Photovoltaic power station equipment intelligent early warning system and method based on multi-source data

Abstract

The invention discloses a photovoltaic power station equipment intelligent early warning system and method based on multi-source data, and relates to the technical field of photovoltaic intelligent operation and maintaining.The photovoltaic power station equipment intelligent early warning method comprises the steps that firstly, meteorological and electrical data of photovoltaic modules are synchronously collected, and cleaning, alignment and normalization processing are conducted; analyzing the real-time slope of the power and irradiance of the component, dividing the component into three time periods of low-irradiation standby, linear working and saturated working, and identifying equipment with serious performance degradation; respectively calculating an insulation degradation coefficient, a performance attenuation index, a voltage mismatch factor and a current consistency index in different time periods, detecting an early insulation fault, evaluating performance attenuation and detecting an electrical mismatch problem; the comprehensive degradation degree score is calculated through weighted average by further combining the accumulated operation time and the historical fault frequency of the photovoltaic module, and three-level dynamic early warning is executed accordingly, so that intelligent evaluation and accurate early warning of the health state of the photovoltaic module are achieved, and the intelligent level of power station operation and maintenance and the fault response timeliness are improved.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic intelligent operation and maintenance technology, specifically to an intelligent early warning system and method for photovoltaic power plant equipment based on multi-source data. Background Technology

[0002] Photovoltaic power generation, as an important component of clean energy, plays a crucial role in achieving the "dual carbon" goal. With the continuous growth of installed capacity and expansion of deployment scale of photovoltaic power plants, the long-term reliable operation and efficient maintenance of power plant equipment have become core issues of concern in the industry. As the basic unit of power generation, photovoltaic modules are exposed to complex natural environments for extended periods, inevitably affected by multiple factors such as sunlight, temperature, humidity, dust, and aging, leading to gradual performance degradation or various failures, such as hot spot effects, wiring faults, insulation deterioration, and power mismatch.

[0003] Currently, the operation and maintenance of photovoltaic power plants still largely rely on traditional methods such as regular inspections and reactive repairs, lacking real-time, accurate perception and intelligent assessment of equipment status. While existing monitoring systems can collect some operational data, they are mostly limited to monitoring single-dimensional electrical parameters, failing to fully integrate multi-source information and lacking targeted analysis of component behavior characteristics under different operating conditions. This makes it difficult to detect early, subtle faults, and potential degradation trends go undetected, with early warnings often lagging behind the actual occurrence of failures.

[0004] Therefore, there is an urgent need for an intelligent early warning system and method that can integrate multi-source data, adapt to different operating conditions, and comprehensively consider the historical status and real-time performance of equipment, so as to achieve early diagnosis, accurate assessment and early warning of the health status of photovoltaic modules, thereby improving the initiative, foresight and intelligence level of power plant operation and maintenance, and ensuring the safe, efficient and long-term stable operation of photovoltaic power plants. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent early warning system and method for photovoltaic power plant equipment based on multi-source data, so as to solve the problems raised in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for intelligent early warning of photovoltaic power plant equipment based on multi-source data, characterized in that the method includes the following steps:

[0008] Step S100: Collect meteorological and electrical multi-source data of photovoltaic modules in photovoltaic power station, clean and align the multi-source data to ensure that signals from different sources are strictly synchronized on the time axis;

[0009] Step S101: Simultaneously collect meteorological and electrical multi-source data in the photovoltaic power station to construct an original data pool. The meteorological multi-source data includes the planar irradiance I of the photovoltaic module and the backsheet temperature T. The electrical multi-source data includes the DC voltage, DC current and output power P of each photovoltaic string and module.

[0010] Step S102: Align the timestamps of all data streams using a clock synchronization protocol, filter and remove abnormal data points, normalize the measured power to the baseline value under standard test conditions, and eliminate the interference of environmental fluctuations on performance comparison.

[0011] Step S200: The output characteristics of photovoltaic modules change nonlinearly with light intensity. By analyzing the slope k of irradiance and DC power in photovoltaic modules, the instantaneous energy conversion efficiency of the modules is characterized. Based on the slope, different physical characteristic time periods are divided and severely misaligned equipment is identified, so as to realize a more refined and adaptive deconstruction of the operating status. A fast channel is embedded to block catastrophic failures in real time.

[0012] Step S201: Calculate the dynamic performance slope k: For each photovoltaic module, within a continuous time window t, use the linear regression method to calculate the real-time slope k(t) between its normalized DC power P(t) and irradiance I(t). The slope k(t) characterizes the instantaneous conversion efficiency of the module under the current environment, and k(t) represents the slope between the DC power P(t) and irradiance I(t) of the photovoltaic module under the current time window t.

[0013] Step S202: Calculate the theoretical slope k of the photovoltaic module model under standard conditions. STC k STC =P STC / I STC In the formula, P STC This is the ideal power output value of current photovoltaic modules, I STC This is the current ideal irradiance value for photovoltaic modules, calculated by comparing the real-time slope k(t) with the theoretical slope k. STC The comparison is used to make auxiliary judgments for the runtime segment, specifically: when k(t) < a1×k STC When a1×k indicates that the component conversion efficiency is extremely low, it may be in standby or severely limited state, and is judged as a low-irradiation standby period; STC ≤k(t)≤a2×k STC When k(t) > a2×k, it indicates that the component is working within the normal linear range, and is determined to be a linear working period; STC When the time corresponds to the power saturation zone, it is determined to be a saturated working period. a1 and a2 are coefficient thresholds, where a1 < a2, and are set by professionals.

[0014] Step S203: Simultaneously, when the photovoltaic module's k(t) remains below k0×k during data points in the linear operating period exceeding θ%, STC When the condition is marked as "severe performance degradation", a high-priority warning is generated directly, and repairs are carried out immediately without waiting for subsequent comprehensive scoring. The θ and k0 mentioned are emergency thresholds, which are set by professionals.

