Photovoltaic system countercurrent prevention method

Abstract

The invention relates to the technical field of countercurrent prevention, in particular to a photovoltaic system countercurrent prevention method, which comprises the following steps of: periodically sampling the voltage of a grid-connected point where a photovoltaic system is connected with a power grid, continuously obtaining an instantaneous measurement value of the voltage, and establishing and updating an instantaneous voltage data cache; and based on the instant voltage data cache, extracting voltage data in the latest fixed time window, calculating the maximum change rate of the voltage amplitude in the time window, and judging and generating a voltage transient trigger signal. According to the invention, on the basis of continuous and periodic sampling of the grid-connected point voltage, an instant voltage data cache is established, and accurate capture of voltage disturbance is realized; by calculating the change rate of the voltage amplitude in the specific time window, the accurate moment when the voltage transient event occurs can be identified, and the real-time performance and accuracy of voltage transient detection are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of anti-counterflow, and particularly to an anti-counterflow method for a photovoltaic system. Background Art

[0002] The technical field of anti-counterflow refers to a series of prevention and control methods and measures proposed for the problem of reverse power flow of electric energy generated during the connection of distributed power sources such as photovoltaic and wind power to the power grid. When the grid load decreases and the power generation of distributed power sources exceeds the local load demand, the excess electric energy will be sent back to the grid, which may cause fluctuations in grid voltage and frequency and affect the operation stability of the grid.

[0003] Existing technologies usually rely on real-time monitoring of voltage, current and power flow directions for dynamic adjustment. However, in scenarios where the grid load suddenly changes or transient disturbances occur rapidly, a continuous and effective instant data cache fails to be established, resulting in the lack of continuity of monitoring data. At the same time, only the power flow direction is concerned while ignoring the duration of the disturbance event and the rate of change of the amplitude, which increases the probability of misjudgment in the determination of the type of disturbance source. As a result, the response mode of anti-counterflow measures is single and lacks flexibility, making it difficult to accurately distinguish between island and non-island disturbances. Therefore, improvements are needed. Summary of the Invention

[0004] The purpose of the present invention is to solve the deficiencies existing in the prior art and propose an anti-counterflow method for a photovoltaic system.

[0005] To achieve the above purpose, the present invention adopts the following technical solution. An anti-counterflow method for a photovoltaic system includes the following steps: Periodically sample the voltage at the grid connection point of the photovoltaic system and the grid, continuously obtain the instantaneous measurement values of the voltage, and establish and update an instant voltage data cache; based on the instant voltage data cache, extract the voltage data within the most recent fixed time window, calculate the maximum rate of change of the voltage amplitude within the time window, and determine whether to generate a voltage transient trigger signal; In response to the generated voltage transient trigger signal, lock the data of the voltage transient occurrence section corresponding to the trigger moment from the instant voltage data cache to form a locked transient voltage segment; based on the locked transient voltage segment, calculate the maximum peak value of the voltage amplitude within the segment and the duration for which the voltage exceeds the normal operating range to obtain a pair of core voltage disturbance parameters; According to the voltage peak amplitude and duration in the pair of core voltage disturbance parameters, match and compare them with a preset reference interval of island event characteristics, determine the degree of conformity between the disturbance event and the real island event, and establish an island event conformity index; based on the island event conformity index, evaluate the source of the current transient disturbance and determine the grid disturbance source classification code; Based on the obtained power grid disturbance source classification code, determine whether the current transient disturbance belongs to a non-island type of power grid fault, determine the response mode of the inverter, and generate an anti-counterflow regulation instruction.

[0006] Preferably, the step of obtaining the instantaneous voltage data cache is as follows: By setting a voltage sampling trigger period, call the voltage sensor at the grid connection point, continuously record the instantaneous measurement values of the grid connection point voltage obtained at each sampling trigger moment, and input all the measurement values into the data temporary storage area in sequence to obtain a sequence of instantaneous measurement values of the grid connection point voltage; According to the sequence of instantaneous measurement values of the grid connection point voltage, enter the voltage information item by item in chronological order, establish a voltage record table structure with time stamps, and perform data overwrite or append updates after each sampling is completed to generate a voltage measurement record set with a time index; Based on the voltage measurement record set with a time index, perform duplicate checking on the newly collected data, dynamically maintain the voltage record set arranged in chronological order, and generate an instantaneous voltage data cache.

[0007] Preferably, the step of obtaining the voltage transient trigger signal is as follows: Based on the instantaneous voltage data cache, trace back from the current time point to a preset fixed time period forward, extract all the voltage amplitude data and corresponding time stamp information arranged in sequence within the fixed time period, and form a voltage amplitude sequence and a time sequence within the fixed time window; According to the voltage amplitude sequence and the time sequence within the fixed time window, calculate the change rate fluctuation intensity of the voltage amplitude; Compare the change rate fluctuation intensity of the voltage amplitude with the set rate fluctuation critical value under stable operating conditions. If the change rate fluctuation intensity of the voltage amplitude is greater than the critical value, it is determined that a transient disturbance has occurred, and a voltage transient trigger signal is generated.

