monitoring units in a peripheral set range as associated monitoring units according to a preset topological relation graph, and sends a collaborative diagnosis activation instruction; s3, the associated monitoring unit executes enhanced data acquisition, and generates a lightweight pre-diagnosis report in combination with local historical operation mode data; s4, each associated monitoring unit sends the report to an event initiation unit, and the event initiation unit constructs a propagation feature model of abnormal fluctuation; s5, generating a preliminary disposal instruction according to the abnormal fluctuation event type output by the model, and sending the preliminary disposal instruction to the affected local monitoring unit for execution; according to the invention, the flexibility and reliability of the photovoltaic power station in a complex operation environment are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power station monitoring, and particularly relates to a distributed monitoring and control method for a photovoltaic power station. Background Art

[0002] With the acceleration of the global energy transformation, the large-scale application of photovoltaic power stations, especially distributed photovoltaic systems, has become a trend. Distributed photovoltaic power stations have the advantages of flexible power consumption and nearby power supply by dispersing photovoltaic strings in diverse scenarios such as rooftops, mountains, and water surfaces. However, they also face challenges such as wide spatial distribution of equipment, strong environmental interference, and complex electrical connections. Efficient monitoring and control are the core links to ensure the safe operation of the power station and improve power generation efficiency. The key lies in real-time perception of local anomalies, accurate judgment of the scope of influence of anomalies, and dynamic adjustment of control strategies.

[0003] Currently, the monitoring and control technology of photovoltaic power stations mainly focuses on centralized polling and fixed zoning modes. In the centralized polling mode, the central management node collects all device data at fixed intervals, judges abnormal situations based on preset thresholds. Once a single parameter exceeds the set range, an alarm is triggered and a disposal instruction is executed. It mainly relies on the central node to periodically poll and collect data, resulting in a lag in abnormal response and a lack of correlation analysis of the operating parameters of surrounding devices, making it difficult to identify the regional propagation risks caused by local anomalies. The fixed zoning mode divides the power station into multiple fixed regions according to physical location or electrical hierarchical relationships. Data interaction and collaborative control are only limited within each region, breaking the natural electrical correlation and spatial proximity between cross-regional devices, making it difficult to identify the diffusion path of anomalies between different regions, thereby affecting the fault location and disposal efficiency. Summary of the Invention

[0004] The purpose of the present invention is to provide a distributed monitoring and control method for a photovoltaic power station to solve the following technical problems: Currently, the monitoring and control technology of photovoltaic power stations mainly focuses on centralized polling and fixed zoning modes, relying on the central node to periodically poll and collect data, making it difficult to identify the diffusion path of anomalies between different regions and unable to achieve distributed collaborative diagnosis.

[0005] The purpose of the present invention can be achieved through the following technical solutions: A distributed monitoring and control method for a photovoltaic power station, where the photovoltaic power station includes a plurality of spatially dispersed photovoltaic strings and corresponding local monitoring units, and the method includes the following steps: S1. Each local monitoring unit continuously monitors the real-time operating parameters of the corresponding photovoltaic string. When any local monitoring unit monitors abnormal fluctuations of the real-time operating parameters that meet the preset trigger conditions, mark this local monitoring unit as an event initiating unit; S2. The event initiation unit uses the preset topological relationship diagram of the photovoltaic power station to take other local monitoring units within a set range around its location as associated monitoring units, and sends a collaborative diagnosis activation instruction to all associated monitoring units; S3. Each associated monitoring unit that receives the collaborative diagnosis activation instruction performs enhanced data collection, and generates a lightweight pre-diagnosis report containing device status tags and anomaly confidence levels based on the data collection results and the stored local historical operation mode data; S4. Each associated monitoring unit sends the lightweight pre-diagnosis report to the event initiation unit, and the event initiation unit constructs a propagation characteristic model of the abnormal fluctuation within the set range around based on the aggregated lightweight pre-diagnosis reports and its own abnormal fluctuation parameters; S5. Generate a preliminary disposal instruction according to the abnormal fluctuation event type output by the propagation characteristic model, and the event initiation unit sends the preliminary disposal instruction to the affected local monitoring unit for execution.

[0006] As a further solution of the present invention: in S1, the abnormal fluctuation of the preset trigger condition is specifically: The absolute value of the change rate of the real-time operation parameters exceeds the first set threshold within n consecutive set sampling periods, and the absolute value of the change rate continues to remain above the second set threshold within the subsequent n set sampling periods; Wherein, n is a natural number, there is a fixed proportional relationship between the first set threshold and the second set threshold, and the first set threshold is greater than the second set threshold; the direction of the change rate remains consistent within the n consecutive set sampling periods, and the duration of the set sampling period is adaptively determined by the electrical time constant of the photovoltaic string.

[0007] As a further solution of the present invention: in S2, the topological relationship diagram of the photovoltaic power station is specifically: During the deployment stage of the photovoltaic power station, record the physical installation coordinates of each local monitoring unit and the electrical connection path of the DC busbar branch to which it belongs; Calculate the Euclidean distance between any two local monitoring units according to the physical installation coordinates; analyze the hierarchical depth of each local monitoring unit in the DC busbar network according to the electrical connection path; fuse the Euclidean distance and the hierarchical depth to generate a topological weight coefficient; the topological weight coefficient characterizes the spatial and electrical association strength between any two local monitoring units; the topological relationship diagram is stored as a graph structure with local monitoring units as nodes and topological weight coefficients as edge attributes.