[0015] Step S300: During the low-irradiance standby period, calculate the insulation degradation coefficient of the photovoltaic module to detect early insulation faults; during the linear operating period, calculate the performance degradation index of the photovoltaic module to assess performance degradation; during the saturated operating period, calculate the voltage mismatch factor and current consistency index of the photovoltaic module to detect series mismatch and connection faults. Through differentiated analysis of the low-irradiance period, linear operating period, and saturated operating period, the system can respectively gain insight into the insulation health status of the module, the degree of general performance degradation, and the series matching consistency, thereby transforming the complex operating status into a series of quantifiable and comparable specific scores, providing multi-dimensional and accurate input for comprehensive evaluation.

[0016] Step S301: During the low-irradiance standby period, the photovoltaic module is close to the open-circuit state. The module acts like a capacitor, and its open-circuit voltage is extremely sensitive to the internal insulation state. Calculate the average open-circuit voltage V of the photovoltaic module during the period. oc,avg Then compared with the benchmark value V oc,ref By comparing and normalizing the relative deviations between the two, the insulation degradation coefficient can be obtained. :

[0017] ;

[0018] In the formula, δ OC Alarm thresholds are set by professionals.

[0019] Step S302: During the linear operating period, the photovoltaic module's behavior is stable and predictable, which is the core window for evaluation. Calculate the ratio PR of the measured power of the photovoltaic module during this period to the current ideal power. raw PR raw =P(t) / P STC Then, by comparing it with the benchmark value, the performance degradation index D is obtained:

[0020] D = max(0, 1-PR);

[0021] In the formula, PR is the normalized performance ratio: the performance degradation index D quantifies the power loss of the module caused by material aging, cell efficiency degradation, uniform dust coverage, etc., and is the core indicator for evaluating the long-term durability of the module.

[0022] ;

[0023] In the formula, For the current photovoltaic module, the current time window PR raw The average value, The average performance ratio of healthy photovoltaic modules during the same period;

[0024] Step S303: During the saturation operating period, the photovoltaic modules operate in the high current output range. According to the characteristics of a series circuit, the loop current is limited by the module with the weakest output capability. At this time, the dispersion of the output voltage of each photovoltaic module is amplified, becoming a key observation window for accurately locating local faults. The relative deviation between the average voltage of the photovoltaic modules and the average voltage of the string is calculated during the period to obtain the voltage mismatch factor VMF.

[0025] ;

[0026] In the formula, V i It is the average DC voltage of the current photovoltaic module, V avg This is the average DC voltage of the photovoltaic string, which refers to the arithmetic mean of the DC voltages of all normally monitored modules in the entire photovoltaic string containing the target module within the current time window. This is the voltage mismatch alarm threshold, set by professionals;

[0027] The relative deviation between the average current of photovoltaic modules and the average current of the string is analyzed during the analysis period to derive the current consistency index (CCI):

[0028] ;

[0029] In the formula, I i I is the average DC current of the target component. avg It is the string average DC current, which refers to the arithmetic mean of the currents of all photovoltaic modules in the photovoltaic string containing the target module within the current time window. This is the current consistency alarm threshold, which is set by professionals;

[0030] The voltage and current consistency index F is derived by combining the voltage mismatch factor and the current consistency index: F = k1 × VMF + k2 × CCI, where k1 and k2 are weighting coefficients, and k1 + k2 = 1.

[0031] Step S400: Calculate the operating life factor by calculating the cumulative effective power generation time of each photovoltaic module; calculate the fault density factor by calculating the historical number of faults of each photovoltaic module, so that the final comprehensive evaluation result has both real-time sensitivity and historical predictability, which is more in line with the equipment management logic in engineering practice.

[0032] Step S401: Calculate the total operating time T of each photovoltaic module during the linear operating period. operThe total runtime is compared with the standard design life T of the current photovoltaic module. norm By comparison, the service life factor τ of the current photovoltaic modules is calculated:

[0033] ;

[0034] Step S402: Count the number of historical failures N for each photovoltaic module since it has been put into operation. fault , and the number of design failures N th In comparison, the failure density factor Φ of the current photovoltaic module is calculated:

[0035] .

[0036] Step S500: After the daily operation is completed, the data from the low-irradiance standby period, the linear working period, and the saturated working period are integrated. The operating years factor and the fault density factor are introduced to calculate the comprehensive degradation score for photovoltaic module fault early warning. Finally, the data analysis conclusions are transformed into clear, graded, and executable operation and maintenance action guidelines.

[0037] Step S501: After the daily operation is completed, the three time periods are integrated using the weighted geometric mean method, and then the service life factor and failure density factor are introduced to calculate the comprehensive deterioration score S:

[0038] ;

[0039] In the formula, c1, c2, and c3 are weighting coefficients, and c1 + c2 + c3 = 1. These are long-term factor coefficients, set by professionals;

[0040] Step S502: Based on the comprehensive degradation score, implement a dynamic hierarchical early warning mechanism, specifically as follows:

[0041] Attention level: S > Q1, indicating possible initial degradation or a trend that needs to be monitored, add to the watchlist, and strengthen data tracking;

[0042] Anomaly level: S > Q2, indicating a clear probability of failure. The system generates a work order, and it is recommended to prioritize this check in the next planned inspection.

[0043] Severity level: If S > Q3, it indicates that the fault is developing rapidly or is already very serious. Immediate on-site intervention should be arranged to prevent the fault from escalating or causing secondary risks.

