A method and system for energy efficiency optimization management of photovoltaic power plants

Abstract

This invention discloses a method and system for energy efficiency optimization management of photovoltaic (PV) power plants, belonging to the field of PV power plant optimization management technology. It includes the following steps: S1, real-time acquisition of PV power plant operation data and data preprocessing; S2, identification of multi-feature mutations in equipment during PV power plant operation, identification of equipment operation anomalies, and generation of anomaly event records; S3, receiving anomaly event records, evaluating the coordinated fluctuation of various characteristic anomalies, identifying the anomaly type, and performing self-healing maintenance adjustments; S4, judging the effectiveness of self-healing maintenance adjustments, and optimizing the algorithm and self-healing strategy based on PV power plant operation data, anomaly event records, and the effectiveness of self-healing maintenance adjustments. This solves the problem that due to the increased data acquisition frequency at the station and string levels, the extremely short duration of anomaly events leads to diluted and missed anomaly signals, making timely manual response difficult.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic power plant optimization management technology, specifically to a method and system for optimizing energy efficiency management of photovoltaic power plants. Background Technology

[0002] With the continuous expansion of photovoltaic (PV) power plant installed capacity and the increasing complexity of operating systems, the number of devices, types of monitoring parameters, and data volume involved in power plant operation are constantly growing. During long-term operation, fluctuations in equipment performance, localized faults, and changes in environmental factors can all affect power generation efficiency and system stability. Therefore, refined monitoring and energy efficiency management of power plant operation status, improving equipment operating efficiency, and reducing energy losses caused by anomalies have gradually become important technical issues in PV power plant operation and maintenance management. How to achieve refined energy efficiency management of all aspects within the power plant through intelligent and digital means, maximizing PV power generation, reducing losses, and improving operational quality, has become a core concern for the industry.

[0003] For example, the invention patent with publication number CN120409903A discloses a method for evaluating the energy efficiency of a photovoltaic power station, including a data acquisition layer module, a dynamic energy efficiency model module, an energy efficiency evaluation module, and a fault location module. The dynamic energy efficiency model module establishes a theoretical power generation model based on AI prediction technology, and this model dynamically updates its parameters according to real-time environmental parameters and equipment status. The energy efficiency evaluation module calculates the actual energy efficiency and implements energy efficiency loss classification based on the theoretical power generation predicted by the dynamic energy efficiency model and the actual collected inverter output power data. The fault location module identifies abnormal strings by the string current dispersion rate and determines the location of hot spots using infrared images. This method relates to the field of photovoltaic power generation technology. By integrating dynamic environmental correction, equipment health diagnosis, and energy efficiency loss classification, it achieves precise and real-time power station energy efficiency management, effectively solving problems such as large evaluation deviations and decreased accuracy over time in existing evaluation technologies.

[0004] For example, invention patent CN116452042A discloses a method and system for intelligent IoT safety supervision of photovoltaic power plants. By real-time and stable collection and display of operational information of the main equipment in the photovoltaic power plant, it provides a comprehensive understanding of the power plant's power generation status. Specifically, it includes real-time monitoring of the following information: meteorological resource data, including irradiance, ambient temperature, module backsheet temperature, wind speed, and wind direction; and power generation data, including daily, monthly, and annual power generation, and real-time power output. This invention can enable photovoltaic power generation companies to achieve a long-term quality control mechanism from the construction phase to the operation phase. Through multiple means such as performance testing, energy efficiency analysis, fault diagnosis, operation and maintenance management, and the research and application of intelligent analysis technologies, it provides strong technical and management support for photovoltaic power generation companies. Simultaneously, problems discovered at subsequent stages can be fed back to earlier stages, shifting the control point forward and achieving a virtuous cycle of quality control, thus realizing the safe, efficient, economical, and stable operation of photovoltaic power plants.

[0005] However, as photovoltaic power plants upgrade to high-frequency, refined data acquisition, the granularity of operation monitoring at the station and string levels has significantly increased. While high-frequency data streams enrich monitoring information, traditional monitoring methods struggle to capture and accurately locate anomalies in a timely manner due to the extremely short duration of anomalies and the ease with which signals are diluted within massive amounts of normal data. This leads to an increased risk of missed anomaly detections and false alarms. Furthermore, relying on manual analysis and response not only lags behind the speed of fault development but also fails to meet the efficiency requirements of intelligent operation and maintenance for large-scale photovoltaic power plants.

[0006] Therefore, in order to address the above problems, there is an urgent need for an energy efficiency optimization management method and system for photovoltaic power plants. Summary of the Invention

[0007] Technical problems to be solved

[0008] To address the shortcomings of existing technologies, this invention provides an energy efficiency optimization management method and system for photovoltaic power plants. It solves the problems that, due to the increased data acquisition frequency at the station and string levels, the duration of abnormal events is extremely short, leading to the dilution and missed detection of abnormal signals, and the difficulty in timely follow-up by manual responses.

[0009] Technical solution

[0010] To achieve the above objectives, the present invention provides the following technical solution: an energy efficiency optimization management method for photovoltaic power plants, comprising the following steps: S1, real-time acquisition of photovoltaic power plant operation data and data preprocessing of the photovoltaic power plant operation data; S2, based on the preprocessed photovoltaic power plant operation data, determining multi-feature mutations of equipment during photovoltaic power plant operation, identifying abnormal operation of photovoltaic power plant equipment based on the multi-feature mutation determination results, and generating abnormal event records; S3, receiving abnormal event records, evaluating the coordinated fluctuation of each feature abnormality based on photovoltaic power plant operation data, identifying the abnormality type based on the coordinated fluctuation of each feature abnormality, and performing self-healing maintenance adjustment based on the abnormality type; S4, combining photovoltaic power plant operation data before and after self-healing maintenance adjustment, determining the self-healing maintenance adjustment effect, and optimizing the algorithm and self-healing strategy based on photovoltaic power plant operation data, abnormal event records, and self-healing maintenance adjustment effect.

[0011] Furthermore, the specific process of real-time acquisition and preprocessing of photovoltaic power plant operation data is as follows: Real-time, high-frequency acquisition of photovoltaic power plant operation data from various core devices at multiple levels, including current, voltage, temperature, power, and power factor; preliminary noise reduction using median filtering, followed by outlier detection and signal interpolation repair; time-series alignment, sliding window normalization, and standardization processing of the photovoltaic power plant operation data; identification and removal of invalid signals through drift rate analysis and sliding standard deviation calculation; simultaneous, primary / backup switching of the data acquisition link based on data stability results; and the establishment of a photovoltaic power plant operation database to store the photovoltaic power plant operation data.

[0012] Furthermore, based on the preprocessed photovoltaic power station operation data, the specific process for determining multi-feature mutations in the photovoltaic power station equipment during operation is as follows: Acquire the current, voltage, and temperature data of the equipment during photovoltaic power station operation; Calculate the mean and standard deviation of the current, voltage, and temperature data respectively based on a sliding time window, obtaining the mean current, standard deviation of current, mean voltage, standard deviation of voltage, mean temperature, and standard deviation of temperature; Divide the difference between the current and the mean current at the current moment by the sum of the standard deviation of current and the minimum constant value, and then square the result to obtain the current normalization deviation value; Divide the difference between the voltage and the mean voltage at the current moment by the sum of the standard deviation of voltage and the minimum constant value, and then square the result to obtain the voltage normalization deviation value; Divide the difference between the temperature and the mean temperature at the current moment by the sum of the standard deviation of temperature and the minimum constant value, and then square the result to obtain the temperature normalization deviation value; Add the current normalization deviation value, the voltage normalization deviation value, and the temperature normalization deviation value together and take the square root to obtain the transient multi-parameter mutation intensity factor.

[0013] Furthermore, the specific process for identifying abnormal operation of photovoltaic power station equipment and generating abnormal event records based on the multi-feature mutation judgment results is as follows: The transient multi-parameter mutation intensity factor of each device is calculated in real time and written into the photovoltaic power station operation database. When the transient multi-parameter mutation intensity factor is greater than or equal to the abnormal threshold, it is determined as a local abnormal event. The current abnormal time, abnormal device, corresponding current, voltage, and temperature data, as well as the mean and standard deviation of the corresponding data, are recorded to generate an abnormal event record. This abnormal event record is then pushed to the next process. When the transient multi-parameter mutation intensity factor is less than the abnormal threshold, photovoltaic power station operation data and transient multi-parameter mutation intensity factors are continuously collected and calculated. Using historical photovoltaic power station operation data and transient multi-parameter mutation intensity factors, the mean, standard deviation, and abnormal threshold are updated using a sliding window adaptive statistical method and an exponential weighted moving average method.