[0008] Preferably, the step of obtaining the locked transient voltage segment is as follows: Based on the trigger moment information carried in the voltage transient trigger signal, retrieve the voltage record position corresponding to the trigger moment in the instantaneous voltage data cache, determine the voltage record index position corresponding to the voltage transient trigger signal generation moment, and generate a voltage transient start index; According to the voltage transient start index, retrieve and extend forward and backward from the instantaneous voltage data cache respectively, track the process of the voltage transient amplitude deviating from the stable state for the first time and recovering to the stable state, determine the start and end index intervals of the voltage transient occurrence, and form a voltage transient occurrence section; Based on the voltage transient occurrence section, all the voltage amplitude data and the corresponding time data in the instant voltage data cache within the voltage transient occurrence section are jointly extracted and stored separately to construct an independent voltage amplitude and time information sequence, obtaining a locked transient voltage segment.

[0009] Preferably, the steps for obtaining the core voltage disturbance parameter pair are as follows: Based on the locked transient voltage segment, all the voltage amplitude data and the corresponding timestamp information within the segment are extracted to construct a structurally matching voltage amplitude sequence and timestamp sequence, obtaining a transient voltage amplitude sequence and a timestamp sequence; According to the transient voltage amplitude sequence and the timestamp sequence, calculate the maximum peak value and count the duration during which the voltage amplitude exceeds the normal operation range; Based on the maximum peak value extracted from the transient voltage amplitude sequence and the duration during which the voltage amplitude exceeds the normal operation range, combine them to form a correspondence between the voltage amplitude peak value and the duration, generating a core voltage disturbance parameter pair.

[0010] Preferably, the steps for obtaining the islanding event compliance index are as follows: According to the voltage amplitude peak value and the cumulative overlimit duration in the core voltage disturbance parameter pair, respectively retrieve the corresponding voltage amplitude threshold range and duration threshold range within the preset islanding event characteristic reference interval, and calculate the differences between the voltage amplitude peak value and the duration and the upper and lower thresholds of the reference interval one by one to obtain the voltage amplitude difference and the duration difference; Based on the voltage amplitude difference and the duration difference, taking the preset islanding event characteristic reference interval as the standard, calculate the relative proportion values of the two differences in their respective threshold intervals item by item, and divide the differences into three levels: high compliance, medium compliance, and low compliance according to the proportion values, obtaining the voltage amplitude matching level and the duration matching level; According to the voltage amplitude matching level and the duration matching level, call the comprehensive compliance level decision table for item-by-item matching, and use the lower value of the two levels as the comprehensive matching level, defining the comprehensive matching level as the compliance degree of the islanding event, generating an islanding event compliance index.

[0011] Preferably, the steps for obtaining the power grid disturbance source classification code are as follows: Based on the islanding event compliance index, extract the compliance level, and combine the voltage change rate and the amplitude jump direction within a fixed interval before and after the triggering moment to form a triple of the compliance level value, the voltage change rate value, and the amplitude change symbol, generating a disturbance characteristic input combination; According to the disturbance characteristic input combination, calculate the disturbance source type identification value; Based on the disturbance source type identification value, perform classification matching with the disturbance type identification reference interval to generate a power grid disturbance source classification code.

[0012] Preferably, the step of obtaining the anti-counterflow regulation instruction is as follows: According to the power grid disturbance source classification code, analyze the coded bit sequence in the power grid disturbance source classification code, interpret each coded bit as three identifiers of disturbance type, disturbance intensity level, and duration level, judge the value of the disturbance type identifier. If the identifier field does not belong to the valid coding range indicating an islanding event, it is determined that the current transient disturbance is a non-islanding type of power grid fault, and a power grid fault type determination result is generated; According to the power grid fault type determination result, combined with the current system operation state, extract the inverter working mode, power grid frequency state, and load power fluctuation value, use the three parameters as judgment conditions for conditional combination reasoning, select the inverter working mode corresponding to the matching conditions, and judge whether it is necessary to limit the inverter injection current, and generate an inverter response mode; Based on the inverter response mode, expand the control actions included in the response mode, list item by item the values set for active power output suppression, reactive power compensation amplitude adjustment, and grid-connected current phase shift, and combine all control action parameters to generate an anti-counterflow regulation instruction.