[0044] A photovoltaic power plant equipment intelligent early warning system based on multi-source data, the system includes a multi-source data acquisition and preprocessing module, a dynamic slope analysis and time period division module, a time period fault feature extraction module, a long-term state factor statistics module, and a comprehensive evaluation and dynamic early warning module;

[0045] The multi-source data acquisition and preprocessing module is responsible for synchronously acquiring meteorological data (module planar irradiance I, backsheet temperature T) and electrical data (string and module DC voltage, current, and power P) from the photovoltaic power station site. All data streams are timestamped using a clock synchronization protocol to eliminate environmental fluctuation interference and construct a clean, comparable "raw data pool," laying the foundation for subsequent accurate analysis.

[0046] The dynamic slope analysis and time period division module analyzes the real-time relationship between the output power and irradiance of each photovoltaic module, calculates the dynamic performance slope for each module, and compares it with the theoretical slope. This intelligently divides the operating period into "low irradiance standby period", "linear working period" and "saturated working period". At the same time, this module has built-in emergency diagnosis logic, which can directly identify and mark modules with "severe performance degradation", trigger the highest priority warning, and realize immediate response to major faults.

[0047] The dynamic slope analysis and time period segmentation module includes a dynamic performance slope calculation unit, a working time period intelligent segmentation unit, and a severely degraded equipment identification unit.

[0048] The dynamic performance slope calculation unit, for each photovoltaic module, uses a linear regression method to fit the relationship between its normalized DC power P(t) and irradiance I(t) within a continuous sliding time window, and calculates the real-time slope k(t). The slope k(t) is a quantitative indicator of the instantaneous conversion efficiency of the module under the current environment, reflecting the electrical power that can be generated per unit irradiance, and is the core parameter for judging the real-time performance status of the module.

[0049] The intelligent division unit for working periods compares the real-time slope k(t) with the theoretical slope k of the photovoltaic module under standard conditions. STC The system compares the data and automatically divides the working state of the components into three characteristic time periods based on preset coefficient thresholds: "low irradiance standby period", "linear working period", and "saturation working period".

[0050] The severely degraded equipment identification unit continuously monitors the equipment during the linear operating period. When the photovoltaic module's slope k(t) remains below the emergency threshold k0*k at linear operating data points exceeding θ%, the unit detects the severely degraded equipment. STC If the component is directly identified as "severely degraded", the system will skip the subsequent comprehensive scoring process and immediately generate the highest priority maintenance warning to ensure that core faults that endanger the power plant's power generation safety are handled with zero delay.

[0051] The time-segmented fault feature extraction module employs a differentiated algorithm strategy to extract refined indicators that characterize specific types of faults. During low-irradiance standby periods, it calculates the insulation degradation coefficient of the photovoltaic module; during linear operating periods, it calculates the performance degradation index of the photovoltaic module; and during saturated operating periods, it calculates the voltage mismatch factor and current consistency index of the photovoltaic module. This enables multi-dimensional health status detection of the photovoltaic module.

[0052] The time-segmented fault feature extraction module includes an insulation degradation coefficient calculation unit, a performance decay index calculation unit, and an electrical parameter mismatch detection unit.

[0053] The insulation degradation coefficient calculation unit utilizes the characteristic that the open-circuit voltage of a photovoltaic module is extremely sensitive to its internal insulation when it is close to open circuit under low irradiance. It calculates the average open-circuit voltage of the module during the low irradiance standby period and compares it with the health benchmark value. The "insulation degradation coefficient γ" is calculated by normalizing the relative deviation. This coefficient can effectively detect early insulation failure hazards caused by moisture intrusion, encapsulation aging, etc.

[0054] The performance degradation index calculation unit calculates the original performance ratio (PR) by comparing the measured power of the photovoltaic module with the ideal power when the photovoltaic module is in the linear operating period. raw Then, it is normalized again with the reference performance ratio of healthy components under the same environment to calculate the "performance degradation index D". This index directly quantifies the long-term power loss caused by material aging, battery efficiency decline, uniform contamination, etc., and is a key indicator for evaluating the durability of components.

[0055] When the photovoltaic modules are operating at saturation, the electrical parameter mismatch detection unit calculates the "voltage mismatch factor (VMF)" by measuring the relative deviation between the module voltage and the average string voltage; and calculates the "current consistency index (CCI)" by measuring the relative deviation between the module current and the average string current. These two factors are then weighted and combined into an "electrical parameter consistency index (F)," which is used to accurately locate series mismatch faults caused by hot spots, shading, poor connections, or severe inconsistencies in module performance.

[0056] The long-term condition factor statistics module establishes an "equipment file" for each photovoltaic module. It calculates the "operating years factor" by statistically analyzing the cumulative effective power generation time, reflecting the degree of natural aging of the equipment. It also calculates the "failure density factor" by statistically analyzing the number of historical failures, reflecting the inherent reliability of the equipment or weak links in operation and maintenance. These two long-term factors will provide important background corrections for the final comprehensive evaluation.

[0057] The long-term state factor statistics module includes an operating years factor calculation unit and a fault density factor calculation unit.

[0058] The operating life factor calculation unit calculates the total operating time T of each photovoltaic module under the "linear operating period". oper Compared with the current standard design life T of photovoltaic modules norm By comparison, the "service life factor τ" was calculated, which quantifies the natural wear and tear process of the equipment over time in a normalized form.

[0059] The fault density factor calculation unit traces all historical fault records of each photovoltaic module since it was put into operation and counts the total number of faults N. fault , and the number of design failures N th In contrast, a "failure density factor Φ" is calculated, which reflects the frequency of failures and amplifies the risk weight of equipment that fails frequently.