[0014] Furthermore, the specific process of receiving abnormal event records and evaluating the coordinated fluctuations of various characteristic anomalies based on photovoltaic power plant operation data is as follows: Abnormal event records are received, specific abnormal equipment and time periods are identified, and based on a sliding time window, the current, power, and temperature data of the equipment during the abnormal period are acquired in real time. Simultaneously, the current, power, and temperature data for the previous d sampling intervals are obtained from the photovoltaic power plant operation database. The power change is obtained by subtracting the power from the power of the previous d sampling intervals at the current moment, and the relative power change rate is obtained by dividing the power change by the power of the previous d sampling intervals. The current at the current moment is subtracted from the current of the previous d sampling intervals. To obtain the current change, divide the current change by the current at the previous d sampling intervals to get the relative rate of change of current; subtract the temperature at the current time from the temperature at the previous d sampling intervals to get the temperature change, and divide the temperature change by the temperature at the previous d sampling intervals to get the relative rate of change of temperature; add the relative rate of change of current and the relative rate of change of temperature to the minimum constant value to get the combined rate of change of temperature and current; divide the relative rate of change of power by the combined rate of change of temperature and current to get the power-thermal-electric integrated change ratio; based on the sliding time window length, accumulate and average the power-thermal-electric integrated change ratios at all times within the window, and take the absolute value to get the power-thermal-electric coupling response ratio.

[0015] Furthermore, the specific process for identifying anomaly types based on the coordinated fluctuations of various characteristic anomalies is as follows: Based on the power thermoelectric coupling response ratio and its changing trend, the anomaly type is determined by jointly analyzing the relative change rates of power, temperature, and current: When the power thermoelectric coupling response ratio is greater than the secondary coupling exceedance threshold, and the relative change rate of temperature is less than or equal to the temperature change threshold and the relative change rate of current is less than or equal to the current change threshold, it is identified as a shading mismatch anomaly; when the power thermoelectric coupling response ratio is greater than the primary coupling exceedance threshold and less than or equal to the secondary coupling exceedance threshold, and the relative change rate of temperature is greater than the temperature change threshold while the relative change rate of current is less than or equal to the current change threshold, it is identified as a thermal runaway anomaly; when the power thermoelectric coupling response ratio is less than or equal to the primary coupling exceedance threshold, and the relative change rate of temperature is greater than the temperature change threshold and the relative change rate of current is greater than the current change threshold, it is identified as a current disturbance anomaly.

[0016] Furthermore, the specific process of self-healing maintenance adjustment based on the anomaly type is as follows: For shading mismatch anomalies, identify the shading mismatched strings, temporarily isolate the abnormal strings through intelligent switches, restart them after a delay, and check the power recovery status; if the anomaly persists after isolation and restart, control the bypass branch to ensure stable power output of the main line; for thermal runaway anomalies, issue derating commands to the equipment with abnormal temperatures, lower the inverter power setpoint, adjust the string output, and reduce the heat load; activate the liquid cooling heat dissipation equipment for local rapid cooling; for current disturbance anomalies, identify the fault point, realize loop switching through the control module, separate the faulty loop, and redistribute the load to the preset redundant power supply line to reduce the impact on the main grid and other branches; link temperature control measures with power reduction to achieve multi-dimensional joint protection; generate a fault maintenance work order, including: power thermoelectric coupling response ratio and corresponding power, temperature and current data, as well as the corresponding self-healing maintenance adjustment process, and synchronously write the fault maintenance work order into the photovoltaic power station operation database.

[0017] Furthermore, combining the photovoltaic power plant operation data before and after self-healing maintenance adjustment, the specific process for judging the effect of self-healing maintenance adjustment is as follows: From the photovoltaic power plant operation database, obtain the power, power factor, and temperature data of the equipment within a window during the abnormal period before self-healing maintenance adjustment; simultaneously obtain the power, power factor, and temperature data of the equipment within the same window length after self-healing maintenance adjustment; based on the sliding time window length, calculate the difference between the power after self-healing maintenance adjustment and the power before self-healing maintenance adjustment at each moment to obtain the power difference value, and sum and average the power differences at all moments to obtain the average power difference value; for each At each time step, the power factor ratio is obtained by dividing the power factor after self-healing maintenance adjustment by the sum of the power factor before self-healing maintenance adjustment and the minimum constant value. The power factor ratios at all times are then summed and averaged to obtain the mean power factor ratio. The mean power difference is multiplied by the mean power factor ratio to obtain the comprehensive self-healing efficiency numerator. Simultaneously, the temperature change value is obtained by calculating the difference between the temperature after self-healing maintenance adjustment and the temperature before self-healing maintenance adjustment. The standard deviation of the temperature change values ​​at all times is calculated and added to the constant to obtain the temperature recovery fluctuation factor. The comprehensive self-healing gain value is obtained by dividing the comprehensive self-healing efficiency numerator by the temperature recovery fluctuation factor.

[0018] Furthermore, based on the photovoltaic power plant's operating data, abnormal event records, and the self-healing maintenance adjustment effect, the specific process for optimizing the algorithm and self-healing strategy is as follows: The comprehensive self-healing gain value is compared with the self-healing threshold. When the comprehensive self-healing gain value is less than the self-healing threshold, the main characteristics of this self-healing maintenance adjustment failure are analyzed, and the self-healing maintenance adjustment parameters are adjusted accordingly, followed by a second self-healing maintenance adjustment. If the initial self-healing maintenance adjustment only adopts limited and mild measures such as parameter fine-tuning, power limiting, and derating of some equipment, the second adjustment will employ enhanced intervention measures such as larger-scale parameter adjustments, wider-range load redistribution, string isolation, loop switching, and forced shutdown reset. If the comprehensive self-healing gain value still fails to meet the standard after multiple self-healing maintenance adjustments, a high-priority work order is generated, and the abnormal trend is pushed to the relevant authorities. The system analyzes potential and generates historical self-healing reports to remind maintenance personnel to conduct follow-up checks. When the comprehensive self-healing gain value is greater than or equal to the self-healing threshold, it continuously writes historical self-healing cases and comprehensive self-healing gain values ​​into the photovoltaic power plant operation database. Statistical analysis is performed on historical self-healing cases and comprehensive self-healing gain values ​​to optimize the selection of self-healing maintenance and adjustment strategies. The entire process of photovoltaic power plant operation data and comprehensive self-healing gain values ​​before and after self-healing maintenance and adjustment is visualized in multiple dimensions, and health classification and trend analysis of batch equipment are performed to generate self-healing effectiveness reports. Historical self-healing maintenance and adjustment data are regularly used to mine abnormal evolution patterns and identify the optimal self-healing maintenance and adjustment strategies through data mining, cluster analysis, pattern recognition, and parameter regression algorithms. The system continuously optimizes and updates various algorithms, thresholds, and self-healing maintenance and adjustment parameters.

[0019] The second aspect of this invention provides an energy efficiency optimization management system for photovoltaic power plants, comprising: a multi-level high-frequency data acquisition and preprocessing module for real-time acquisition of photovoltaic power plant operation data and preprocessing the photovoltaic power plant operation data; an anomaly signal detection and multi-scale analysis module for judging multi-feature mutations of equipment during photovoltaic power plant operation based on the preprocessed photovoltaic power plant operation data, identifying photovoltaic power plant equipment operation anomalies based on the multi-feature mutation judgment results, and generating anomaly event records; a local anomaly attribution and self-healing maintenance module for receiving anomaly event records, evaluating the coordinated fluctuation of various feature anomalies based on photovoltaic power plant operation data, identifying anomaly types based on the coordinated fluctuation of various feature anomalies, and performing self-healing maintenance adjustments based on the anomaly types; and a self-healing feedback and optimization management module for judging the self-healing maintenance adjustment effect by combining photovoltaic power plant operation data before and after self-healing maintenance adjustment, and optimizing algorithms and self-healing strategies based on photovoltaic power plant operation data, anomaly event records, and self-healing maintenance adjustment effects.

[0020] Beneficial effects

[0021] The present invention has the following beneficial effects:

[0022] (1) This invention acquires photovoltaic power station operation data in real time at multiple levels, and combines median filtering, outlier detection, signal interpolation repair, drift rate analysis and sliding standard deviation processing to effectively eliminate invalid signals and intelligently switch between main and backup channels, which greatly improves data quality and the timeliness and stability of perception.

[0023] (2) This invention constructs a transient multi-parameter mutation intensity factor to perform standardized fusion analysis of multiple physical quantities such as current, voltage and temperature, thereby achieving millisecond-level accurate identification of local abnormal events in high-frequency data streams and improving the sensitivity and accuracy of anomaly detection.