[0008] As a further solution of the present invention: in S3, the enhanced data collection and historical operation mode data are specifically: The enhanced data acquisition is to increase the sampling frequency to an integer multiple of the regular monitoring frequency within the validity period of the collaborative diagnosis activation instruction; synchronously collect the output current and output voltage parameters of the photovoltaic string, and perform time alignment processing on the collected current and voltage parameter sequences; The local historical operation mode data is the current-voltage characteristic curve cluster of the photovoltaic string under the same solar irradiance interval as the current environment, and the current-voltage characteristic curve cluster includes a standard operating curve and an allowable deviation boundary.

[0009] As a further solution of the present invention: in S3, the process of generating a lightweight pre-diagnosis report including device status tags and anomaly confidence levels is as follows: Map the current-voltage parameter sequence obtained by enhanced data acquisition to the current-voltage characteristic curve cluster space in the local historical operation mode data; Calculate the dynamic time warping distance between the current-voltage parameter sequence and the standard operating curve; assign device status tags according to whether the dynamic time warping distance exceeds the allowable deviation boundary of the characteristic curve cluster; the device status tags include in-boundary status tags, boundary fluctuation status tags, and out-of-boundary status tags; Calculate the anomaly confidence level value based on the number of times and the duration ratio of the current parameter sequence crossing the allowable deviation boundary; the lightweight pre-diagnosis report includes device status tags, anomaly confidence level values, and local monitoring unit identifiers.

[0010] As a further solution of the present invention: in S4, the process of constructing a propagation characteristic model for abnormal fluctuations within a preset range around is as follows: Extract the device status tags and anomaly confidence level values in all lightweight pre-diagnosis reports; establish a spatial coordinate system centered on the event initiation unit according to the preset photovoltaic power station topology diagram; In the spatial coordinate system, convert the anomaly confidence level value of each associated monitoring unit into a spatial density field quantity value; the spatial density field quantity value exponentially decays as the Euclidean distance between the associated monitoring unit and the event initiation unit increases; Overlay the spatial density field quantity values of all associated monitoring units to generate a spatial density distribution cloud map of abnormal fluctuations. At the same time, determine the hierarchical depth of each associated monitoring unit in the DC busbar path, and calculate the gradient slope of the anomaly confidence level value changing with the hierarchical depth; Combine the morphological characteristics of the spatial density distribution cloud map with the sign and absolute value of the gradient slope to output a set of propagation characteristic model parameters.

[0011] As a further solution of the present invention: in S5, the process of the propagation characteristic model judging the type of abnormal fluctuation event is as follows: If the spatial density distribution cloud map presents a single-peak shape and the peak is located at the coordinates of the event initiation unit, and the gradient slope is zero, it is determined as a highly localized independent device event; If the spatial density distribution cloud map presents a multi-peak shape or a continuous diffusion shape, and the gradient slope is negative, it is determined as a regional event with forward electrical propagation characteristics; If the spatial density distribution cloud map presents an irregular distribution shape, and the gradient slope is positive, it is determined as a sign of a systematic event with reverse electrical propagation characteristics; the forward electrical propagation means that the abnormal fluctuation diffuses along the normal current direction, and the reverse electrical propagation means that the abnormal fluctuation diffuses against the current direction.

[0012] As a further solution of the present invention: in the S5, the process of generating the preliminary disposal instruction is as follows: When it is determined as a highly localized independent device event, the preliminary disposal instruction only includes the output power limit operation for the photovoltaic string associated with the event initiation unit; When it is determined as a regional event with forward electrical propagation, the instruction includes the global power reduction coefficient of the busbar branch where the event initiation unit and all monitoring units associated with the out-of-bound state labels are located; When it is determined as a sign of a systematic event with reverse electrical propagation, the instruction includes the emergency disconnection command for all photovoltaic strings in the DC busbar area to which the event initiation unit belongs and the alarm trigger signal of the central management node; The thresholds of the output power limit operation, the global power reduction coefficient, and the emergency disconnection command are all dynamically calculated through the propagation characteristic model parameter set.

[0013] As a further solution of the present invention: in the S5, after the abnormal fluctuation event is disposed of, the set of actually affected local monitoring units is recorded; the overlap degree between the set of local monitoring units and the set of associated monitoring units determined according to the topology relationship diagram is calculated, and the overlap degree is calculated using the Jaccard similarity coefficient formula; According to the degree of deviation of the overlap degree from the expected threshold, the weight coefficients of the relevant edges in the topology relationship diagram are adaptively adjusted, and the adjustment range of the weight coefficient is positively correlated with the overlap degree deviation value, and the expected threshold is determined by the statistical mean of historical events.