[0060] After each day's operation, the comprehensive assessment and dynamic early warning module integrates fault characteristic indicators from three different working periods. These indicators include insulation degradation coefficient, performance decay index, voltage mismatch factor, and current consistency index, and incorporates factors such as service life and fault density. A comprehensive and objective "comprehensive degradation score S" is calculated using a weighted geometric average method. Based on this comprehensive degradation score S, the system executes a dynamic hierarchical early warning mechanism, classifying early warnings into "attention level," "abnormal level," and "severe level," and associating them with different operation and maintenance response strategies to form a complete closed loop from data analysis to operation and maintenance actions.

[0061] Compared with the prior art, the beneficial effects of the present invention are:

[0062] 1. Achieve accurate, early, multi-dimensional fault diagnosis and proactive warning: The system integrates meteorological and electrical multi-source data and intelligently divides three typical time periods based on the dynamic slope of irradiance-power. Targeted indicators are calculated for each time period, and through dynamic comparison with theoretical values ​​and historical data, this method can not only effectively eliminate environmental fluctuation interference but also accurately identify various hidden dangers such as early insulation faults, performance degradation, and series mismatch. It achieves proactive, graded warnings from "attention" to "serious," significantly improving the timeliness and accuracy of fault detection.

[0063] 2. Improve the scientific nature and efficiency of operation and maintenance decision-making, and promote the intelligent transformation of operation and maintenance mode: The system automatically integrates the evaluation results of multiple time periods every day, and introduces the operating years factor and historical fault density factor to calculate the comprehensive deterioration score, providing a quantitative health status profile for each component, so that operation and maintenance work can be transformed from traditional periodic inspection and post-event handling to data-driven predictive maintenance. Attached Figure Description

[0064] Figure 1 This is a flowchart illustrating the application of the present invention to an intelligent early warning system for photovoltaic power plant equipment based on multi-source data;

[0065] Figure 2 This is a schematic diagram of the structure of an intelligent early warning method for photovoltaic power station equipment based on multi-source data, according to the present invention. Detailed Implementation

[0066] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0067] Example: Figures 1-2 As shown, this invention provides a technical solution. A photovoltaic power station is set up with 100 photovoltaic strings, each string containing 20 photovoltaic modules, totaling 2000 modules. A photovoltaic power station equipment intelligent early warning method based on multi-source data is applied to module A-01. The method includes the following steps:

[0068] Step S100: Collect meteorological and electrical multi-source data of photovoltaic modules in photovoltaic power station, clean and align the multi-source data to ensure that signals from different sources are strictly synchronized on the time axis;

[0069] Step S101: Simultaneously collect meteorological and electrical multi-source data in the photovoltaic power station to construct an original data pool. The meteorological multi-source data includes the planar irradiance I of the photovoltaic module and the backsheet temperature T. The electrical multi-source data includes the DC voltage, DC current and output power P of each photovoltaic string and module.

[0070] Step S102: Align the timestamps of all data streams using a clock synchronization protocol, filter and remove abnormal data points, normalize the measured power to the baseline value under standard test conditions, and eliminate the interference of environmental fluctuations on performance comparison.

[0071] Step S200: The output characteristics of photovoltaic modules change nonlinearly with light intensity. By analyzing the slope k of irradiance and DC power in photovoltaic modules, the instantaneous energy conversion efficiency of the modules is characterized. Based on the slope, different physical characteristic time periods are divided and severely misaligned equipment is identified, so as to realize a more refined and adaptive deconstruction of the operating status. A fast channel is embedded to block catastrophic failures in real time.

[0072] Step S201: Calculate the dynamic performance slope k: For each photovoltaic module, within a continuous time window t, use the linear regression method to calculate the real-time slope k(t) between its normalized DC power P(t) and irradiance I(t). The slope k(t) characterizes the instantaneous conversion efficiency of the module under the current environment, and k(t) represents the slope between the DC power P(t) and irradiance I(t) of the photovoltaic module under the current time window t.

[0073] Step S202: Calculate the theoretical slope k of the photovoltaic module model under standard conditions. STC k STC =P STC / I STC In the formula, P STC This is the ideal power output value of current photovoltaic modules, I STC This is the current ideal irradiance value for photovoltaic modules, calculated by comparing the real-time slope k(t) with the theoretical slope k. STC The comparison is used to make auxiliary judgments for the runtime segment, specifically: when k(t) < a1×k STC When a1×k indicates that the component conversion efficiency is extremely low, it may be in standby or severely limited state, and is judged as a low-irradiation standby period; STC ≤k(t)≤a2×k STC When k(t) > a2×k, it indicates that the component is working within the normal linear range, and is determined to be a linear working period; STC When the time corresponds to the power saturation zone, it is determined to be a saturated working period. a1 and a2 are coefficient thresholds, where a1 < a2, and are set by professionals.

[0074] Step S203: Simultaneously, when the photovoltaic module's k(t) remains below k0×k during data points in the linear operating period exceeding θ%, STC When the condition is marked as "severe performance degradation", a high-priority warning is generated directly, and repairs are carried out immediately without waiting for subsequent comprehensive scoring. The θ and k0 mentioned are emergency thresholds, which are set by professionals.

[0075] Step S300: During the low-irradiance standby period, calculate the insulation degradation coefficient of the photovoltaic module to detect early insulation faults; during the linear operating period, calculate the performance degradation index of the photovoltaic module to assess performance degradation; during the saturated operating period, calculate the voltage mismatch factor and current consistency index of the photovoltaic module to detect series mismatch and connection faults. Through differentiated analysis of the low-irradiance period, linear operating period, and saturated operating period, the system can respectively gain insight into the insulation health status of the module, the degree of general performance degradation, and the series matching consistency, thereby transforming the complex operating status into a series of quantifiable and comparable specific scores, providing multi-dimensional and accurate input for comprehensive evaluation.