[0024] (3) Based on the power thermoelectric coupling response ratio and the rate of change of multiple parameters, the present invention intelligently classifies the abnormal types and realizes the self-healing control of the abnormal types of shielding mismatch, thermal runaway and current disturbance through a variety of means such as intelligent switching, derating command, liquid cooling heat dissipation and soft switching, which greatly improves the fault handling and recovery capabilities.

[0025] (4) This invention quantifies the effectiveness of self-healing maintenance by comprehensively evaluating the self-healing gain value, and combines sliding time window and multi-dimensional data comparison to realize the visualization analysis, health classification and trend tracking of the entire process of self-healing maintenance adjustment, and continuously improves the level of self-healing strategy and operation and maintenance intelligence based on pattern recognition and parameter optimization of historical data.

[0026] Of course, any product implementing this invention does not necessarily need to achieve all of the advantages described above at the same time. Attached Figure Description

[0027] Figure 1 A flowchart illustrating an energy efficiency optimization management method for photovoltaic power plants;

[0028] Figure 2 This is a structural diagram of an energy efficiency optimization management system for photovoltaic power plants;

[0029] Figure 3 A trend chart of transient multi-parameter energy mutation factors for photovoltaic power plants;

[0030] Figure 4 A bar chart comparing the overall self-healing gain values ​​of each device. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. As those skilled in the art will understand, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Please see Figures 1-4 This invention provides a technical solution: a method and system for energy efficiency optimization management of photovoltaic power plants, such as... Figure 1 As shown, the process includes the following steps: S1, real-time acquisition of photovoltaic power station operation data and data preprocessing of the photovoltaic power station operation data; S2, based on the preprocessed photovoltaic power station operation data, determining multi-feature mutations of the equipment during photovoltaic power station operation, identifying abnormal operation of photovoltaic power station equipment based on the multi-feature mutation judgment results, and generating abnormal event records; S3, receiving abnormal event records, evaluating the coordinated fluctuation of each feature abnormality based on the photovoltaic power station operation data, identifying the abnormality type based on the coordinated fluctuation of each feature abnormality, and performing self-healing maintenance adjustment according to the abnormality type; S4, combining the photovoltaic power station operation data before and after self-healing maintenance adjustment, determining the self-healing maintenance adjustment effect, and optimizing the algorithm and self-healing strategy based on the photovoltaic power station operation data, abnormal event records, and self-healing maintenance adjustment effect.

[0033] Specifically, the real-time acquisition and preprocessing of photovoltaic power plant operation data involves the following steps: Multi-level, real-time, high-frequency acquisition of photovoltaic power plant operation data from various core devices. This multi-level acquisition refers to continuously acquiring operational status data at high sampling frequencies at the station level, string level, and other different device levels, ensuring the timeliness and completeness of monitoring. The photovoltaic power plant operation data includes current, voltage, temperature, power, and power factor. Current and voltage are fundamental parameters reflecting the health of electrical circuits and load status; temperature can be used to monitor equipment overheating and localized hotspots; and power and power factor directly reflect power generation capacity and power quality. Median filtering is used for initial noise reduction of the photovoltaic power plant operation data, followed by outlier detection and signal interpolation repair. Outlier detection identifies extreme anomalies and false alarms from sensors, while signal interpolation repair uses mathematical interpolation algorithms to supplement lost and abnormal data, ensuring temporal continuity. The photovoltaic power plant operation data undergoes time-series alignment, sliding window normalization, and standardization. Drift rate analysis and sliding standard deviation calculations are performed on the data to identify and remove invalid signals. Drift rate analysis determines whether there is a long-term trend shift in sensor data, while sliding standard deviation calculation measures the degree of short-term data fluctuation. The combination of these two methods effectively identifies and removes invalid signals caused by sensor failures and channel drift. Simultaneously, based on data stability results, a primary / backup switching mechanism is implemented for the data acquisition link. When abnormal or unstable signals are detected in the primary channel, the system switches to a pre-set backup acquisition channel, improving the redundancy and reliability of the data acquisition link. A photovoltaic power plant operation database is established to store the operation data, providing unified data support and a management foundation for subsequent anomaly detection, self-healing decision-making, and historical analysis.

[0034] This implementation scheme effectively improves the accuracy and timeliness of raw data by acquiring photovoltaic power plant operation data in real time at multiple levels and combining various data preprocessing techniques such as median filtering, outlier detection, signal interpolation, time series alignment, sliding window normalization, and standardization. Through drift rate analysis and sliding standard deviation calculation, invalid signals can be identified and eliminated in a timely manner. Simultaneously, the primary and backup channels are automatically switched based on data stability to ensure the continuity and reliability of data acquisition. Finally, all processed data is uniformly stored in the photovoltaic power plant operation database, providing a solid data foundation for subsequent anomaly detection and intelligent operation and maintenance.

[0035] Specifically, based on the preprocessed photovoltaic power station operation data, the specific process for judging multi-characteristic sudden changes in equipment during photovoltaic power station operation is as follows: Obtain the current, voltage, and temperature data of the equipment during photovoltaic power station operation; based on a sliding time window, calculate the mean and standard deviation of the current, voltage, and temperature data respectively, obtaining the mean current, standard deviation current, mean voltage, standard deviation voltage, mean temperature, and standard deviation temperature; the mean reflects the average level of the parameter within the window, and the standard deviation measures the fluctuation range of the parameter; both are used together to measure whether the current data deviates from the normal range. The difference between the current and the mean current at the current moment is divided by the sum of the standard deviation current and the minimum constant value, and then squared to obtain the current normalization deviation value; the difference between the voltage and the mean voltage at the current moment is divided by the sum of the standard deviation voltage and the minimum constant value, and then squared to obtain the voltage normalization deviation value; the difference between the temperature and the mean temperature at the current moment is divided by the sum of the standard deviation temperature and the minimum constant value, and then squared to obtain the temperature normalization deviation value; where the minimum constant value is 0.001, used to prevent numerical instability caused by a zero denominator and a minimum value. The transient multi-parameter mutation intensity factor is obtained by adding the normalized deviation values ​​of current, voltage, and temperature, and taking the square root. This factor comprehensively reflects the degree of abnormal mutation of multiple characteristic parameters at the same time. It is a key indicator for realizing joint anomaly detection of multi-dimensional parameters and can be used to quantitatively identify instantaneous mutations and abnormal states in equipment operation.

[0036] The specific formula for the transient multi-parameter mutation intensity factor is as follows:

[0037] ;

[0038] In the formula, This represents the transient multi-parameter energy mutation factor at the current moment, used for anomaly detection in the high-frequency data stream of photovoltaic power plants. By standardizing and comprehensively aggregating the deviations of three physical quantities—current, voltage, and temperature—from the local window mean at the current moment, it characterizes the intensity of the coordinated mutation of multiple physical quantities at this moment. The larger the value, the more significantly these feature data deviate from the normal fluctuation range at the current moment, and the easier it is to judge them as abnormal; The current at the current moment reflects the actual current flow at the sampling points such as strings, junction boxes and inverters at the current moment, and is an important basis for detecting whether the circuit is smooth, load changes and string faults. This represents the average current, which serves as the normal baseline for the current in the current window, making it easier to determine deviations. It represents the standard deviation of the current, reflecting the recent fluctuation range of the current, and provides a scale normalization benchmark for anomaly detection; It represents the voltage at the current moment, reflecting the voltage level at the current sampling point, and is an important signal for monitoring line health, electrical contact reliability, and arc risk; This represents the average voltage, which serves as the baseline for normal voltage operation during the current time period and provides a reference for deviation identification. It represents the voltage standard deviation, measures the voltage fluctuation level within the current window, and provides a unified standard for anomaly detection in different scenarios; It indicates the temperature at the previous moment, reflecting the current thermal state of each device node, and is an important signal for detecting overheating, thermal runaway, and environmental anomalies. It represents the average temperature, reflecting the temperature level over a normal period, and provides a benchmark for judging current temperature anomalies; It represents the standard deviation of temperature, measures the normal fluctuation range of recent temperature, and provides a standardized basis for temperature anomaly detection; This represents a very small constant value, with a value of 0.001. It represents the current normalization deviation value, which measures the degree of deviation of the current from the recent average level at the current moment, reflects the strength of the current abnormal fluctuation, and is used to identify short-term drastic changes in current and eliminate the impact of different current baselines in different strings. It represents the voltage normalization deviation value, measures the degree of deviation of the current voltage from the recent average level, reflects the strength of the current voltage abnormal fluctuation, identifies instantaneous voltage abnormalities such as poor contact and arcing problems, and eliminates misjudgments caused by different voltage operating ranges. It represents the temperature normalization deviation value, which measures the degree of deviation of the current temperature from the recent average level, reflects the strength of the current abnormal temperature fluctuation, and quickly detects temperature anomalies such as junction box overheating and local hot spots in components, thereby improving the sensitivity of detecting potential hazards such as thermal runaway.