[0014] The beneficial effects of the present invention: The present invention achieves an innovative breakthrough in the realization of a rippling collaborative diagnosis based on local abnormal events and a dynamic response path for power station topology adaptation. Through a local abnormal triggering mechanism, the present invention changes the periodic passive acquisition mode of centralized polling. When a single local monitoring unit detects abnormal fluctuations, it immediately triggers the collaborative diagnosis of surrounding associated monitoring units, significantly shortening the abnormal response time and achieving a leap from minute-level polling to second-level triggering. Relying on a dynamic topology relationship diagram that integrates physical distance and electrical hierarchy, the static barriers of fixed partitions are broken, enabling the abnormal diagnosis to be upgraded from regional isolated analysis to cross-regional spatial and electrical correlation analysis, effectively identifying the diffusion trend of abnormalities in the spatial dimension and the transmission path at the electrical hierarchy, and improving the accuracy of fault location. By adopting a lightweight pre-diagnosis report mechanism, each associated unit only transmits key information such as status tags and abnormal confidence levels, reducing data transmission redundancy. At the same time, through a hierarchical disposal strategy, power limit operations are performed for local events, gradient adjustments are implemented for regional events, and global warnings are issued for system risks, avoiding the coarseness of traditional one-size-fits-all disposals and achieving dynamic matching of the accurate classification of abnormal properties and disposal strategies. The present invention constructs a complete closed loop of triggering, collaboration, diagnosis, disposal, and optimization, comprehensively improving the distributed photovoltaic power station's real-time perception, propagation judgment, and accurate disposal capabilities for abnormal events, and enhancing the flexibility and reliability of the photovoltaic power station in a complex operating environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The present invention will be further described below with reference to the accompanying drawings.

[0016] Figure 1 It is a flowchart of the present invention. DETAILED IMPLEMENTATION MANNER

[0017] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0018] Please refer to Figure 1 As shown, the present invention is a distributed monitoring and control method for a photovoltaic power station. The photovoltaic power station includes a plurality of photovoltaic strings with dispersed positions and corresponding local monitoring units, and includes the following steps: Step S1, Abnormal Fluctuation Detection and Event Initiation Unit Marking Each local monitoring unit continuously collects real-time operating parameters of the photovoltaic string, including current, voltage, component temperature, etc. When the change rate of any parameter significantly exceeds the first-level threshold in multiple consecutive sampling cycles, and remains above the second-level threshold in the subsequent cycle and the direction is consistent, it is judged as an abnormal fluctuation, triggering the unit to become an event initiator unit to avoid misjudgment caused by environmental interference.

[0019] Step S2: Positioning of associated units based on topological relationship diagram The event initiating unit determines the collaborative scope based on the preset topological relationship diagram. The diagram integrates the physical installation coordinates of each unit with the electrical hierarchy of the DC convergence path, characterizes spatial proximity through Euclidean distance, reflects electrical association through hierarchy depth, and generates a topological weight coefficient. According to the weight or distance threshold, the associated monitoring units within a specific range of the surrounding area are screened out, such as units with a physical distance of less than 50 meters and similar electrical hierarchy, and collaborative diagnosis activation instructions are sent.

[0020] Step S3: Enhanced data collection and pre-diagnosis report generation After receiving the command, the associated monitoring unit increases the sampling frequency to several times the normal frequency, synchronously collects current and voltage parameters, and performs time alignment. The historical current-voltage characteristic curve cluster of the same irradiance interval is called, the real-time parameter sequence is compared with the historical data, and the dynamic time distance is calculated. Based on whether the parameter exceeds the boundary and the degree of deviation, a lightweight report containing normal, suspicious, abnormal status labels and abnormal confidence is generated, retaining only key diagnostic information.

[0021] Step S4: Construction of abnormal propagation feature model The event initiating unit summarizes the report and its own data, establishes a spatial coordinate system with itself as the center, converts the abnormal confidence into the spatial density field value that decays with distance, generates a spatial density distribution cloud map, and analyzes the abnormal diffusion form. At the same time, combined with the hierarchical depth of each unit in the confluence path, calculates the gradient slope of the confidence with the hierarchical change, and determines whether the abnormality propagates along the current direction or against the current direction.

[0022] Step S5: Generation and execution of hierarchical disposal instructions According to the propagation characteristic model, if it is determined to be a local independent event, only the power limit operation is performed on the strings associated with the event initiating unit; if it is a regional event with forward propagation, the power reduction is implemented on the bus branch where the event initiating unit and the associated units with abnormal status are located; if it is a systemic risk with reverse propagation, all strings in the area to which it belongs are instructed to be urgently disconnected and an alarm is sent to the central node. After handling, the system adaptively adjusts the topology weight coefficient according to the overlap between the actual affected units and the topology screening units to optimize the subsequent coordination efficiency.

[0023] In a preferred embodiment of the present invention, in S1, the abnormal fluctuation of the preset trigger condition is specifically: The key operating parameters of the photovoltaic strings, such as output current and voltage, are sampled at high frequency. The sampling period is not fixed, but is adaptively determined according to the electrical time constant of the photovoltaic strings. The electrical time constant is an important parameter that reflects the response speed of the strings in the dynamic process. For example, the electrical time constant of the strings with long cables or large-capacity capacitors is large, and the corresponding sampling period will be automatically extended to match the actual dynamic characteristics of the equipment, avoiding abnormal omissions or misjudgments due to improper sampling frequency.