[0013] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, based on continuous periodic sampling of the grid connection point voltage, an instant voltage data cache is established, realizing precise capture of voltage disturbances; by calculating the voltage amplitude change rate within a specific time window, the exact moment when a voltage transient event occurs can be identified, improving the real-time performance and accuracy of voltage transient detection. On this basis, transient voltage segments are extracted, and disturbance parameters such as the maximum peak value of the voltage amplitude and the over-limit duration are obtained, improving the accuracy of voltage disturbance event feature extraction; further, multi-dimensional matching of core voltage disturbance parameters and the reference interval of islanding event characteristics is adopted to construct an islanding event compliance index, improving the recognition efficiency and accuracy of power grid disturbance types, and avoiding misjudgment of disturbance events; according to the classification of disturbance sources, through analysis and numerical response strategy output, a highly targeted anti-counterflow regulation instruction is dynamically generated, effectively avoiding unnecessary power losses caused by the blind response of the inverter, and enhancing the stable support ability and safety guarantee ability of distributed power sources for power grid operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a step schematic diagram of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] In order to make the objectives, technical solutions, and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0016] Please refer to Figure 1 , the present invention provides a technical solution, a method for preventing reverse current in a photovoltaic system, including the following steps: Periodically sample the voltage at the grid connection point where the photovoltaic system is connected to the grid, continuously obtain the instantaneous measurement values of the voltage, and establish and update the real-time voltage data cache; based on the real-time voltage data cache, extract the voltage data within the most recent fixed time window, calculate the maximum change rate of the voltage amplitude within the time window, and determine and generate a voltage transient trigger signal; In response to the generated voltage transient trigger signal, lock the data of the voltage transient occurrence section corresponding to the trigger moment from the real-time voltage data cache to form a locked transient voltage segment; based on the locked transient voltage segment, calculate the maximum peak value of the voltage amplitude within the segment and the duration for which the voltage exceeds the normal operating range, and obtain a pair of core voltage disturbance parameters; According to the voltage peak amplitude and duration in the pair of core voltage disturbance parameters, match and compare them with the preset reference interval of islanding event characteristics, determine the degree of coincidence between the disturbance event and the real islanding event, and establish an islanding event coincidence index; based on the islanding event coincidence index, evaluate the source of the current transient disturbance and determine the grid disturbance source classification code; Based on the obtained grid disturbance source classification code, determine whether the current transient disturbance belongs to a non-islanding type of grid fault, determine the response mode of the inverter, and generate a reverse current prevention adjustment instruction.

[0017] The steps for obtaining the real-time voltage data cache are as follows: By setting the voltage sampling trigger period, call the voltage sensor at the grid connection point, continuously record the instantaneous measurement values of the grid connection point voltage obtained at each sampling trigger moment, and sequentially input all the measurement values into the data temporary storage area to obtain a sequence of instantaneous measurement values of the grid connection point voltage; According to the sequence of instantaneous measurement values of the grid connection point voltage, enter the voltage information item by item in chronological order, establish a voltage record table structure with time stamps, and perform data overwrite or append updates after each sampling to generate a set of voltage measurement records with time indexes; Based on the set of voltage measurement records with time indexes, perform duplicate checking on the newly collected data, dynamically maintain the set of voltage records arranged in chronological order, and generate a real-time voltage data cache.

[0018] Specifically, by setting the voltage sampling trigger period, for example, for a grid with a nominal frequency of 50 Hz, in order to capture fast voltage fluctuations that may indicate early faults or disturbances, the sampling frequency is set to 200 times the fundamental frequency of the grid, that is, the voltage sampling trigger period is specifically set to Seconds, that is, 0.1 milliseconds. The selection of this period is based on the statistical analysis of historical power grid disturbance data. Specifically, it analyzes the rate of change of the voltage waveform at the initial stage of the disturbance event, and selects the minimum time interval that can ensure capturing the characteristics of at least 99.5% of the starting stage of such fast transient processes. According to this 0.1-millisecond voltage sampling trigger period, it drives the voltage sensor installed at the grid connection point to work. This voltage sensor can be an opto-isolated voltage transformer with a nanosecond-level response time, or a measurement unit composed of a high-precision voltage dividing resistor network and a high-speed 16-bit analog-to-digital converter (ADC). At each sampling trigger moment defined by this period, for example, at the absolute timestamp , the voltage sensor synchronously captures the instantaneous analog waveform values of the three-phase (or single-phase, depending on the system configuration) voltages at the grid connection point, and immediately converts them into digital quantities through the ADC. For example, at this moment, it is recorded as Ua = 220.12 volts, Ub = 219.58 volts, Uc = 220.31 volts. Immediately afterwards, at the next sampling moment , Ua = 220.05 volts, Ub = 219.63 volts, Uc = 220.25 volts are recorded again. These continuously acquired digital data points containing the instantaneous values of each phase voltage are sent one by one and in real time into a pre-allocated memory buffer with the characteristics of first-in-first-out (FIFO). This buffer serves as a data temporary storage area, and its capacity setting needs to be sufficient to accommodate all the sampling data of at least two complete power frequency cycles. Taking a 50-hertz power grid as an example, a power frequency cycle is 20 milliseconds, so the number of sampling points included in two cycles is . If each sampling point contains three-phase voltage values and each phase voltage is represented by a 4-byte floating-point number, then the buffer size is at least. All these sequentially arranged measurement values form a temporally continuous and ordered list of original voltage readings in the data temporary storage area, thus obtaining a sequence of instantaneous measurement values of the voltage at the grid connection point.