[0039] In this embodiment, Table 1 is a data table of transient multi-parameter energy mutation factors. It records in detail the current, mean current, standard deviation of current, voltage, mean voltage, standard deviation of voltage, temperature, mean temperature, standard deviation of temperature, and transient multi-parameter energy mutation factors at five time points. At time 1, the current is 8.2, the mean current is 8.0, the standard deviation of the current is 0.10, the voltage is 520, the mean voltage is 518, the standard deviation of the voltage is 2.0, the temperature is 39.0, the mean temperature is 38.7, the standard deviation of the temperature is 0.20, and the transient multi-parameter energy mutation factor is 2.674; at time 2, the current is 8.6, the mean current is 8.2, the standard deviation of the current is 0.15, the voltage is 525, the mean voltage is 520, the standard deviation of the voltage is 2.2, the temperature is 41.5, the mean temperature is 39.2, the standard deviation of the temperature is 0.23, and the transient multi-parameter energy mutation factor is 10.551; at time 3, the current is 7.9, the mean current is 8.0, the standard deviation of the current is 0.12, the voltage is 516, the mean voltage is 518, and the voltage... The standard deviation is 2.1, the temperature is 37.8, the mean temperature is 38.5, the standard deviation of temperature is 0.21, and the transient multi-parameter energy mutation factor is 3.549; the current at time 4 is 9.0, the mean current is 8.5, the standard deviation of current is 0.17, the voltage is 529, the mean voltage is 524, the standard deviation of voltage is 2.4, the temperature is 43.2, the mean temperature is 40.1, the standard deviation of temperature is 0.25, and the transient multi-parameter energy mutation factor is 12.862; the current at time 5 is 8.1, the mean current is 8.0, the standard deviation of current is 0.11, the voltage is 517, the mean voltage is 518, the standard deviation of voltage is 2.0, the temperature is 38.2, the mean temperature is 38.7, the standard deviation of temperature is 0.20, and the transient multi-parameter energy mutation factor is 2.692.

[0040] Table 1 Transient Multiparameter Energy Mutation Factor Data Table

[0041]

[0042] like Figure 3 The figure shows the trend of transient multi-parameter energy mutation factor for photovoltaic power plants. It illustrates the changing trends of the transient multi-parameter mutation intensity factor at five time points, with a dashed line indicating an anomaly threshold of 12. The broken line represents the transient multi-parameter energy mutation factor after integrating the normalized deviations of multiple physical quantities such as current, voltage, and temperature at each time point, reflecting the degree of mutation in the equipment's operating state at each time point. Data points in the figure are connected by solid black lines to highlight the changes in mutation intensity between time points. Based on Table 1 and... Figure 3As can be seen, the transient multi-parameter energy mutation factor at time 4 reached 12.862, significantly exceeding the abnormal threshold, indicating a significant abnormal change in the equipment's operating state at time 4, requiring close monitoring and timely intervention. The mutation factors at other times were all below the threshold, especially at times 1, 3, and 5, indicating that the equipment's operating state at these times was basically within the normal fluctuation range, without any obvious abnormalities. The figure clearly shows that the transient multi-parameter energy mutation factor fluctuated dramatically at different times, especially showing significant peaks at times 2 and 4, while remaining relatively low at other times, highlighting the sensitivity to transient anomalies in multiple physical quantities and the ability to respond in real time.

[0043] In this implementation plan, by analyzing multi-feature data of current, voltage, and temperature from photovoltaic power plant equipment during operation, the mean and standard deviation are calculated in real time based on a sliding time window. A standardized deviation processing method is then employed to accurately capture instantaneous abnormal fluctuations in each feature quantity. Through multi-feature normalized deviation fusion, a transient multi-parameter mutation intensity factor is formed. This not only effectively avoids anomalies in single parameters but also improves the sensitivity and accuracy of detecting collaborative anomalies in multiple physical quantities. This provides a more scientific and reliable quantitative basis for subsequent equipment condition assessment and rapid anomaly early warning, enhancing the real-time perception and accurate identification capabilities of photovoltaic power plant operational anomalies.

[0044] Specifically, the process of identifying abnormal operation of photovoltaic power station equipment and generating abnormal event records based on multi-feature mutation judgment results is as follows: The transient multi-parameter mutation intensity factor of each device is calculated in real time and written into the photovoltaic power station operation database. When the transient multi-parameter mutation intensity factor is greater than or equal to the abnormal threshold, it is determined as a local abnormal event. The current abnormal time, abnormal device, corresponding current, voltage, and temperature data, as well as the mean and standard deviation of the corresponding data, are recorded to generate an abnormal event record. The abnormal threshold is a judgment baseline set based on historical data statistical analysis, used to distinguish between normal fluctuations and abnormal mutations. A local abnormal event refers to a short-term, drastic deviation of equipment parameters within a certain time window, which may indicate equipment failure, abnormal environment, or sudden changes in operating status. The abnormal event record is then pushed to the next process. When the transient multi-parameter mutation intensity factor is less than the abnormal threshold, photovoltaic power station operation data and transient multi-parameter mutation intensity factors are continuously collected and calculated. Using historical photovoltaic power station operation data and transient multi-parameter mutation intensity factors, the mean, standard deviation, and abnormal threshold are updated using a sliding window adaptive statistical method and an exponentially weighted moving average method. Among them, the sliding window adaptive statistical method is a dynamic calculation method that can automatically adjust statistics according to changes in data within the window, improving the algorithm's flexibility and anti-interference ability. The exponentially weighted moving average method is a commonly used time-series data smoothing technique that assigns higher weights to the latest data, smoothing short-term fluctuations and effectively improving the timeliness and adaptability of outlier thresholds and algorithm parameters.

[0045] This implementation scheme achieves highly sensitive automatic identification of photovoltaic power plant operational anomalies through real-time calculation and dynamic monitoring of transient multi-parameter mutation intensity factors for each device. Comparing the mutation intensity factors with anomaly thresholds allows for timely identification and recording of detailed information about abnormal events, providing accurate data for subsequent self-healing maintenance. Simultaneously, high-frequency data acquisition and mutation factor calculation are continuously performed when no anomalies are detected, constantly optimizing the mean, standard deviation, and threshold settings to ensure the adaptability and accuracy of anomaly detection. Overall, this improves the real-time performance and reliability of anomaly identification.

[0046] Specifically, the process of receiving abnormal event records and assessing the coordinated fluctuations of various characteristic anomalies based on photovoltaic power plant operation data is as follows: Abnormal event records are received, specific abnormal equipment and time periods are identified, and based on a sliding time window, the current, power, and temperature data of the equipment during the abnormal period are acquired in real time. Simultaneously, the current, power, and temperature data for the previous d sampling intervals are retrieved from the photovoltaic power plant operation database; 'd' represents the number of sampling intervals, set according to the actual acquisition frequency and monitoring requirements, used to compare the changing trends of the current and historical states. The power change is obtained by subtracting the power of the previous d sampling intervals from the current power, and the relative rate of change of power is obtained by dividing the power change by the power of the previous d sampling intervals. Similarly, the current change is obtained by subtracting the current of the previous d sampling intervals from the current current, and the relative rate of change of current is obtained by dividing the current change by the current of the previous d sampling intervals. Likewise, the temperature change is obtained by subtracting the temperature of the previous d sampling intervals from the current temperature, and the relative rate of change of temperature is obtained by dividing the temperature change by the temperature of the previous d sampling intervals. The relative rate of change measures the dynamic deviation of parameters from historical benchmarks during the abnormal period, reflecting the severity of the suddenness of the anomaly. The relative rate of change of current and the relative rate of change of temperature are added to the minimum constant value to obtain the composite rate of change of temperature and current. The composite rate of change of temperature and current comprehensively evaluates the synchronous fluctuation characteristics of electrical and thermal parameters. The power-thermal-electrical integrated mutation ratio is obtained by dividing the relative rate of change of power by the composite rate of change of temperature and current. The power-thermal-electrical integrated mutation ratio is used to describe whether the power mutation has obtained a coordinated response of temperature and current, and is the core indicator of anomaly attribution analysis. Based on the sliding time window length, the power-thermal-electrical integrated mutation ratio at all times within the window is accumulated and averaged, and the absolute value is taken to obtain the power-thermal-electrical coupling response ratio. The power-thermal-electrical coupling response ratio can quantitatively reflect the normalized intensity of power anomalies under the background of coordinated thermal and electrical features, which helps to accurately identify subsequent anomaly types and self-healing regulation.