[0076] Step S301: During the low-irradiance standby period, the photovoltaic module is close to the open-circuit state. The module acts like a capacitor, and its open-circuit voltage is extremely sensitive to the internal insulation state. Calculate the average open-circuit voltage V of the photovoltaic module during the period. oc,avg Then compared with the benchmark value V oc,ref By comparing and normalizing the relative deviations between the two, the insulation degradation coefficient can be obtained. :

[0077] ;

[0078] In the formula, δ OC Alarm thresholds are set by professionals.

[0079] Step S302: During the linear operating period, the photovoltaic module's behavior is stable and predictable, which is the core window for evaluation. Calculate the ratio PR of the measured power of the photovoltaic module during this period to the current ideal power. raw PR raw =P(t) / P STC Then, by comparing it with the benchmark value, the performance degradation index D is obtained:

[0080] D = max(0, 1-PR);

[0081] In the formula, PR is the normalized performance ratio: the performance degradation index D quantifies the power loss of the module caused by material aging, cell efficiency degradation, uniform dust coverage, etc., and is the core indicator for evaluating the long-term durability of the module.

[0082] ;

[0083] In the formula, For the current photovoltaic module, the current time window PR raw The average value, The average performance ratio of healthy photovoltaic modules during the same period;

[0084] Step S303: During the saturation operating period, the photovoltaic modules operate in the high current output range. According to the characteristics of a series circuit, the loop current is limited by the module with the weakest output capability. At this time, the dispersion of the output voltage of each photovoltaic module is amplified, becoming a key observation window for accurately locating local faults. The relative deviation between the average voltage of the photovoltaic modules and the average voltage of the string is calculated during the period to obtain the voltage mismatch factor VMF.

[0085] ;

[0086] In the formula, V i It is the average DC voltage of the current photovoltaic module, V avgThis is the average DC voltage of the photovoltaic string, which refers to the arithmetic mean of the DC voltages of all normally monitored modules in the entire photovoltaic string containing the target module within the current time window. This is the voltage mismatch alarm threshold, set by professionals;

[0087] The relative deviation between the average current of photovoltaic modules and the average current of the string is analyzed during the analysis period to derive the current consistency index (CCI):

[0088] ;

[0089] In the formula, I i I is the average DC current of the target component. avg It is the string average DC current, which refers to the arithmetic mean of the currents of all photovoltaic modules in the photovoltaic string containing the target module within the current time window. This is the current consistency alarm threshold, which is set by professionals;

[0090] The voltage and current consistency index F is derived by combining the voltage mismatch factor and the current consistency index: F = k1 × VMF + k2 × CCI, where k1 and k2 are weighting coefficients, and k1 + k2 = 1.

[0091] Step S400: Calculate the operating life factor by calculating the cumulative effective power generation time of each photovoltaic module; calculate the fault density factor by calculating the historical number of faults of each photovoltaic module, so that the final comprehensive evaluation result has both real-time sensitivity and historical predictability, which is more in line with the equipment management logic in engineering practice.

[0092] Step S401: Calculate the total operating time T of each photovoltaic module during the linear operating period. oper The total runtime is compared with the standard design life T of the current photovoltaic module. norm By comparison, the service life factor τ of the current photovoltaic modules is calculated:

[0093] ;

[0094] Step S402: Count the number of historical failures N for each photovoltaic module since it has been put into operation. fault , and the number of design failures N th In comparison, the failure density factor Φ of the current photovoltaic module is calculated:

[0095] .

[0096] Step S500: After the daily operation is completed, the data from the low-irradiance standby period, the linear working period, and the saturated working period are integrated. The operating years factor and the fault density factor are introduced to calculate the comprehensive degradation score for photovoltaic module fault early warning. Finally, the data analysis conclusions are transformed into clear, graded, and executable operation and maintenance action guidelines.

[0097] Step S501: After the daily operation is completed, the three time periods are integrated using the weighted geometric mean method, and then the service life factor and failure density factor are introduced to calculate the comprehensive deterioration score S:

[0098] ;

[0099] In the formula, c1, c2, and c3 are weighting coefficients, and c1 + c2 + c3 = 1. These are long-term factor coefficients, set by professionals;

[0100] Step S502: Based on the comprehensive degradation score, implement a dynamic hierarchical early warning mechanism, specifically as follows:

[0101] Attention level: S > Q1, indicating possible initial degradation or a trend that needs to be monitored, add to the watchlist, and strengthen data tracking;

[0102] Anomaly level: S > Q2, indicating a clear probability of failure. The system generates a work order, and it is recommended to prioritize this check in the next planned inspection.

[0103] Severity level: If S > Q3, it indicates that the fault is developing rapidly or is already very serious. Immediate on-site intervention should be arranged to prevent the fault from escalating or causing secondary risks.

[0104] A photovoltaic power plant equipment intelligent early warning system based on multi-source data, the system includes a multi-source data acquisition and preprocessing module, a dynamic slope analysis and time period division module, a time period fault feature extraction module, a long-term state factor statistics module, and a comprehensive evaluation and dynamic early warning module;

[0105] The multi-source data acquisition and preprocessing module is responsible for synchronously acquiring meteorological data (module planar irradiance I, backsheet temperature T) and electrical data (string and module DC voltage, current, and power P) from the photovoltaic power station site. All data streams are timestamped using a clock synchronization protocol to eliminate environmental fluctuation interference and construct a clean, comparable "raw data pool," laying the foundation for subsequent accurate analysis.

[0106] The dynamic slope analysis and time period division module analyzes the real-time relationship between the output power and irradiance of each photovoltaic module, calculates the dynamic performance slope for each module, and compares it with the theoretical slope. This intelligently divides the operating period into "low irradiance standby period", "linear working period" and "saturated working period". At the same time, this module has built-in emergency diagnosis logic, which can directly identify and mark modules with "severe performance degradation", trigger the highest priority warning, and realize immediate response to major faults.