[0024] For the judgment of parameter change rate, a hierarchical threshold mechanism is adopted: for example, when the absolute value of the change rate of a parameter in 5 consecutive sampling cycles suddenly exceeds the first set threshold, such as the current change rate exceeds 20% / s of the rated value, and continues to maintain above the second set threshold, such as 10% / s, in the subsequent 5 cycles, and the change direction remains consistent, that is, continuous rise or continuous decline, it is judged as an effective abnormal fluctuation. The first set threshold is significantly higher than the second set threshold, usually the former is about 2 times the latter, forming a dual judgment logic of jump-maintenance: the first set threshold is used to identify sudden abnormal changes, and the second set threshold is used to exclude transient interference. For example, when the current of the photovoltaic string continues to drop due to hidden cracks in the components, its change rate will first quickly break through the first threshold, and then maintain at the second threshold level during the fault development stage, triggering the event initiation unit mark; while the parameter fluctuations caused by the natural environment, such as the irradiance change caused by short-term cloud cover, its change rate may increase for a short time, but it is difficult to maintain above the second threshold, so it is effectively filtered.

[0025] In another preferred embodiment of the present invention, in S2, the photovoltaic power station topology diagram is specifically: During the deployment phase of the power station, the precise physical installation coordinates of each local monitoring unit are first recorded through the geographic information system, such as longitude and latitude or X / Y coordinates in the plane coordinate system. At the same time, the DC bus branch connection path of each unit is analyzed through the electrical design drawings to clarify the hierarchical relationship between the string and the combiner box and the inverter. The strings under the same combiner box belong to the same electrical level, and the combiner box close to the inverter has a higher level. For example, the string level is 1, the combiner box level to which it belongs is 2, and the connected inverter level is 3.

[0026] Based on the physical installation coordinates, the system calculates the Euclidean distance between any two local monitoring units, which intuitively reflects the spatial proximity of the devices. Adjacent units may be affected by the same environmental factors, such as local shadows and wind speed changes, or there may be a possibility of fault conduction at the physical level. For example, bracket vibration may cause poor contact between adjacent strings. At the same time, according to the electrical connection path, the hierarchical depth of each unit in the DC bus network is analyzed. The smaller the difference in hierarchical depth, the closer the units are in electrical connection, and the greater the direct impact of fault propagation. For example, a string fault will directly affect the input parameters of the busbar box, and then affect the operation of the inverter.

[0027] To generate a topological weight coefficient that can comprehensively characterize the association strength between units, the system quantitatively fuses the Euclidean distance and the hierarchical depth. The reciprocal of the Euclidean distance is used as the spatial association factor. The closer the distance, the higher the spatial association factor. The absolute value of the hierarchical depth difference is used as the electrical association factor. The closer the hierarchy, the higher the electrical association factor. Through weighted calculation, a topological weight coefficient is finally formed. For example, two units with a distance of 50 meters and a hierarchical difference of 1 have a higher weight coefficient than two units with a distance of 100 meters and a hierarchical difference of 2. The topological relationship graph is stored with local monitoring units as nodes and topological weight coefficients as edge attributes to form graph structure data. Each node can quickly query its surrounding high-weight associated nodes, providing an accurate data basis for determining the scope of collaborative diagnosis in abnormal events.

[0028] This topological relationship graph supports a dynamic update mechanism. When there are changes in the physical configuration of the power station, such as adding new strings, replacing busbar boxes, or adjusting electrical connections, the central management node will re-collect the coordinate information of relevant units, analyze the new electrical connection path, recalculate the hierarchical depth and Euclidean distance, generate updated weight coefficients, and distribute them to all local monitoring units to ensure that the topological model is always consistent with the actual power station structure. This dual-attribute modeling method breaks the static limitations of the traditional fixed partition mode, enabling the system to dynamically determine the scope of collaborative diagnosis based on the real spatial layout and electrical connection strength of the devices. For example, when an abnormality occurs in a certain string, the system will not only trigger the surrounding strings that are physically adjacent for collaborative diagnosis, but also link the electrical associated devices upstream and downstream of the same busbar branch to comprehensively capture the diffusion trend of the abnormality in the spatial dimension and the transmission path in the electrical hierarchy, effectively identifying abnormal propagation characteristics in complex scenarios such as a string fault spreading upstream to the inverter along the busbar branch or the surrounding strings being affected by mechanical vibration due to spatial proximity.

[0029] In another preferred embodiment of the present invention, in the S3, the enhanced data collection and historical operation mode data are specifically: Enhanced data acquisition is a high-frequency data acquisition mechanism executed during the effective period of the collaborative diagnosis activation instruction. Its core lies in precisely capturing the dynamic details of abnormal fluctuations by enhancing the sampling density and synchronous multi-parameter acquisition. Specifically, after receiving the collaborative diagnosis instruction, the system increases the data acquisition frequency from the original regular frequency to an integer multiple of the regular frequency. For example, it increases from once per minute to 10 times per second, enabling the capture of high-frequency change characteristics during abnormal events, such as instantaneous signals like sudden current drops or voltage jumps at the initial stage of component hot spot formation. At the same time, the system synchronously acquires the output current and output voltage parameters of the photovoltaic string because, as parameters with strong physical correlations, the synchronous change trends of current and voltage can reveal the true operating state of the device. For example, when an internal series resistance increase fault occurs in the string, the current decreases while the voltage may show non-linear changes. Through synchronous analysis, it is possible to effectively distinguish between fault types and environmental interferences. During the acquisition process, the system performs time alignment processing on the current and voltage parameter sequences, ensuring the complete synchronization of multi-parameter data in the time dimension through a unified timestamp calibration mechanism to avoid analysis misjudgments caused by sampling timing deviations.