[0047] The specific formula for the power thermoelectric coupling response ratio is as follows:

[0048] ;

[0049] In the formula, The power thermoelectric coupling response ratio represents the normalized abrupt change intensity of power under the combined influence of temperature and current variations, i.e., the attribution characteristics of the thermoelectric response to power anomalies. A significant increase indicates that the power anomaly was not significantly responded to by temperature and current, which is most likely due to shielding or mismatch. Conversely, if there are drastic changes in temperature and current, it can indicate faults and safety hazards, providing a scientific classification basis for subsequent self-healing regulation and cause tracing. The formula for calculating the change in power is: It measures short-term power fluctuations; Indicates the preceding The power at each sampling interval provides a historical benchmark for power changes; The formula for representing the change in temperature is: It reflects short-term temperature rises and falls; Indicates the preceding The temperature at each sampling interval serves as a historical reference for temperature changes; The formula for representing the change in current is: Used for sensing sudden current changes and anomalies; Indicates the preceding The current at each sampling interval provides a historical benchmark for current changes; This represents a very small constant value, with a value of 0.001. Indicates the length of the sliding time window; This indicates the number of sampling periods and defines the time scale for calculating changes. It represents the relative rate of change of power, measures the intensity of the instantaneous relative change in power within the current window, reflects the short-term large changes in local power generation capacity and load, and is highly sensitive to abnormal output due to blocking, mismatch, string faults. It represents the combined rate of change of temperature and current, comprehensively assesses the relative changes of temperature and current, reflects electrical and thermal response capabilities, distinguishes multiple types of anomalies such as thermal runaway, short circuit, and overload, and improves the accuracy of anomaly classification and identification.

[0050] In this implementation plan, by dynamically tracking and comparing multi-feature data of equipment current, power, and temperature during abnormal periods, and combining historical data from the previous d sampling intervals, the coordinated fluctuations of each feature during abnormal periods can be quantified in detail. Using relative change rate and temperature-current composite change rate indices, a power-thermal-electric coupling response ratio is formed, and statistical averaging is performed through a sliding window to effectively extract the multi-parameter response characteristics of abnormal events. The power-thermal-electric coupling response ratio not only improves the scientific rigor and discriminative power of anomaly attribution but also provides a more accurate basis for subsequent anomaly type identification and self-healing adjustment, thereby enhancing the photovoltaic power plant's comprehensive judgment and intelligent response capabilities to complex multi-source anomalies.

[0051] Specifically, the process for identifying anomaly types based on the coordinated fluctuations of various characteristic anomalies is as follows: Based on the power thermoelectric coupling response ratio and its changing trend, the anomaly type is determined by jointly analyzing the relative change rates of power, temperature, and current. When the power thermoelectric coupling response ratio is greater than the secondary coupling exceedance threshold, and the relative change rate of temperature is less than or equal to the temperature change threshold and the relative change rate of current is less than or equal to the current change threshold, it is identified as a shading mismatch anomaly. The secondary coupling exceedance threshold is a high-risk criterion; if the power anomaly is obvious but the temperature and current do not fluctuate accordingly, it is mostly due to performance degradation caused by local shading of components and series-parallel mismatch. When the power thermoelectric coupling response ratio is greater than the primary coupling exceedance threshold and less than or equal to the secondary coupling exceedance threshold, and the relative change rate of temperature is greater than the temperature change threshold while the relative change rate of current is less than or equal to the current change threshold, it is identified as a thermal runaway anomaly. The primary coupling exceedance threshold is a lower-risk criterion; abnormal temperature fluctuations without significant current changes are common in equipment heat dissipation anomalies and local hot spot risks. When the power thermoelectric coupling response ratio is less than or equal to the first-level coupling over-limit threshold, and the relative temperature change rate is greater than the temperature change threshold and the relative current change rate is greater than the current change threshold, it is identified as a current disturbance anomaly. Current disturbance anomalies usually manifest as faults in electrical circuits, violent current fluctuations accompanied by temperature changes, such as string short circuits and junction box faults.

[0052] This implementation plan achieves precise classification of anomaly types by classifying and judging the power thermoelectric coupling response ratio and its changing trend, combined with the relative change rates of power, temperature, and current. Based on the multi-feature coordinated fluctuation characteristics, it distinguishes various typical anomaly modes such as shading mismatch, thermal runaway, and current disturbance, effectively improving the scientific rigor and relevance of anomaly attribution. This not only enhances the photovoltaic power plant's intelligent identification capability for different anomaly mechanisms but also provides a solid foundation for the automatic matching and response of subsequent differentiated self-healing maintenance and adjustment measures, further promoting the refinement and intelligence of photovoltaic power plant operation and maintenance management.

[0053] Specifically, the self-healing maintenance and adjustment process based on the anomaly type is as follows: For shading mismatch anomalies, the shading mismatched strings are identified, and the abnormal strings are temporarily isolated using intelligent switches to achieve rapid response and precise management of string on / off states. After a delay, the strings are restarted, and the power recovery status is checked. Restarting and checking after isolation helps to troubleshoot transient and persistent faults, improving the reliability of self-healing judgment. If the anomaly persists after isolation and restart, the bypass branch is controlled to ensure stable power output of the main circuit. The bypass branch is an auxiliary conduction circuit connected in parallel in the photovoltaic array, which can be activated when the main branch is abnormal, ensuring that the overall power generation capacity is not affected by single-point failure. For thermal runaway anomalies, a derating command is issued to the equipment with abnormal temperature, lowering the inverter power setpoint, adjusting the string output, and reducing the heat load. The liquid cooling heat dissipation equipment is activated for localized rapid cooling. The derating command refers to reducing the output power of the abnormal equipment by controlling the reduction of heat generation. The liquid cooling heat dissipation equipment is a high-efficiency heat dissipation unit based on fluid circulation, which can achieve rapid cooling of equipment nodes and effectively prevent the temperature from rising continuously and causing greater risks. For current disturbance-type anomalies, the fault point is identified, and the control module switches the circuit to isolate the faulty circuit and redistribute the load to the pre-set redundant power supply lines, reducing the impact on the main grid and other branches. The control module is an integrated intelligent control unit that can complete the isolation of the faulty circuit and the switching of redundant lines. Redundant power supply lines refer to backup electrical circuits pre-configured to improve reliability. This localizes the impact of the fault, improving the system's resilience and power supply continuity. Combined temperature control and power reduction measures achieve multi-dimensional joint protection. Temperature control measures include, but are not limited to, air cooling, liquid cooling, and intelligent ventilation. Power reduction actively reduces the output power of some equipment to alleviate load and prevent thermal runaway. Multi-dimensional joint protection simultaneously ensures equipment safety from both thermal and electrical perspectives. A fault maintenance work order is generated, including: the power thermoelectric coupling response ratio and corresponding power, temperature, and current data, as well as the corresponding self-healing maintenance adjustment process. The fault maintenance work order is synchronously written into the photovoltaic power plant operation database. The fault maintenance work order provides standardized and structured data support for subsequent operation and maintenance and traceability, enabling closed-loop management of the entire anomaly handling process.

[0054] This implementation plan can match optimal self-healing maintenance and adjustment measures for different types of anomalies. For shading mismatch anomalies, string isolation and restart are achieved through intelligent switches, and bypass diodes can be activated to ensure the power output of the main system. For thermal runaway anomalies, derating commands can be issued and liquid cooling equipment can be activated to quickly reduce equipment temperature and alleviate heat load. For current disturbance anomalies, soft switching and temperature control are used in conjunction to effectively isolate faulty circuits and provide multi-faceted protection. Fault maintenance work orders are generated and recorded throughout the process to ensure traceability of anomaly responses and standardized management. This achieves automated, differentiated, and closed-loop management of self-healing responses, improving the operation and maintenance efficiency and safety resilience of photovoltaic power plants.