[0107] The dynamic slope analysis and time period segmentation module includes a dynamic performance slope calculation unit, a working time period intelligent segmentation unit, and a severely degraded equipment identification unit.

[0108] The dynamic performance slope calculation unit, for each photovoltaic module, uses a linear regression method to fit the relationship between its normalized DC power P(t) and irradiance I(t) within a continuous sliding time window, and calculates the real-time slope k(t). The slope k(t) is a quantitative indicator of the instantaneous conversion efficiency of the module under the current environment, reflecting the electrical power that can be generated per unit irradiance, and is the core parameter for judging the real-time performance status of the module.

[0109] The intelligent division unit for working periods compares the real-time slope k(t) with the theoretical slope k of the photovoltaic module under standard conditions. STC The system compares the data and automatically divides the working state of the components into three characteristic time periods based on preset coefficient thresholds: "low irradiance standby period", "linear working period", and "saturation working period".

[0110] The severely degraded equipment identification unit continuously monitors the equipment during the linear operating period. When the photovoltaic module's slope k(t) remains below the emergency threshold k0*k at linear operating data points exceeding θ%, the unit detects the severely degraded equipment. STC If the component is directly identified as "severely degraded", the system will skip the subsequent comprehensive scoring process and immediately generate the highest priority maintenance warning to ensure that core faults that endanger the power plant's power generation safety are handled with zero delay.

[0111] The time-segmented fault feature extraction module employs a differentiated algorithm strategy to extract refined indicators that characterize specific types of faults. During low-irradiance standby periods, it calculates the insulation degradation coefficient of the photovoltaic module; during linear operating periods, it calculates the performance degradation index of the photovoltaic module; and during saturated operating periods, it calculates the voltage mismatch factor and current consistency index of the photovoltaic module. This enables multi-dimensional health status detection of the photovoltaic module.

[0112] The time-segmented fault feature extraction module includes an insulation degradation coefficient calculation unit, a performance decay index calculation unit, and an electrical parameter mismatch detection unit.

[0113] The insulation degradation coefficient calculation unit utilizes the characteristic that the open-circuit voltage of a photovoltaic module is extremely sensitive to its internal insulation when it is close to open circuit under low irradiance. It calculates the average open-circuit voltage of the module during the low irradiance standby period and compares it with the health benchmark value. The "insulation degradation coefficient γ" is calculated by normalizing the relative deviation. This coefficient can effectively detect early insulation failure hazards caused by moisture intrusion, encapsulation aging, etc.

[0114] The performance degradation index calculation unit calculates the original performance ratio (PR) by comparing the measured power of the photovoltaic module with the ideal power when the photovoltaic module is in the linear operating period. raw Then, it is normalized again with the reference performance ratio of healthy components under the same environment to calculate the "performance degradation index D". This index directly quantifies the long-term power loss caused by material aging, battery efficiency decline, uniform contamination, etc., and is a key indicator for evaluating the durability of components.

[0115] When the photovoltaic modules are operating at saturation, the electrical parameter mismatch detection unit calculates the "voltage mismatch factor (VMF)" by measuring the relative deviation between the module voltage and the average string voltage; and calculates the "current consistency index (CCI)" by measuring the relative deviation between the module current and the average string current. These two factors are then weighted and combined into an "electrical parameter consistency index (F)," which is used to accurately locate series mismatch faults caused by hot spots, shading, poor connections, or severe inconsistencies in module performance.

[0116] The long-term condition factor statistics module establishes an "equipment file" for each photovoltaic module. It calculates the "operating years factor" by statistically analyzing the cumulative effective power generation time, reflecting the degree of natural aging of the equipment. It also calculates the "failure density factor" by statistically analyzing the number of historical failures, reflecting the inherent reliability of the equipment or weak links in operation and maintenance. These two long-term factors will provide important background corrections for the final comprehensive evaluation.

[0117] The long-term state factor statistics module includes an operating years factor calculation unit and a fault density factor calculation unit.

[0118] The operating life factor calculation unit calculates the total operating time T of each photovoltaic module under the "linear operating period". oper Compared with the current standard design life T of photovoltaic modules norm By comparison, the "service life factor τ" was calculated, which quantifies the natural wear and tear process of the equipment over time in a normalized form.

[0119] The fault density factor calculation unit traces all historical fault records of each photovoltaic module since it was put into operation and counts the total number of faults N. fault , and the number of design failures N thIn contrast, a "failure density factor Φ" is calculated, which reflects the frequency of failures and amplifies the risk weight of equipment that fails frequently.

[0120] After each day's operation, the comprehensive assessment and dynamic early warning module integrates fault characteristic indicators from three different working periods. These indicators include insulation degradation coefficient, performance decay index, voltage mismatch factor, and current consistency index, and incorporates factors such as service life and fault density. A comprehensive and objective "comprehensive degradation score S" is calculated using a weighted geometric average method. Based on this comprehensive degradation score S, the system executes a dynamic hierarchical early warning mechanism, classifying early warnings into "attention level," "abnormal level," and "severe level," and associating them with different operation and maintenance response strategies to form a complete closed loop from data analysis to operation and maintenance actions.