[0030] The construction of local historical operation mode data is based on the normal operation characteristics of the photovoltaic string under the same environmental conditions. Considering that solar irradiance is a key environmental factor affecting the output characteristics of the photovoltaic string, the system classifies and stores historical operation data according to irradiance intervals, such as dividing it into multiple intervals like [0 - 200 W / m²], [200 - 400 W / m²], [400 - 600 W / m²], etc. Each interval corresponds to a cluster of current-voltage characteristic curves. Each cluster of curves contains a standard operating curve and allowable deviation boundaries: The standard operating curve is generated by statistically averaging historical normal operation data, that is, data during periods without faults and without shadow occlusion, representing the ideal operating state of the string in this irradiance interval; the allowable deviation boundaries are determined based on the standard deviation or percentile of historical data. For example, centered on the standard curve, a current fluctuation range of ±5% or a voltage fluctuation range of ±3% is extended to form a boundary interval containing normal operation fluctuations. This classification and modeling method enables the system to quickly call the matching historical mode data for real-time comparison under the current environmental irradiance, improving the accuracy and timeliness of diagnosis.

[0031] In a preferred case of this embodiment, in step S3, the process of generating a lightweight pre-diagnosis report including device status tags and abnormal confidence levels is as follows: First, map the current-voltage parameter sequence obtained by enhanced data acquisition to the corresponding historical current-voltage characteristic curve cluster space to establish the spatial mapping relationship between real-time data and historical patterns. Subsequently, calculate the distance between the real-time parameter sequence and the standard operating curve through the Dynamic Time Warping (DTW) algorithm. This algorithm can effectively match time series with different speeds or phases by elastically adjusting the time axis alignment method, and is suitable for analyzing the shape differences of current-voltage curves. If the calculated Dynamic Time Warping distance is less than the threshold of the allowable deviation boundary, it is determined that the current parameter sequence is within the boundary, and a boundary-in status label is assigned, indicating that the string is operating normally; if the distance exceeds the boundary threshold but does not completely deviate from the curve cluster range, it is determined as a boundary fluctuation status label, indicating that the operating status is suspicious and there may be early fault signs; if the distance significantly exceeds the boundary and continuously deviates from the standard curve, it is determined as a boundary-out status label, indicating that a clear abnormality has occurred in the string.

[0032] The calculation of the abnormal confidence value is based on the frequency and duration ratio of the parameter sequence crossing the allowable deviation boundary. Specifically, the system counts the number of times the parameter sequence crosses the boundary and the proportion of time outside the boundary during the enhanced data acquisition period, and generates a quantization value between 0 and 1 through weighted calculation. For example, if the parameter sequence crosses the boundary 3 times within 10 seconds and the cumulative proportion of time outside the boundary reaches 40%, the abnormal confidence can be determined to be 0.6, indicating a medium probability of abnormality. The lightweight pre-diagnosis report only contains key information such as the device status label, abnormal confidence value, local monitoring unit identifier, and timestamp, and eliminates redundant data to reduce the communication transmission load. For example, the data volume of a single report can be controlled within hundreds of bytes to ensure fast transmission and processing in a distributed network.

[0033] In another preferred embodiment of the present invention, the process of constructing the propagation characteristic model of abnormal fluctuations within a surrounding set range in S4 is as follows: First, the event initiation unit extracts key information from the lightweight pre-diagnosis reports of all associated monitoring units, including the device status label and abnormal confidence value, which reflect the response degree of each unit to the abnormal event. Subsequently, based on the preset topological relationship diagram of the photovoltaic power station, with the physical installation coordinates of the event initiation unit as the origin, a two-dimensional or three-dimensional space coordinate system is established, and the positions of each associated monitoring unit are mapped into this coordinate system to form a local space network centered on the event initiation point.

[0034] In the space coordinate system, the abnormal confidence value is converted into a spatial density field value, and this conversion process follows the principle of distance attenuation: for monitoring units closer to the event initiation unit, their abnormal confidence contributes more to the spatial density; conversely, the farther the distance, the contribution decreases exponentially. Specifically, the spatial density field value is calculated through the confidence value and the distance attenuation function. For example, for every 10-meter increase in distance, the density contribution value decays by 50%, so as to simulate the physical law that the abnormal influence weakens with the expansion of spatial distance. By superimposing the spatial density field values of all associated monitoring units, the system generates a spatial density distribution cloud map of abnormal fluctuations. This cloud map visually presents the aggregation and diffusion characteristics of the anomaly in physical space in the form of color shades or contour lines. If the cloud map shows a single-peak shape and the peak is near the coordinates of the event initiation unit, it indicates that the anomaly is highly localized; if there are multiple peaks or a continuous diffusion shape, it suggests that the anomaly may spread to the surrounding areas.

[0035] Meanwhile, the system analyzes the hierarchical depth of each associated monitoring unit in the DC busbar path. The hierarchical depth is defined as the number of nodes of the unit from the string end in the electrical link of string - junction box - inverter. For example, the string level is 1, the associated junction box level is 2, and the connected inverter level is 3. By calculating the gradient slope of the abnormal confidence value changing with the hierarchical depth, the propagation direction of the anomaly in the electrical hierarchy can be judged: if the slope is negative, it means that the confidence decreases with the increase of the level, indicating that the anomaly may propagate forward from the low-level string to the high-level equipment; if the slope is positive, the confidence increases with the increase of the level, suggesting that the anomaly propagates reversely from the inverter to the string against the current direction. The absolute value of the gradient slope reflects the propagation intensity, and the larger the absolute value, the more significant the transfer effect of the anomaly between electrical hierarchies.