[0055] Specifically, the process of judging the effect of self-healing maintenance by combining the photovoltaic power plant operation data before and after self-healing maintenance is as follows: obtain the power, power factor and temperature data of the equipment within a window during the abnormal period before self-healing maintenance from the photovoltaic power plant operation database; at the same time, obtain the power, power factor and temperature data of the equipment within the same window length after self-healing maintenance; by comparing the operation data before and after self-healing maintenance within the same time window, the actual impact of the regulation measures on the equipment performance and status can be dynamically evaluated. Based on the sliding time window length, the power difference is calculated by comparing the power after self-healing maintenance adjustment with the power before self-healing maintenance adjustment at each moment. The power difference values ​​at all moments are then summed and averaged to obtain the mean power difference. This power difference reflects the improvement effect of the adjustment measures on power generation output. Using the average value avoids interference from occasional fluctuations and improves the stability of the evaluation. For each moment, the power factor ratio is obtained by dividing the power factor after self-healing maintenance adjustment by the sum of the power factor before self-healing maintenance adjustment and the minimum constant value. The power factor ratio measures the degree of improvement in equipment energy utilization after self-healing adjustment. The minimum constant value is 0.001 to prevent calculation errors caused by a zero denominator. Anomalies are identified, and the power factor ratios at all times are summed and averaged to obtain the mean power factor ratio. The mean power difference is multiplied by the mean power factor ratio to obtain the comprehensive self-healing efficiency numerator. The comprehensive self-healing efficiency numerator integrates two dimensions: improved power generation capacity and improved power quality, thus more comprehensively reflecting the self-healing effect. Simultaneously, the temperature change value is obtained by calculating the difference between the temperature after self-healing maintenance and the temperature before self-healing maintenance. The standard deviation of the temperature change values ​​at all times is calculated and added to a constant to obtain the temperature recovery fluctuation factor. The temperature change value and its standard deviation quantify the stability of temperature fluctuations after self-healing adjustment. The constant is used to improve the stability of the denominator and prevent the standard deviation from being too small and amplifying the result. The comprehensive self-healing gain value is obtained by dividing the comprehensive self-healing efficiency numerator by the temperature recovery fluctuation factor. The comprehensive self-healing gain value serves as the core indicator for evaluating the overall effect of self-healing maintenance. A larger value indicates a more significant improvement in equipment performance, energy efficiency, and temperature stability, providing quantitative support for adjusting self-healing strategies and judging their effectiveness.

[0056] The specific formula for the overall self-healing gain value is as follows:

[0057] ;

[0058] In the formula, The overall self-healing gain value reflects the power increase, power factor improvement, and temperature stability brought about by self-healing maintenance regulation. It is used to judge the quality of self-healing effect and is a key indicator for quantitative evaluation of multi-parameter self-healing effectiveness. The larger the overall self-healing gain value, the more significant the performance improvement and effective suppression of temperature fluctuations brought about by self-healing maintenance regulation, and the better the self-healing effect. It represents the power output after self-healing maintenance adjustment, reflecting the power generation capacity of the equipment after the self-healing action, and is a direct basis for evaluating the improvement of the self-healing effect; This represents the power before self-healing maintenance adjustment, serving as a baseline before self-healing and used for comparison with the power after self-healing to reflect the extent of performance improvement. It represents the power factor after self-healing maintenance adjustment, and evaluates the quality of the output power after self-healing. The closer it is to 1, the higher the energy utilization rate. This indicates the power factor before self-healing maintenance adjustment, used for comparison to show the degree of power factor improvement after self-healing; This represents a very small constant value, with a value of 0.001. This indicates the temperature after self-healing maintenance and adjustment; Indicates the temperature before self-healing maintenance and adjustment; This represents the temperature recovery fluctuation factor, which measures the volatility of temperature changes within a window and reflects whether the self-healing action has led to an improvement in thermal stability. A large denominator indicates large temperature fluctuations and unstable self-healing; a small denominator indicates stable temperature recovery and a controlled self-healing process. This represents the average power difference, which measures the improvement in power after self-healing compared to before self-healing. It reflects the direct energy gain brought about by self-healing. The larger the value, the better the equipment recovery and energy efficiency after self-healing. It represents the power factor ratio mean, which measures the average improvement of the power factor after self-healing compared to before self-healing. The larger the value, the better the quality of power output, which is beneficial for long-term stable operation.

[0059] In this embodiment, Table 2 is a data table of the comprehensive self-healing gain values ​​of each device. The data for five devices were recorded in detail, including power before and after self-healing maintenance adjustment, power factor before and after self-healing maintenance adjustment, temperature before and after self-healing maintenance adjustment, and overall self-healing gain. Specifically, for device 1, the power before self-healing maintenance adjustment was 101, the power after self-healing maintenance adjustment was 109, the power factor before self-healing maintenance adjustment was 0.96, the power factor after self-healing maintenance adjustment was 0.99, the temperature before self-healing maintenance adjustment was 44.8°C, the temperature after self-healing maintenance adjustment was 42.3°C, and the overall self-healing gain was 4.057. For device 2, the power before self-healing maintenance adjustment was 88, the power after self-healing maintenance adjustment was 92, the power factor before self-healing maintenance adjustment was 0.93, the power factor after self-healing maintenance adjustment was 0.97, the temperature before self-healing maintenance adjustment was 46.0°C, the temperature after self-healing maintenance adjustment was 43.7°C, and the overall self-healing gain was 2.051. For device 3, the power before self-healing maintenance adjustment was... The self-healing rate is 120, the power after self-healing maintenance adjustment is 125, the power factor before self-healing maintenance adjustment is 0.95, the power factor after self-healing maintenance adjustment is 0.99, the temperature before self-healing maintenance adjustment is 48.5, the temperature after self-healing maintenance adjustment is 45.1, and the overall self-healing gain value is 2.562; the power before self-healing maintenance adjustment for device 4 is 99, the power after self-healing maintenance adjustment is 100, the power factor before self-healing maintenance adjustment is 0.98, and the power factor after self-healing maintenance adjustment is... 1.00, the temperature before self-healing maintenance adjustment was 45.7, the temperature after self-healing maintenance adjustment was 44.8, and the overall self-healing gain value was 0.502; the power before self-healing maintenance adjustment for device 5 was 111, the power after self-healing maintenance adjustment was 117, the power factor before self-healing maintenance adjustment was 0.94, the power factor after self-healing maintenance adjustment was 0.97, the temperature before self-healing maintenance adjustment was 47.9, the temperature after self-healing maintenance adjustment was 44.0, and the overall self-healing gain value was 3.044.

[0060] Table 2. Comprehensive Self-Healing Gain Values ​​of Each Device

[0061]

[0062] like Figure 4 As shown, this is a bar chart comparing the overall self-healing gain values ​​of various devices. Each bar represents one device, and the height of the bar and the numerical label above it reflect the overall self-healing gain value of each device after self-healing maintenance adjustments, intuitively enabling batch grading of the self-healing effects of multiple devices. A higher overall self-healing gain value indicates a better overall performance improvement and temperature fluctuation suppression effect brought about by the self-healing maintenance measures. (Based on Table 2 and...) Figure 4It can be seen that Device 1 has the highest overall self-healing gain value, indicating that the self-healing maintenance and adjustment measures have the most significant effect on Device 1, bringing about a noticeable increase in power and optimization of temperature. Devices 5 and 3 are next, also showing good self-healing effects. Device 2's overall self-healing gain value is at a moderate level, indicating that the self-healing maintenance and adjustment measures have limited effect on improving Device 2. Device 4's overall self-healing gain value is significantly lower than that of the other devices, suggesting that Device 4's self-healing maintenance and adjustment is ineffective and may have persistent anomalies, requiring further optimization of self-healing parameters and strategies.

[0063] This implementation plan achieves a multi-dimensional quantitative evaluation of self-healing effectiveness through windowed comparative analysis of key operating data on equipment power, power factor, and temperature before and after self-healing maintenance adjustments. The mean power difference and mean power factor ratio are used to characterize the improvement in energy efficiency and power quality after self-healing, while the temperature recovery fluctuation factor reflects the improvement in temperature stability during the adjustment process. Finally, a comprehensive self-healing gain value is constructed to effectively measure the overall effect of self-healing measures. This comprehensively reflects the improvement in equipment performance and safety brought about by self-healing adjustments, providing a scientific basis for subsequent optimization of self-healing strategies and differentiated operation and maintenance.