[0019] Based on the sequence of instantaneous measurement values of the voltage at the grid connection point obtained in the previous step, the system starts from the starting position of this sequence and extracts each instantaneous voltage measurement value one by one in strict chronological order. For each measurement value taken out from the sequence, for example, the instantaneous voltage value obtained at a specific sampling moment is (if it is a three-phase system, then represents a set of three-phase voltage values at this moment, such as ), the system will forcibly associate a high-precision timestamp with it. This timestamp comes from a hardware real-time clock (RTC) that is closely synchronized with the central processing unit (CPU). This RTC is synchronized with an external time server through the Network Time Protocol (NTP). For example, if a certain voltage value The acquisition occurs at the moment when the RTC shows the time as "18:01:38.123456 seconds on May 7, 2025", then this complete time information will be used as the timestamp of this data point. Subsequently, each information unit containing the precise timestamp and the corresponding instantaneous voltage value (or polyphase voltage values) is organized and entered into a structured data record. This data record appears as a logically tabular structure in memory, and each row represents the data at a sampling moment, including at least the following column fields: "Globally Unique ID" (for example, a 64-bit integer that monotonically increases starting from 1, used to uniquely identify each record), "Precise Timestamp" (represented as an integer in nanosecond precision of Coordinated Universal Time (UTC) or a specific high-precision floating-point number, such as 1678886400.123456789 seconds), "Phase A Voltage Value" (floating-point number, unit: volt), "Phase B Voltage Value" (floating-point number, unit: volt), "Phase C Voltage Value" (floating-point number, unit: volt), and an optional "Data Quality Flag Bit" (for example, a byte, used to mark whether the data has been verified, whether there are abnormalities, etc.). After each new instantaneous voltage measurement value and its corresponding timestamp data are prepared, the system will update this record set according to the preset storage management strategy. To ensure the timeliness of analysis and control memory occupancy, this set is implemented as a circular buffer with a fixed size, and its size is set to , for example, if it is necessary to analyze the power grid dynamics in the most recent 10 seconds and the sampling period is 0.1 milliseconds, then the record set will store . When a new data record enters, if the current number of records has reached , then the earliest stored record (i.e., the record with the smallest ID or the earliest timestamp) will be overwritten by the new record. This continuous input and update operation dynamically generates a set of voltage measurement records with time indices.

[0020] Based on the set of time-indexed voltage measurement records generated and continuously updated in the previous step, whenever a new voltage measurement record (including the precise timestamp and the corresponding instantaneous voltage value) is passed from the data acquisition process and is about to be added to the final real-time data cache, the system will first perform a duplicate check on this new record. The core logic of the duplicate check is mainly based on the uniqueness of the timestamp and the close comparison with the timestamps of the existing records in the set. Specifically, when executing, it will retrieve the record with the latest timestamp in the cache and compare the timestamp of the new record with that of this latest record. If the two are exactly the same, then the voltage values will be further compared. If the voltage values of all phases are also exactly the same (or the difference is within the basic noise fluctuation range of the instrument, for example, the absolute difference is less than 0.01 volts), then this new record is determined to be duplicate data and is discarded by the system. In addition, to handle the "pseudo-non-duplicate" situations that may be caused by slight jitters in system clock synchronization or extremely minor offsets in the sampling moments, a very small time difference tolerance threshold is set. , and its value is usually set according to the stability of the system clock and the expected maximum sampling jitter. For example, if it is known that the short-term stability of the system RTC is better than 5 microseconds, then it can be set to 0.005 milliseconds (i.e., 5 microseconds), which is much smaller than the sampling period of 0.1 milliseconds, ensuring that only real-time point conflicts are processed. If the absolute value of the difference between the timestamp of the new record and the timestamp of any existing record in the cache is less than this , and the corresponding instantaneous voltage values of each phase also show a high degree of similarity. Specifically, it is judged that the absolute average value of the voltage differences of each phase is less than a preset tiny voltage change threshold (this threshold is set based on the empirical value of 1.5 times the noise standard deviation of the sensor in the stable state plus one standard deviation. For example, if the noise standard deviation is 0.008 volts, then the threshold is ), then this new record may also be identified as potential duplicate or abnormal data and is processed according to the preset rules (such as, keeping the record with the earlier timestamp, or marking it for further manual verification). By default, the timestamp is used as the primary criterion to ensure the uniqueness of the data at each precise time point in the cache. After successfully passing the duplicate check, this newly confirmed valid and unique voltage record will be officially integrated into a memory data structure dedicated to fast reading and analysis. This structure constitutes the core of the real-time voltage data cache. The cache always maintains the characteristic that all records are strictly sorted in ascending order by timestamp internally. If this cache is implemented based on, for example, a circular array or a doubly linked list with a size of as mentioned above, then the safe addition of new data points (after duplicate checking) may trigger the removal of the oldest data point (if the cache is full), thus maintaining the constant size of the cache and the timeliness of the data. This series of duplicate checking and dynamic maintenance operations ensure the final output real-time voltage data cache.