[0121] Example: For photovoltaic module A-01, within the time window of 10:00–14:00 on the same day, I(t) = 850W / m 2 The normalized power P(t) = 450W, and the calculated slope is k(t) = 0.529. The theoretical slope k is set as follows. STC =550 / 1000=0.55, let a1=0.2, a2=0.9, k0=0.6, a1×k STC =0.2×0.55=0.11, a2×k STC =0.9×0.55=0.495, because k(t)>a2×k STC This is considered a saturated work period;

[0122] Set the average voltage V of the component during the current saturation operating period. i =30.5V, string average voltage V avg =32.0V, calculated as follows: =0.469, setting the average current I of the component during the current saturation operating period. i =9.8A, average string current I avg =10.0A, calculated as follows: =0.133, set the weighting coefficients k1=k2=0.5, and obtain the voltage and current consistency index F=0.5×0.469+0.5×0.133=0.301;

[0123] Set the total operating time T of photovoltaic module A-01 oper =5 years, design life T norm =25 years, number of historical failures N fault =2, Design failure count N th =5, calculated as follows , ;

[0124] Set the insulation degradation coefficient of photovoltaic module A-01 during the low-irradiance standby period of the day. =0.72, performance degradation index D=0.036, c1=0.2, c2=0.5, c3=0.3, Q1=0.3, Q2=0.6, Q3=0.9, long-term factor coefficient α=0.1, the calculated comprehensive degradation score S=0.2523×1.4×(1+0.1×0.2)=0.36, which is between Q1(0.3) and Q2(0.6), the warning level is attention level, and component A-01 is added to the observation list and monitored more closely.

[0125] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A method for intelligent early warning of photovoltaic power station equipment based on multi-source data, characterized in that: The method includes the following steps: Step S100: Collect multi-source data on meteorology and electrical conditions from photovoltaic modules in the photovoltaic power station; Step S200: Within a continuous time window, calculate the real-time slope of DC power and irradiance of the photovoltaic module, and divide the time window into low irradiance standby period, linear working period and saturated working period in sequence by comparing with the theoretical slope. For photovoltaic modules with severe performance degradation in the linear period, trigger the highest priority warning. Step S300: During the low-irradiance standby period, calculate the insulation degradation coefficient of the photovoltaic module; during the linear operating period, calculate the performance degradation index of the photovoltaic module; during the saturated operating period, calculate the voltage mismatch factor and current consistency index of the photovoltaic module. Step S400: Calculate the operating life factor by calculating the cumulative effective power generation time of each photovoltaic module; calculate the fault density factor by calculating the historical number of faults of each photovoltaic module; Step S500: After the daily operation ends, the time period scores of low-irradiance standby period, linear working period and saturated working period are integrated, and the operating years factor and fault density factor are introduced to calculate the comprehensive degradation score for photovoltaic module fault early warning.

2. The intelligent early warning method for photovoltaic power station equipment based on multi-source data according to claim 1, characterized in that: Step S200 includes: Step S201: For each photovoltaic module, within a continuous time window t, the real-time slope k(t) between the DC power P(t) and irradiance I(t) of the photovoltaic module is normalized and calculated using a linear regression method. The real-time slope k(t) characterizes the instantaneous conversion efficiency of the photovoltaic module in the current window. Step S202: Calculate the theoretical slope k of the photovoltaic module model under standard conditions. STC By comparing the real-time slope k(t) with the theoretical slope k STC The comparison is used to make auxiliary judgments for the runtime segment, specifically: when k(t) < a1×k STC When a1×k is determined to be a low-irradiation standby period; STC ≤k(t)≤a2×k STC When k(t) > a2×k, it is determined to be a linear working period; STC When a certain time period is defined as a saturated working period, a1 and a2 are coefficient thresholds, where a1 < a2, and are set by professionals. Step S203: Within the current time window, when the real-time slope k(t) under the linear working period is continuously lower than k0×k STC If the duration exceeds θ% of the linear working period, it is marked as "severe performance degradation", generating the highest priority warning for immediate repair. θ and k0 are emergency thresholds set by professionals.

3. The intelligent early warning method for photovoltaic power station equipment based on multi-source data according to claim 1, characterized in that: Step S300 includes the following steps: Step S301: During the low-irradiance standby period, the photovoltaic module is close to the open-circuit state. Calculate the average open-circuit voltage V of the photovoltaic module. oc,avg Then compared with the reference voltage V oc,ref By comparing and normalizing the relative deviations between the two, the insulation degradation coefficient can be obtained. The average open-circuit voltage is the average DC voltage measured by the photovoltaic module during the low-irradiance standby period, and the reference voltage is the average DC voltage calculated by a healthy photovoltaic module under the same environment. Step S302: During the linear operating period, the photovoltaic module behaves stably. Calculate the ratio PR of the average DC power of the photovoltaic module to the current ideal power. raw The performance degradation index D is obtained, and the ideal power is the theoretical optimal output power of the photovoltaic module. Step S303: During the saturation operating period, the photovoltaic module operates in the high current output range. Calculate the relative deviation between the average DC voltage of the photovoltaic module and the average string voltage to obtain the voltage mismatch factor VMF. The relative deviation between the average current of the photovoltaic module and the average current of the string during the saturated operating period is analyzed to obtain the current consistency index (CCI). The voltage and current consistency index F is derived by combining the voltage mismatch factor and the current consistency index: F = k1 × VMF + k2 × CCI, where k1 and k2 are weighting coefficients, and k1 + k2 = 1.

4. The intelligent early warning method for photovoltaic power station equipment based on multi-source data according to claim 1, characterized in that: Step S400 includes the following steps: Step S401: Calculate the total operating time of each photovoltaic module under the linear operating period, compare the total operating time with the standard design life of the current photovoltaic module, and calculate the operating life factor of the current photovoltaic module; Step S402: Count the number of historical failures of each photovoltaic module since it was put into operation, and calculate the failure density factor of the current photovoltaic module by comparing it with the design failure number.