[0064] Specifically, based on photovoltaic power plant operation data, abnormal event records, and self-healing maintenance adjustment effects, the specific process for optimizing the algorithm and self-healing strategy is as follows: The comprehensive self-healing gain value is compared with the self-healing threshold. When the comprehensive self-healing gain value is less than the self-healing threshold, the main characteristics of the current self-healing maintenance adjustment failure are analyzed, and the self-healing maintenance adjustment parameters are adjusted accordingly. A second self-healing maintenance adjustment is then performed. If the initial self-healing maintenance adjustment only adopts limited and mild measures such as parameter fine-tuning, power limiting, and derating of some equipment, the second adjustment will employ enhanced intervention measures such as larger-scale parameter adjustments, wider-range load redistribution, string isolation, loop switching, and forced shutdown reset. Mild measures include small-scale adjustment of setpoints, while enhanced interventions include large-scale switching, isolation of faulty strings, forced power outages, and other enhanced measures, which can maximize the probability of successful self-healing and safety resilience. If the overall self-healing gain value still fails to meet the standard after multiple self-healing maintenance adjustments, a high-priority work order is generated for alarm, dispatch, and scheduling of rapid manual intervention to improve decision-making efficiency. Anomaly trend analysis and historical self-healing reports are pushed to assist in tracing the root cause of anomalies and remind maintenance personnel to conduct a review. When the overall self-healing gain value is greater than or equal to the self-healing threshold, only historical self-healing cases and overall self-healing gain values ​​are continuously written into the photovoltaic power station operation database. Statistical analysis is performed on historical self-healing cases and overall self-healing gain values, including summarizing and comparing typical cases and the distribution of good and bad effects to optimize the selection of self-healing maintenance adjustment strategies. The entire process of photovoltaic power plant operation data and comprehensive self-healing gain value before and after maintenance adjustment is visualized in multiple dimensions. Health grading and trend analysis are performed on batch equipment. Health grading classifies equipment status based on self-healing gain value and anomaly frequency parameters. Trend analysis helps identify systemic hidden dangers and generates self-healing effectiveness reports. Historical data on self-healing maintenance adjustment and photovoltaic power plant operation is periodically used to mine anomaly evolution patterns and identify optimal self-healing maintenance adjustment strategies through data mining, cluster analysis, pattern recognition, and parameter regression algorithms. This continuously optimizes and updates algorithms, thresholds, and self-healing maintenance adjustment parameters. Data mining, cluster analysis, pattern recognition, and parameter regression are typical algorithmic tools in big data analysis, capable of deeply exploring the intrinsic relationship between equipment anomalies and self-healing adjustment, continuously improving self-learning and adaptive capabilities, and achieving dynamic iterative optimization of algorithms, thresholds, and strategies.

[0065] In this implementation plan, adaptive optimization of self-healing maintenance adjustment parameters and strategies is achieved through dynamic comparison of the comprehensive self-healing gain value and the self-healing threshold. When the self-healing effect fails to meet expectations, failure characteristics are analyzed, parameters are adjusted, and stronger intervention measures are switched to ensure timely closed-loop handling of anomalies. For anomalies that fail to self-heal multiple times, high-priority work orders are generated and trend analysis is pushed to improve the initiative and intelligence level of anomaly handling. For compliant self-healing cases, historical data is continuously accumulated, and self-healing strategies are continuously optimized through self-learning and parameter recommendation. Combining data mining, clustering, and pattern recognition intelligent algorithms, the evolution patterns of anomalies and optimal self-healing strategies are continuously explored, constantly improving the adaptability of algorithms and the intelligence level of operation and maintenance decisions, helping photovoltaic power plants achieve efficient and autonomous energy efficiency optimization management.

[0066] Reference Figure 2 As shown, the second aspect of this invention provides an energy efficiency optimization management system for photovoltaic power plants, applied to the aforementioned energy efficiency optimization management method for photovoltaic power plants. The system includes: a multi-level high-frequency data acquisition and preprocessing module for real-time acquisition of photovoltaic power plant operation data and preprocessing the data; an anomaly signal detection and multi-scale analysis module for judging multi-feature mutations in equipment during photovoltaic power plant operation based on the preprocessed operation data, identifying operational anomalies based on the multi-feature mutation judgment results, and generating anomaly event records; a local anomaly attribution and self-healing maintenance module for receiving anomaly event records, evaluating the coordinated fluctuation of various anomalies based on photovoltaic power plant operation data, identifying anomaly types based on the coordinated fluctuation of various anomalies, and performing self-healing maintenance adjustments based on the anomaly types; and a self-healing feedback and optimization management module for judging the self-healing maintenance adjustment effect by combining photovoltaic power plant operation data before and after self-healing maintenance adjustment, and optimizing algorithms and self-healing strategies based on photovoltaic power plant operation data, anomaly event records, and the self-healing maintenance adjustment effect.

[0067] This implementation scheme achieves real-time, accurate perception and intelligent anomaly diagnosis of photovoltaic power plant operation status through the organic synergy of multi-level high-frequency data acquisition and preprocessing, abnormal signal detection and multi-scale analysis, local anomaly attribution and self-healing maintenance, and self-healing feedback and optimization management. It can not only sensitively capture sudden anomalies with multiple characteristics of equipment and promptly identify and classify anomaly types, but also realize automated self-healing maintenance and adjustment based on anomaly types. Furthermore, through quantitative evaluation of the self-healing effect, it continuously optimizes algorithm parameters and self-healing strategies, thereby improving the operational safety, intelligent operation and maintenance, and overall energy efficiency of the photovoltaic power plant.

[0068] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0069] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to the specific implementations described. As those skilled in the art will understand, 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 method for optimizing energy efficiency management in photovoltaic power plants, characterized in that, Includes the following steps: S1 collects real-time photovoltaic power station operation data and performs data preprocessing on the photovoltaic power station operation data; S2, based on the preprocessed photovoltaic power station operation data, determine the multi-feature mutations of the equipment during the operation of the photovoltaic power station, identify the abnormal operation of the photovoltaic power station equipment according to the multi-feature mutation judgment results, and generate abnormal event records; The specific process for determining multi-feature mutations of equipment during photovoltaic power plant operation based on preprocessed photovoltaic power plant operation data is as follows: Acquire current, voltage, and temperature data of the equipment during photovoltaic power plant operation; based on a sliding time window, calculate the mean and standard deviation of the current, voltage, and temperature data respectively to obtain the mean current, standard deviation of current, mean voltage, standard deviation of voltage, mean temperature, and standard deviation of temperature; The current normalization deviation is obtained by dividing the difference between the current and the mean current by the sum of the current standard deviation and the minimum constant value, and then squaring the result. Similarly, the voltage normalization deviation is obtained by dividing the difference between the voltage and the mean voltage by the sum of the voltage standard deviation and the minimum constant value, and then squaring the result. Finally, the temperature normalization deviation is obtained by dividing the difference between the temperature and the mean temperature by the sum of the temperature standard deviation and the minimum constant value, and then squaring the result. The transient multi-parameter mutation intensity factor is obtained by adding the normalized deviation values ​​of current, voltage, and temperature, and then taking the square root of the sum. S3 receives abnormal event records, evaluates the coordinated fluctuation of various characteristic anomalies based on photovoltaic power station operation data, identifies the anomaly type based on the coordinated fluctuation of various characteristic anomalies, and performs self-healing maintenance and adjustment according to the anomaly type. The specific process of receiving abnormal event records and evaluating the coordinated fluctuation of various characteristic anomalies based on photovoltaic power plant operation data is as follows: Receive abnormal event records, identify specific abnormal devices and time periods, and acquire the current, power and temperature data of the devices during the abnormal period in real time based on a sliding time window. At the same time, acquire the current, power and temperature data of the previous d sampling intervals from the photovoltaic power plant operation database. The power change is obtained by subtracting the power from the power of the previous d sampling intervals from the power at the current moment, and the relative rate of change of power is obtained by dividing the power change by the power of the previous d sampling intervals. The current change is obtained by subtracting the current from the current of the previous d sampling intervals from the current at the current moment, and the relative rate of change of current is obtained by dividing the current change by the current of the previous d sampling intervals. The temperature change is obtained by subtracting the temperature from the temperature of the previous d sampling intervals from the temperature at the current moment, and the relative rate of change of temperature is obtained by dividing the temperature change by the temperature of the previous d sampling intervals. The relative rate of change of current and the relative rate of change of temperature are added to the minimum constant value to obtain the combined rate of change of temperature and current; the power thermoelectric combined mutation ratio is obtained by dividing the relative rate of change of power by the combined rate of change of temperature and current; based on the sliding time window length, the power thermoelectric combined mutation ratio at all times within the window is accumulated and averaged, and the absolute value is taken to obtain the power thermoelectric coupling response ratio. The specific process for identifying anomaly types based on the coordinated fluctuation of various characteristic anomalies is as follows: Based on the power thermoelectric coupling response ratio and its changing trend, anomaly types are identified through joint analysis of the relative change rates of power, temperature, and current. When the power thermoelectric coupling response ratio is greater than the secondary coupling exceedance threshold, and the relative temperature change rate is less than or equal to the temperature change threshold and the relative current change rate is less than or equal to the current change threshold, it is identified as a shading mismatch anomaly. When the power thermoelectric coupling response ratio is greater than the first-level coupling exceedance threshold and less than or equal to the second-level coupling exceedance threshold, and the relative temperature change rate is greater than the temperature change threshold while the relative current change rate is less than or equal to the current change threshold, it is identified as a thermal runaway anomaly. When the power thermoelectric coupling response ratio is less than or equal to the first-level coupling exceedance threshold, and the relative temperature change rate is greater than the temperature change threshold and the relative current change rate is greater than the current change threshold, it is identified as a current disturbance anomaly. S4 combines the photovoltaic power plant operation data before and after self-healing maintenance adjustment to determine the self-healing maintenance adjustment effect. Based on the photovoltaic power plant operation data, abnormal event records, and self-healing maintenance adjustment effect, the algorithm and self-healing strategy are optimized.