[0036] Finally, the propagation feature model combines the morphological features of the spatial density distribution cloud map, such as single-peak, multi-peak, diffusion, etc., with the positive and negative signs and absolute values of the gradient slope, and outputs a parameter set including the spatial diffusion mode, electrical propagation direction, and abnormal aggregation degree. For example, if the cloud map shows a single-peak concentration and the gradient slope is zero, it means that the anomaly is confined near the event initiation unit and does not propagate upstream and downstream electrically; if the cloud map shows multi-peak diffusion and the gradient slope is negative, it indicates that the anomaly not only spreads outward in space but also propagates to high-level equipment along the current direction. These parameters provide a key basis for subsequent determination of abnormal event types and generation of disposal strategies, enabling the system to distinguish local independent events, regional propagation events, or signs of systemic risks, and avoiding the diagnostic limitations caused by traditional methods relying only on single-dimensional data.

[0037] In another preferred embodiment of the present invention, in S5, the process by which the propagation feature model determines the type of abnormal fluctuation event is as follows: When the spatial density distribution cloud diagram presents a single peak shape, and the peak value is strictly located at the coordinate position of the event initiating unit, and the gradient slope of the abnormal confidence changing with the depth of the DC confluence path level is zero, it indicates that the abnormality is limited to the photovoltaic string associated with the event initiating unit and has not spread to the surrounding space or electrical upstream and downstream. This situation is usually caused by a single equipment failure, such as hidden cracks in the internal components of the string, local poor contact, etc., and its impact range is strictly limited to the lowest level of physical location and electrical connection. For example, a string has abnormal current due to loose terminals, and the monitoring data of the surrounding adjacent strings has no significant change, and the parameters of high-level equipment such as the junction box and inverter remain normal. At this time, the gradient slope is zero, and the spatial density is concentrated at the location of the faulty string.

[0038] If the spatial density distribution cloud map presents a multi-peak shape or a continuous diffusion shape, it means that the anomaly has broken through the scope of a single device and spread to other physically adjacent strings; at the same time, the gradient slope is a negative value, indicating that the anomaly confidence decreases with the increase of the DC bus path level, that is, along the current direction, from the string to the combiner box and the inverter direction. This propagation mode usually corresponds to regional faults, such as the failure of the internal components of the combiner box causing multiple strings hanging below it to be affected at the same time, or local shadows causing power imbalance of adjacent strings and triggering a chain reaction through electrical connections. For example, the input end of a combiner box has poor contact, and the three strings hanging below it are all detected with current anomalies, and the anomaly confidence gradually decreases from the string level to the combiner box level. The spatial density cloud map shows a multi-peak distribution centered on the combiner box, which is determined to be a regional event of positive electrical propagation.

[0039] When the spatial density distribution cloud map shows an irregular distribution pattern and the gradient slope is positive, the anomaly confidence increases with the level, indicating that the anomaly may originate from high-level equipment, such as inverters and the grid side, and propagate to the lower-level strings in the opposite direction of the current. This pattern has systemic risk characteristics and may be caused by inverter failures, grid voltage fluctuations, or energy storage system anomalies. Its impact range is not limited to a specific area, but spreads to the upstream strings through electrical connections. For example, a short circuit failure on the DC side of the inverter may cause a sudden drop in the voltage of all the junction boxes connected to it, thereby affecting the normal operation of hundreds of strings. At this time, the spatial density cloud map shows that the anomaly distribution has no obvious aggregation center, and the gradient slope is positive, indicating that the anomaly propagates from the inverter to the junction box and the string in the opposite direction.

[0040] In a preferred case of this embodiment, in S5, the process of generating the preliminary treatment instruction is: For highly localized anomalies, the preliminary disposal instructions only act on the PV strings associated with the event initiation unit, and perform output power limiting operations. For example, when the current of a certain string is continuously lower than the standard value due to the hot spot effect, the system limits its maximum power output, such as limiting the output to 70% of the rated value, to avoid further development of the fault while maintaining the normal operation of other strings. The limiting threshold is dynamically calculated through the propagation characteristic model parameters, combined with the anomaly confidence level and historical operation data, to ensure that the limiting can suppress the fault without overly affecting the overall power generation.

[0041] For regional anomalies with forward electrical propagation, the instructions cover the event initiation unit and the busbars where all associated monitoring units with the status label outside the boundary are located, and implement global power reduction. The power reduction coefficient is determined according to the spatial distribution density of the anomaly confidence level and the electrical level propagation intensity. For example, when 30% of the strings in a certain busbar are marked as outside the boundary status and the spatial density cloud map shows a diffusion trend, the system may reduce the overall power of this busbar by 20% to reduce the risk of fault diffusion. At the same time, a warning is sent to the upper-level electrical node to prompt it to pay attention to the change of input parameters and make preparations for further regulation.