[0021] The steps to obtain the voltage transient trigger signal are as follows: Based on the real-time voltage data cache, trace back from the current time point to a preset fixed time period forward, extract all the voltage amplitude data arranged in sequence and the corresponding timestamp information within the fixed time period, and form a voltage amplitude sequence and a time sequence within the fixed time window; According to the voltage amplitude sequence and the time sequence within the fixed time window, calculate the change rate fluctuation intensity of the voltage amplitude. The calculation formula is: ; Wherein, is the change rate fluctuation intensity of the voltage amplitude, is the voltage amplitude at the th moment, is the timestamp at the th moment, is the total number of voltage measurement values within the time window; Compare the change rate fluctuation intensity of the voltage amplitude with the set rate fluctuation critical value under stable operating conditions. If the change rate fluctuation intensity of the voltage amplitude is greater than the critical value, it is determined that a transient disturbance has occurred, and a voltage transient trigger signal is generated.

[0022] Specifically, based on the real-time voltage data cache generated in the previous system stage, the system performs an analysis every analysis cycle (for example, every 10 milliseconds). First, based on the current accurate system time point, trace back to a preset fixed time period in the historical data direction. The length of this fixed time period, for example, is set to 100 milliseconds, which is determined by statistical analysis of the duration characteristics of typical power grid transient events (such as voltage dips caused by short-circuit faults, large load switching, etc.). The goal is to ensure that the window can completely capture the core dynamic processes of most transient events while taking into account the calculation efficiency. Specifically, the system filters out all voltage records in the real-time voltage data cache whose timestamps fall within the interval from "the current time point" to "the current time point minus 100 milliseconds". Subsequently, from these filtered records, the voltage amplitude data contained in each record and the precisely corresponding sampling timestamp information are extracted. Among them, for the voltage amplitude data, if it is a three-phase system, it refers to the positive-sequence voltage amplitude or the voltage amplitude of a specific phase. These extracted voltage amplitude data and the corresponding timestamp information are arranged in the order of the timestamps, and no order adjustment or data interpolation operations are allowed to ensure that the original time sequence characteristics of the data are retained, thereby forming a voltage amplitude sequence and a corresponding time sequence containing continuous sampling points within the fixed time window.

[0023] Formula: , The advantage of the formula is that by calculating the fluctuation intensity of the voltage amplitude change rate, it can sensitively capture the severe or abnormal changes in the grid voltage. Its core idea is to quantify the "instability" or "acceleration" of the voltage change trend. Compared with only monitoring the voltage amplitude itself or a single change rate, it is more sensitive to the initial stage of voltage disturbance, especially for those complex transient processes where the change rate itself is also rapidly changing. For example, during the initial stage of a near - zone fault or islanding effect, not only will the voltage deviate from the normal value, but the deviation speed and acceleration will also show a specific pattern. Through the square - sum and square - root operations, similar to calculating the standard deviation, this formula can amplify and accumulate these second - order change effects, enabling even small fluctuations indicating potential instability to be effectively detected. At the same time, through normalization processing, the fluctuation intensities under different - length windows or different numbers of sampling points are made comparable to a certain extent.

[0024] The parameter is obtained as follows: represents the total number of voltage measurement values included in the current fixed time window. This value is directly determined according to the length of the voltage amplitude sequence within the fixed time window formed in the previous step. For example, if the set fixed time period is 100 milliseconds and the previously configured voltage sampling trigger period (i.e., the time interval between two adjacent sampling points) of the system is 0.1 milliseconds, then the total number of voltage measurement values in the time window is calculated by the formula: sample points. Here, adding 1 is because if both ends of the time period include sampling points, and in this embodiment, .

[0025] The parameter is obtained as follows: represents the voltage amplitude corresponding to the th sampling moment ( ranging from 1 to ) within the fixed time window. These voltage amplitude data are directly sourced from the voltage amplitude sequence formed within the fixed time window, which is extracted from the real - time voltage data cache. The unit of the voltage amplitude is volts. For example, for a case, the five consecutive voltage amplitudes extracted from the voltage amplitude sequence are: , , , , .