2. The energy efficiency optimization management method for photovoltaic power plants according to claim 1, characterized in that, The specific process of real-time acquisition of photovoltaic power plant operation data and data preprocessing of the photovoltaic power plant operation data is as follows: The photovoltaic power station operates by collecting real-time, high-frequency data from various core devices at multiple levels. The photovoltaic power station operating data includes current, voltage, temperature, power, and power factor. Median filtering is used to perform preliminary noise reduction on photovoltaic power plant operation data, and outlier detection and signal interpolation repair are performed; time-series alignment, sliding window normalization and standardization are performed on photovoltaic power plant operation data; By performing drift rate analysis and sliding standard deviation calculation on photovoltaic power plant operation data, invalid signals are identified and eliminated. At the same time, based on the data stability results, a primary / backup switch is performed on the data acquisition link; Establish a photovoltaic power plant operation database to store photovoltaic power plant operation data.

3. The energy efficiency optimization management method for photovoltaic power plants according to claim 1, characterized in that, The specific process of identifying abnormal operation of photovoltaic power station equipment based on the multi-feature mutation judgment result and generating abnormal event records is as follows: The transient multi-parameter mutation intensity factor of each device is calculated in real time and written into the photovoltaic power station operation database. When the transient multi-parameter mutation intensity factor is greater than or equal to the abnormal threshold, it is determined as a local abnormal event. The current abnormal time, abnormal device, corresponding current, voltage and temperature data, as well as the mean and standard deviation of the corresponding data are recorded to generate an abnormal event record. And push the abnormal event record to the next process; When the transient multi-parameter mutation intensity factor is less than the anomaly threshold, the photovoltaic power station operation data and transient multi-parameter mutation intensity factor are continuously collected and calculated. Using historical photovoltaic power station operation data and transient multi-parameter mutation intensity factor, the mean, standard deviation and anomaly threshold are updated through sliding window adaptive statistical method and exponential weighted moving average method.

4. The energy efficiency optimization management method for photovoltaic power plants according to claim 1, characterized in that, The specific process of self-healing maintenance and adjustment based on the anomaly type is as follows: For shading mismatch anomalies, identify the shading mismatch string and temporarily isolate the abnormal string through a smart switch. After a delay, restart the switch and check the power recovery status. If the anomaly persists after isolation and restart, control the bypass branch to conduct to ensure stable power output of the main circuit. For thermal runaway-type anomalies, a derating command is issued to the equipment with abnormal temperature to lower the inverter power setpoint, adjust the string output, and reduce the heat load; the liquid cooling equipment is activated for localized rapid cooling. For current disturbance-type anomalies, the fault point is identified, and the circuit switching is realized through the control module to separate the faulty circuit and redistribute the load to the preset redundant power supply line to reduce the impact on the main power grid and other branches. By combining temperature control measures with power reduction, multi-dimensional joint protection is achieved; A fault maintenance work order is generated, including the power thermoelectric coupling response ratio and the corresponding power, temperature and current data, as well as the corresponding self-healing maintenance adjustment process. The fault maintenance work order is then synchronously written into the photovoltaic power plant operation database.

5. The energy efficiency optimization management method for photovoltaic power plants according to claim 1, characterized in that, The specific process for judging the effectiveness of self-healing maintenance by combining the photovoltaic power station operation data before and after self-healing maintenance is as follows: From the photovoltaic power plant operation database, obtain the power, power factor, and temperature data of the equipment within a window during the abnormal period before self-healing maintenance adjustment; at the same time, obtain the power, power factor, and temperature data of the equipment within the same window length after self-healing maintenance adjustment. Based on the sliding time window length, the power difference is calculated by comparing the power after self-healing maintenance adjustment with the power before self-healing maintenance adjustment at each moment. The power difference values ​​at all moments are then summed and averaged to obtain the average power difference. For each moment, the power factor ratio is obtained by dividing the power factor after self-healing maintenance adjustment by the sum of the power factor before self-healing maintenance adjustment and the minimum constant value. The power factor ratio ratio is then summed and averaged at all moments to obtain the average power factor ratio. The average power difference is multiplied by the average power factor ratio to obtain the comprehensive self-healing efficiency numerator. Simultaneously, the temperature change value is obtained by calculating the difference between the temperature after self-healing maintenance adjustment and the temperature before self-healing maintenance adjustment. The standard deviation of the temperature change value at all times is calculated and added to a constant to obtain the temperature recovery fluctuation factor. The comprehensive self-healing gain value is obtained by dividing the comprehensive self-healing efficiency numerator by the temperature recovery fluctuation factor.

6. The energy efficiency optimization management method for photovoltaic power plants according to claim 1, characterized in that, The specific process of optimizing the algorithm and self-healing strategy based on photovoltaic power plant operation data, abnormal event records, and self-healing maintenance adjustment effects is as follows: The overall self-healing gain value is compared with the self-healing threshold. When the overall self-healing gain value is less than the self-healing threshold, the main characteristics of the self-healing maintenance adjustment failure are analyzed, the self-healing maintenance adjustment parameters are adjusted accordingly, and a second self-healing maintenance adjustment is performed. If the initial self-healing maintenance adjustment only adopts limited and mild measures such as parameter fine-tuning, power limiting, and derating of some equipment, the second adjustment will adopt enhanced intervention measures such as larger parameter adjustments, wider load redistribution, string isolation, loop switching, and forced shutdown and reset. If the overall self-healing gain value still fails to meet the standard after multiple self-healing maintenance adjustments, a high-priority work order is generated, and an abnormal trend analysis and self-healing history report are pushed to remind the operation and maintenance personnel to review the issue. When the comprehensive self-healing gain value is greater than or equal to the self-healing threshold, only historical self-healing cases and comprehensive self-healing gain values ​​are continuously written into the photovoltaic power station operation database. Statistical analysis is performed on historical self-healing cases and comprehensive self-healing gain values ​​to optimize the selection of self-healing maintenance and adjustment strategies. The entire process of photovoltaic power plant operation data and comprehensive self-healing gain value before and after self-healing maintenance adjustment is visualized in multiple dimensions. Health classification and trend analysis of batch equipment are carried out to generate self-healing effectiveness reports. Historical data on self-healing maintenance adjustment and photovoltaic power plant operation are regularly used to mine abnormal evolution patterns and identify the optimal self-healing maintenance adjustment strategy through data mining, cluster analysis, pattern recognition and parameter regression algorithms. The algorithms, thresholds and self-healing maintenance adjustment parameters are continuously optimized and updated.

7. An energy efficiency optimization management system for photovoltaic power plants, employing the energy efficiency optimization management method for photovoltaic power plants as described in any one of claims 1-6, characterized in that, include: A multi-level high-frequency data acquisition and preprocessing module is used to acquire photovoltaic power plant operation data in real time and perform data preprocessing on the photovoltaic power plant operation data; The abnormal signal detection and multi-scale analysis module is used to determine multi-feature mutations of equipment during the operation of photovoltaic power plants based on preprocessed photovoltaic power plant operation data, identify abnormal operation of photovoltaic power plant equipment based on the results of multi-feature mutation judgment, and generate abnormal event records. The local anomaly attribution and self-healing maintenance module is used to receive anomaly event records, evaluate the coordinated fluctuation of various characteristic anomalies based on photovoltaic power plant operation data, identify the anomaly type based on the coordinated fluctuation of various characteristic anomalies, and perform self-healing maintenance adjustment according to the anomaly type. The self-healing feedback and optimization management module is used to combine the photovoltaic power plant operation data before and after self-healing maintenance and adjustment to judge the effect of self-healing maintenance and adjustment, and to optimize the algorithm and self-healing strategy based on the photovoltaic power plant operation data, abnormal event records and the effect of self-healing maintenance and adjustment.

Images