[0042] If it is determined as a systematic risk with reverse electrical propagation, the disposal instructions are upgraded to emergency disconnection and global warning: First, cut off the connection between all PV strings in the DC busbar area where the event initiation unit belongs and the power grid to prevent the anomaly from spreading to the power grid through the inverter; Second, send an emergency event report containing complete summary information and preliminary analysis conclusions to the central management node to trigger the full-station fault response process, such as starting the standby power supply and dispatching maintenance personnel to the site for investigation. The triggering conditions and disconnection scope of the emergency disconnection command are determined by the absolute value of the gradient slope and the degree of irregular spatial distribution in the propagation characteristic model to ensure rapid isolation of potential fault sources before the risk escalates.

[0043] The above disposal mechanism deeply binds the anomaly propagation characteristics and disposal strategies, avoiding the disadvantages of the traditional one-size-fits-all mode, preventing both over-disposal caused by local faults and under-disposal under systematic risks. At the same time, the dynamically calculated threshold mechanism enables the disposal instructions to adapt to the severity of different fault scenarios. For example, for the same localized event, the limiting amplitude of the string with an anomaly confidence level of 0.9 will be greater than that of the string with a confidence level of 0.6, reflecting the technical concept of precise control.

[0044] In another preferred embodiment of the present invention, in S5, after the abnormal event disposal is completed, record the set of actually affected local monitoring units, which is dynamically determined based on whether each unit performs operations such as power limiting and disconnection during the disposal process, and truly reflects the equipment range involved in the anomaly propagation.

[0045] Subsequently, calculate the overlap degree between the set of actually affected units and the preset set of associated monitoring units in the topological relationship diagram, and use the Jaccard similarity coefficient to quantify the matching degree between the two. Compare the overlap degree with the preset expected threshold: if the overlap degree is significantly lower than expected, it indicates that the weight coefficients of the relevant units in the topological relationship diagram fail to accurately reflect the actual association strength and need to be adjusted. The adjustment logic is as follows: Dynamically correct the weight coefficients of the relevant edges according to the overlap degree deviation value. The greater the deviation, the higher the adjustment amplitude. For example, for units that are actually affected but not included in the association set, the weight coefficients between them and the event initiation unit will be increased to enhance the relevance during subsequent screening.

[0046] The adjusted topological weight coefficients are directly applied to subsequent abnormal events, making the selection of associated monitoring units more in line with the actual propagation range. For example, if the actually affected units in a certain regional fault exceed the initial association set, the system will increase the electrical level weight or spatial proximity weight between the units in the corresponding area to avoid screening omissions in similar events. This mechanism dynamically adapts to changes in the characteristics of power station equipment, layout adjustments, or new fault modes through a closed loop of handling, feedback, and optimization, solving the lag problem of traditional static topological models, upgrading the screening of associated units from relying on a fixed logic of initial deployment to dynamic learning combined with historical experience, and continuously improving the prediction accuracy of the system for abnormal propagation and the collaborative diagnosis efficiency.

[0047] The above has described a specific embodiment of the present invention in detail, but the described content is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made according to the scope of the present invention application should still fall within the scope covered by the patent of the present invention.

Claims

1. A distributed monitoring and control method for a photovoltaic power station, the photovoltaic power station comprising a plurality of photovoltaic strings dispersed in different locations and corresponding local monitoring units, characterized in that, Including the following steps: S1. Each local monitoring unit continuously monitors the real-time operation parameters of the corresponding photovoltaic string. When any local monitoring unit detects an abnormal fluctuation of the real-time operation parameters that meets the preset trigger condition, mark this local monitoring unit as an event initiating unit; S2. The event initiating unit, according to the preset topological relationship diagram of the photovoltaic power station, takes other local monitoring units within a set range around its location as associated monitoring units, and sends a collaborative diagnosis activation instruction to all associated monitoring units; S3. Each associated monitoring unit that receives the collaborative diagnosis activation instruction performs enhanced data collection, and generates a lightweight pre-diagnosis report containing device status tags and abnormal confidence levels according to the data collection results and the stored local historical operation mode data; S4. Each associated monitoring unit sends the lightweight pre-diagnosis report to the event initiating unit. The event initiating unit constructs a propagation characteristic model of the abnormal fluctuation within the set range around based on the aggregated lightweight pre-diagnosis reports and its own abnormal fluctuation parameters; S5. Generate a preliminary disposal instruction according to the abnormal fluctuation event type output by the propagation characteristic model, and the event initiating unit sends the preliminary disposal instruction to the affected local monitoring units for execution.

2. A distributed monitoring and control method for a photovoltaic power station according to claim 1, characterized in that, In S1, the abnormal fluctuation of the preset trigger condition is specifically: The absolute value of the change rate of the real-time operation parameters exceeds the first set threshold within consecutive n set sampling periods, and the absolute value of the change rate continuously remains above the second set threshold within the subsequent n set sampling periods; Wherein, n is a natural number, there is a fixed proportional relationship between the first set threshold and the second set threshold, and the first set threshold is greater than the second set threshold; the direction of the change rate remains consistent within the consecutive n set sampling periods, and the duration of the set sampling period is adaptively determined by the electrical time constant of the photovoltaic string.

3. A distributed monitoring and control method for a photovoltaic power station according to claim 1, characterized in that, In S2, the topological relationship diagram of the photovoltaic power station is specifically: During the deployment stage of the photovoltaic power station, record the physical installation coordinates of each local monitoring unit and the electrical connection path of the affiliated DC busbar branch; Calculate the Euclidean distance between any two local monitoring units according to the physical installation coordinates; analyze the hierarchical depth of each local monitoring unit in the DC busbar network according to the electrical connection path; fuse the Euclidean distance and the hierarchical depth to generate a topological weight coefficient; the topological weight coefficient represents the spatial and electrical association strength between any two local monitoring units; the topological relationship diagram is stored as a graph structure with local monitoring units as nodes and topological weight coefficients as edge attributes.