[0026] The parameter is obtained as follows: represents the sampling instants ( from 1 to ), the corresponding exact timestamps, which are directly derived from the time series within the formed fixed time window, and this sequence corresponds one-to-one with the voltage amplitude sequence. For example, for the above case, if the voltage sampling trigger period is 0.1 millisecond, and the starting timestamp of the window is 0 millisecond, then the corresponding timestamp sequence is: , , , , .

[0027] Calculation process: Taking , and the example data given in the previous parameter acquisition step as an example: V, ms; Substituting the parameters into the formula to calculate and obtain 41.4327. This result indicates that within the selected time window containing 5 sampling points, the fluctuation intensity of the voltage amplitude change rate is approximately 41.4327 volts per square millisecond. This value quantifies the severity of the change in the voltage change rate, that is, the average fluctuation amplitude of the voltage "acceleration". The larger the value, the more unstable the voltage change, and the more frequent or severe the change in the change trend.

[0028] According to the calculated fluctuation intensity of the voltage amplitude change rate , for example, its value is 41.4327 volts per square millisecond, compare it with a pre-set critical value of the rate fluctuation characterizing the stable operation state of the power grid. The setting of this rate fluctuation critical value is based on the values of a large number of historical power grid operation data under fault-free and undisturbed conditions through statistical analysis. The specific method is: collect the data under stable operating conditions for at least one continuous month or longer to form a sample set containing millions of values, calculate the average value and standard deviation of this sample set. The rate fluctuation critical value is then set as a threshold that can distinguish normal fluctuations from abnormal disturbances with a high confidence level (for example, 99.95%). For example, it is set as , where is a coefficient selected based on the balance of the desired detection sensitivity and false alarm rate. For example , if through the above statistics, the calculated average value of under the stable state is 0.5 volts per square millisecond, and the standard deviation is 0.2 volts per square millisecond, then the rate fluctuation critical value Volts per square millisecond, when the fluctuation intensity of the change rate of the latest calculated voltage amplitude (such as 41.4327 volts per square millisecond) is greater than this rate fluctuation critical value (such as 1.3 volts per square millisecond), the system determines that a voltage transient disturbance has occurred in the current power grid and immediately generates a voltage transient trigger signal, which is a Boolean type flag or an event object containing information such as the trigger time and the trigger value, etc., to initiate subsequent detailed disturbance analysis and response processes.

[0029] The steps for obtaining the locked transient voltage segment are as follows: Based on the trigger time information carried in the voltage transient trigger signal, retrieve the voltage record position corresponding to the trigger time in the immediate voltage data cache, determine the voltage record index position corresponding to the voltage transient trigger signal generation time, and generate a voltage transient start index; According to the voltage transient start index, retrieve and extend forward and backward from the immediate voltage data cache respectively, track the process of the voltage transient amplitude first deviating from the stable state and recovering to the stable state, determine the start and end index intervals of the voltage transient occurrence, and form a voltage transient occurrence section; Based on the voltage transient occurrence section, extract and store separately all the voltage amplitude data and the corresponding time data in the immediate voltage data cache within the voltage transient occurrence section, construct an independent voltage amplitude and time information sequence, and obtain the locked transient voltage segment.

[0030] Specifically, based on the precise trigger time information carried in the voltage transient trigger signal generated in the previous stage. For example, the signal indicates that a possible voltage transient was detected at the absolute system timestamp "18:01:38 and 123456 microseconds on May 7, 2025". Immediately utilize this trigger time information to perform a retrieval operation in the immediately updated voltage data cache that was previously constructed. This immediately updated voltage data cache is a collection that stores a large number of voltage measurement records with precise timestamps in chronological order. The goal of the retrieval is to locate the voltage record that exactly matches the trigger time timestamp given in the voltage transient trigger signal, or is the closest within an allowed extremely small time error (for example, set to one-tenth of the sampling period. If the sampling period is 0.1 milliseconds, the error tolerance is 0.01 milliseconds, and this tolerance is set according to the system clock synchronization accuracy and the jitter characteristics of the sampling device). The specific retrieval process can use a binary search algorithm (if the records in the cache are strictly sorted based on timestamps and support random access) or through a hash index (if the timestamp is used as a key value), or perform a one-by-one comparison in a simple sequential storage structure. Once this voltage record that exactly corresponds or is closest to the trigger time is found, the system extracts the storage location identifier of this record in the immediately updated voltage data cache. This identifier can be an array index, a linked list node pointer, or a unique row number of a database record. For example, if the immediately updated voltage data cache is an array containing 1 million records and the record corresponding to the trigger time is located at the 753421st position in the array, then this position 753421 is determined as the voltage record index position corresponding to the voltage transient trigger signal generation time, and a voltage transient start index is generated based on this.