5. The intelligent early warning method for photovoltaic power station equipment based on multi-source data according to claim 1, characterized in that: Step S500 includes the following steps: Step S501: After the daily operation is completed, calculate the insulation degradation coefficient for the low-irradiance standby period, linear operating period, and saturated operating period. The performance degradation index D and the voltage and current consistency index F are combined using a weighted geometric average method to integrate the scores of the three time periods. Then, the service life factor and the fault density factor are introduced to calculate the comprehensive deterioration score S. Step S502: Based on the comprehensive degradation score S, implement a dynamic hierarchical early warning mechanism.

6. The intelligent early warning method for photovoltaic power station equipment based on multi-source data according to claim 1, characterized in that: Step S100 includes the following steps: Step S101: Simultaneously collect meteorological and electrical multi-source data in the photovoltaic power station to construct an original data pool. The meteorological multi-source data includes the planar irradiance I of the photovoltaic module and the backsheet temperature T. The electrical multi-source data includes the DC voltage, DC current and output power P of each photovoltaic string and module. Step S102: Align the timestamps of all data streams using a clock synchronization protocol, filter and remove abnormal data points, normalize the measured power to the baseline value under standard test conditions, and eliminate the interference of environmental fluctuations on performance comparison.

7. A photovoltaic power station equipment intelligent early warning system based on multi-source data, characterized in that: The system includes a multi-source data acquisition and preprocessing module, a dynamic slope analysis and time period segmentation module, a time period fault feature extraction module, a long-term state factor statistics module, and a comprehensive evaluation and dynamic early warning module. The multi-source data acquisition and preprocessing module is responsible for synchronously acquiring meteorological and electrical data from the photovoltaic power station site, and aligning the timestamps of all data streams through a clock synchronization protocol. The dynamic slope analysis and time period division module analyzes the real-time relationship between the output power and irradiance of each photovoltaic module, calculates the dynamic performance slope for each module, and compares it with the theoretical slope. The operating period is divided into "low irradiance standby period", "linear working period" and "saturated working period". The module has built-in emergency diagnosis logic, which can directly identify and mark modules with "severe performance degradation" and trigger the highest priority warning. The time-segmented fault feature extraction module calculates the insulation degradation coefficient of the photovoltaic module during the low-irradiance standby period; During the linear operating period, calculate the performance degradation index of the photovoltaic module; During the saturated operating period, calculate the voltage mismatch factor and current consistency index of the photovoltaic module; The long-term state factor statistics module calculates the "operating years factor" by statistically analyzing the cumulative effective power generation time; and calculates the "fault density factor" by statistically analyzing the number of historical faults. After each day's operation, the comprehensive assessment and dynamic early warning module integrates fault characteristic indicators from three different working periods. These indicators include insulation degradation coefficient, performance decay index, voltage mismatch factor, and current consistency index, and incorporates factors such as service life and fault density. A weighted geometric average method is used to calculate the "comprehensive degradation score S". Based on the comprehensive degradation score S, the system executes a dynamic hierarchical early warning mechanism, classifying early warnings into "attention level", "abnormal level", and "severe level", and associating them with different operation and maintenance response strategies.

8. The intelligent early warning system for photovoltaic power station equipment based on multi-source data according to claim 7, characterized in that: The dynamic slope analysis and time period segmentation module includes a dynamic performance slope calculation unit, a working time period intelligent segmentation unit, and a severely degraded equipment identification unit. The dynamic performance slope calculation unit calculates the real-time slope k(t) for each photovoltaic module by using a linear regression method to fit the relationship between its normalized DC power P(t) and irradiance I(t) within a continuous sliding time window. The intelligent division unit for working periods compares the real-time slope k(t) with the theoretical slope k of the photovoltaic module under standard conditions. STC The system performs a comparison and automatically divides the component's working state into three characteristic time periods based on preset coefficient thresholds: "Low-irradiation standby period", "linear working period", "saturation working period"; The severely degraded equipment identification unit continuously monitors the equipment during the linear operating period. When the photovoltaic module's slope k(t) remains below the emergency threshold k0*k at linear operating data points exceeding θ%, the unit detects the severely degraded equipment. STC If the component is directly identified as "severely degraded," the system will skip the subsequent comprehensive scoring process and immediately generate a high-priority repair warning.

9. The intelligent early warning system for photovoltaic power station equipment based on multi-source data according to claim 7, characterized in that: The time-segmented fault feature extraction module includes an insulation degradation coefficient calculation unit, a performance decay index calculation unit, and an electrical parameter mismatch detection unit. The insulation degradation coefficient calculation unit calculates the average open-circuit voltage of the component during the low-irradiation standby period and compares it with the health benchmark value. The "insulation degradation coefficient γ" is calculated by normalizing the relative deviation. The performance degradation index calculation unit calculates the original performance ratio (PR) by comparing the measured power of the photovoltaic module with the ideal power when the photovoltaic module is in the linear operating period. raw Then, it is normalized twice with the reference performance ratio of healthy components under the same environment to calculate the "performance degradation index D". When the photovoltaic module is operating at saturation, the electrical parameter mismatch detection unit calculates the relative deviation between the module voltage and the string average voltage to obtain the "voltage mismatch factor VMF"; it calculates the relative deviation between the module current and the string average current to obtain the "current consistency index CCI", and the two are weighted and merged into the "electrical parameter consistency index F".

10. The intelligent early warning system for photovoltaic power station equipment based on multi-source data according to claim 7, characterized in that: The long-term state factor statistics module includes an operating years factor calculation unit and a fault density factor calculation unit. The operating life factor calculation unit calculates the total operating time T of each photovoltaic module under the "linear operating period". oper Compared with the current standard design life T of photovoltaic modules norm By comparing the results, the "operating years factor τ" is calculated. The fault density factor calculation unit traces all historical fault records of each photovoltaic module since it was put into operation and counts the total number of faults N. fault , and the number of design failures N th In comparison, the "fault density factor Φ" is calculated.

Images