4. A distributed monitoring and control method for a photovoltaic power station according to claim 1, wherein In S3, the enhanced data collection and historical operation mode data are specifically: The enhanced data collection is to increase the sampling frequency to an integer multiple of the normal monitoring frequency within the validity period of the collaborative diagnosis activation instruction; synchronously collect the output current and output voltage parameters of the photovoltaic string, and perform time alignment processing on the collected current and voltage parameter sequences; The local historical operation mode data is a cluster of current-voltage characteristic curves of a photovoltaic string under the same solar irradiance interval as the current environment. The cluster of current-voltage characteristic curves includes a standard operating curve and allowable deviation boundaries.

5. A distributed monitoring and control method for a photovoltaic power station according to claim 4, characterized in that, In step S3, the process of generating a lightweight pre-diagnosis report including device status tags and anomaly confidence levels is as follows: Map the current-voltage parameter sequence obtained by enhanced data acquisition to the space of the cluster of current-voltage characteristic curves in the local historical operation mode data; Calculate the dynamic time warping distance between the current-voltage parameter sequence and the standard operating curve; assign a device status tag according to whether the dynamic time warping distance exceeds the allowable deviation boundary of the characteristic curve cluster. The device status tags include in-boundary status tags, boundary fluctuation status tags, and out-of-boundary status tags; Calculate the anomaly confidence level value based on the number of times and the duration ratio of the current parameter sequence crossing the allowable deviation boundary. The lightweight pre-diagnosis report includes device status tags, anomaly confidence level values, and local monitoring unit identifiers.

6. A distributed monitoring and control method for a photovoltaic power station according to claim 1, characterized in that, In step S4, the process of constructing a propagation characteristic model for abnormal fluctuations within a surrounding set range is as follows: Extract the device status tags and anomaly confidence level values from all lightweight pre-diagnosis reports; establish a spatial coordinate system centered on the event initiation unit according to the preset photovoltaic power station topology diagram; In the spatial coordinate system, convert the anomaly confidence level value of each associated monitoring unit into a spatial density field value; The spatial density field value exponentially decays as the Euclidean distance between the associated monitoring unit and the event initiation unit increases; Superimpose the spatial density field values of all associated monitoring units to generate a spatial density distribution cloud map of abnormal fluctuations. At the same time, determine the hierarchical depth of each associated monitoring unit in the DC busbar path, and calculate the gradient slope of the anomaly confidence level value changing with the hierarchical depth; Combining the morphological characteristics of the spatial density distribution cloud map with the sign and absolute value of the gradient slope, output a set of propagation characteristic model parameters.

7. A distributed monitoring and control method for a photovoltaic power station according to claim 6, characterized in that, In step S5, the process of the propagation characteristic model determining the type of abnormal fluctuation event is as follows: If the spatial density distribution cloud map presents a single-peak morphology and the peak is located at the coordinates of the event initiation unit, and at the same time the gradient slope is zero, it is determined as a highly localized independent device event; If the spatial density distribution cloud map presents a multi-peak morphology or a continuous diffusion morphology, and the gradient slope is negative, it is determined as a regional event with a forward electrical propagation characteristic; If the spatial density distribution cloud map presents an irregular distribution morphology, and the gradient slope is positive, it is determined as a sign of a systematic event with a reverse electrical propagation characteristic. The forward electrical propagation means that the abnormal fluctuation diffuses along the normal current direction, and the reverse electrical propagation means that the abnormal fluctuation diffuses against the current direction.

8. A distributed monitoring and control method for a photovoltaic power station according to claim 7, characterized in that, In step S5, the process of generating a preliminary disposal instruction is as follows: When it is determined as a highly localized independent device event, the preliminary disposal instruction only includes the output power limiting operation for the photovoltaic string associated with the event initiation unit; When it is determined as a regional event with forward electrical propagation, the instruction includes the global power reduction coefficient of the busbar branches where the event initiation unit and all monitoring units with out-of-boundary status tags are located; When it is determined as a sign of a systematic event of reverse electrical propagation, the instruction includes an emergency disconnection command for all PV strings in the DC bus area to which the event initiation unit belongs and an alarm trigger signal for the central management node; The thresholds of the output limiting operation, the global power reduction factor, and the emergency disconnection command are all dynamically calculated through the propagation characteristic model parameter set.

9. A distributed monitoring and control method for a photovoltaic power station according to claim 1, characterized in that, In S5, after the abnormal fluctuation event is handled, the set of actually affected local monitoring units is recorded; the overlap degree between the set of local monitoring units and the set of associated monitoring units determined according to the topological relationship diagram is calculated, and the overlap degree is calculated using the Jaccard similarity coefficient formula; According to the degree of deviation of the overlap degree from the expected threshold, the weight coefficients of relevant edges in the topological relationship diagram are adaptively adjusted, and the adjustment amplitude of the weight coefficient is positively correlated with the overlap degree deviation value, and the expected threshold is determined by the statistical mean of historical events.

Images