[0031] Based on the voltage transient start index determined in the previous step, for example, the index value is 753421, which points to the data point where the transient was first detected in the immediately updated voltage data cache. The system uses this point as the center and conducts a stretching retrieval in both the forward (historical data) and backward (subsequent data) directions of the time series, aiming to define the complete duration of the entire voltage transient event. In specific operations, the system will check the voltage amplitude data before and after the voltage transient start index one by one and compare it with the upper and lower threshold values of the preset power grid voltage stable state. For example, for an AC system with a nominal RMS voltage of 230 volts, its corresponding peak value is approximately 325 volts. The lower limit of the voltage amplitude (here referring to the short-term RMS value or envelope value) during stable operation can be set to 90% of the nominal RMS voltage, that is , and the upper limit can be set to 110% of the nominal RMS voltage, that is , when the system retrieves data by tracing back from the voltage transient start index, it searches for the first sampling point where the voltage amplitude continuously remains within the above-mentioned stable range ([207V, 253V]) for a preset minimum stable duration (for example, lasting for 20 milliseconds, corresponding to a complete cycle of a 50Hz power grid. If the sampling period is 0.1 milliseconds, then 200 consecutive sampling points need to meet the condition). The index of this point is recorded as the actual start index of the transient occurrence. Similarly, the system retrieves data backward (in the direction of increasing time) from the voltage transient start index to find the first sampling point where the voltage amplitude re-enters and continuously remains within the stable range ([207V, 253V]) for the same preset minimum stable duration. The index of this point is then recorded as the actual end index of the transient occurrence. Through this two-way tracing and the judgment based on continuous stability, the start and end index ranges covered by the entire process of the voltage transient from starting to deviate from the stable state to finally fully recovering to the stable state are determined, thus forming the voltage transient occurrence section.

[0032] Based on the voltage transient occurrence section accurately determined in the previous step, this section is jointly defined by an actual start index and an actual end index. For example, the start index is 753000 and the end index is 755500, which means the transient event starts from the 753000th record in the real-time voltage data cache and ends at the 755500th record. Next, the system will browse all the data records within this section and completely copy out all the voltage amplitude data and their corresponding precise time data (timestamps) located between this start index and end index (including both) from the real-time voltage data cache. This extraction process ensures that the voltage information and time information of each sampling point during the transient are captured without omission. For example, if the voltage transient occurrence section covers 2501 sampling points (from index 753000 to 755500), then the system will extract these 2501 voltage amplitude readings and the corresponding 2501 timestamps. These extracted voltage amplitude data and timestamp data will then be organized into two independent but strictly time-ordered sequences or lists: a voltage amplitude sequence (for example, a floating-point array, storing them in sequence ), and a timestamp sequence (for example, a long integer or high-precision floating-point array, storing them in sequence ). These two newly constructed sequences are jointly stored in a dedicated memory area or data structure to form a data copy separated from the main cache. This copy only contains data related to the currently identified specific transient event, thus obtaining a locked transient voltage segment with clear structure and focused content.

[0033] The steps to obtain the core voltage disturbance parameter pair are as follows: Based on the locked transient voltage segment, extract all the voltage amplitude data and the corresponding timestamp information within the segment, construct a voltage amplitude sequence and a timestamp sequence with matching structures, and obtain the transient voltage amplitude sequence and the timestamp sequence; According to the transient voltage amplitude sequence and the timestamp sequence, calculate the maximum peak value and count the duration for which the voltage amplitude exceeds the normal operating range. The calculation formula is: ; where, represents the duration for which the voltage amplitude exceeds the normal operating range in the th transient, is the sampling time interval for each sampling point, is the voltage amplitude at the th sampling point, is the upper limit of voltage operation, is the lower limit of voltage operation, is the total number of sampling points in the voltage amplitude sequence; Based on the maximum peak value extracted from the transient voltage amplitude sequence and the duration for which the voltage amplitude exceeds the normal operating range, combine them to form the corresponding relationship between the voltage amplitude peak value and the duration, and generate the core voltage disturbance parameter pair.

[0034] Specifically, based on the locked transient voltage segment obtained in the previous stage, which contains all the instantaneous voltage amplitude data sampled at high frequency during a specific transient event and its corresponding timestamp information, the system first extracts these two corresponding original data sequences from the segment, namely the original instantaneous voltage amplitude sequence and the original timestamp sequence. Then, to facilitate the assessment of voltage quality in line with industry standards, the system processes the original instantaneous voltage amplitude sequence to construct a transient voltage amplitude sequence that is more suitable for characterizing the severity of voltage events in terms of structure. The specific processing method is to divide the locked transient voltage segment into a number of consecutive complete cycles or half - cycle windows in time according to the preset nominal grid frequency (for example, 50 Hz, with a corresponding period of 20 milliseconds). For each such window, calculate the root mean square (RMS) value of all the instantaneous voltage amplitudes within the window. For example, if the segment duration is 100 milliseconds and the sampling period is 0.1 milliseconds, there are 1000 instantaneous sample points in the segment. Calculate one RMS value every 20 milliseconds (200 sample points), and 5 RMS voltage amplitudes will be obtained. These calculated per - cycle (or per - half - cycle) RMS values constitute the new "transient voltage amplitude sequence". At the same time, generate a corresponding timestamp for each calculated RMS value, which can be the end moment or the central moment of the corresponding calculation window, thus forming a matching "transient timestamp sequence". These two sequences together form the transient voltage amplitude sequence and the timestamp sequence required for subsequent analysis.

[0035] Formula: The advantage of the formula is that it provides a standardized and quantifiable method to evaluate the severity of voltage transient events, especially by precisely calculating the cumulative duration during which the voltage magnitude exceeds the normal operating range , which directly reflects the time that electrical equipment endures abnormal voltage and is one of the key indicators for judging whether the equipment may be damaged or whether the system needs to take protective measures. The conditional judgment in the formula clearly defines what is "beyond the normal operating range". Combining the operation of multiplying the indicator function by 1, the complex conditional judgment is transformed into simple counting, making the calculation of the total duration intuitive and easy to implement.

[0036] Parameter is obtained as follows: represents the time interval or duration represented by each sampling point in the transient voltage magnitude sequence when calculating . In the current context, since is the root mean square (RMS) value calculated from the original high-frequency sampled instantaneous voltage data within a fixed time window (for example, one power grid cycle), so is the length of this fixed time window used to calculate the RMS value. For example, if the power grid frequency is 50 Hz and its period is 20 milliseconds, and it is set to calculate one RMS voltage magnitude per complete cycle , then has a value of 20 milliseconds, that is seconds.

[0037] Parameter is obtained as follows: represents the voltage magnitude of the th sampling point in the transient voltage magnitude sequence, where each value is the root mean square (RMS) value calculated from the original high-frequency instantaneous voltage within the time window (for example, one power grid cycle), in volts (V). This sequence represents the per-cycle (or per-window) variation of the effective voltage during the transient event. For example, for a transient event lasting 100 milliseconds, if , then values will be obtained. A specific example sequence is: , , , , .

[0038] Parameter is obtained as follows: Represents the upper threshold value of the allowable voltage amplitude (here referring to the RMS value) during normal operation of the power grid. The setting of this value is strictly based on power quality standards. For example, for a low-voltage distribution system with 220V (phase voltage) or 230V (phase voltage, standard in some regions), the allowable long-term voltage deviation is usually or of the nominal voltage. Here, is selected as an example. If the nominal phase voltage is 230V, then is calculated as: .

[0039] Parameter is obtained as follows. Represents the lower threshold value of the allowable voltage amplitude (here referring to the RMS value) during normal operation of the power grid. Similar to , the setting of this value is also based on relevant power quality standards, using the same reference standards and nominal voltage as . If the nominal phase voltage is 230V and the allowable negative deviation is 10%, then is calculated as: .

[0040] Parameter is obtained as follows. Denotes the total number of sampling points in the transient voltage amplitude sequence used for calculating , that is, the length of this sequence. This value is jointly determined by the total duration of the transient event and the time window used to calculate each value. Specifically, is equal to the total duration of the locked transient voltage segment divided by , and then rounded (usually rounded up or determined according to the specific window division logic). For example, if the total duration of a transient segment is 100 milliseconds and is set to 20 milliseconds (one power grid cycle), then sampling points.

[0041] The "maximum peak value" refers to finding the maximum RMS value from the transient voltage amplitude sequence , denoted by . The calculation method is to traverse all elements in the sequence and find its maximum value. For example, for the sequence V, its maximum peak value .

[0042] Calculation process: Taking , , , as an example, the transient voltage amplitude sequence is: V.

[0043] Calculate : For : (false) (false). The condition is false. The contribution is 0. For : (false) (true). The condition is true. The contribution is 1. For : (false) (true). The condition is true. The contribution is 1. For : (false) (false). The condition is false. The contribution is 0. For : (false) (false). The condition is false. The contribution is 0.

[0044] Sum .

[0045] Calculate : ; i.e., 40 milliseconds. This result indicates that in the transient event of this example, the cumulative duration during which the voltage amplitude (RMS value) exceeds the preset normal operating range (207V to 253V) is 40 milliseconds, and this value is a key parameter for quantifying the severity of the transient disturbance. Combining with the maximum peak (RMS) calculated previously, they jointly constitute the core description of the amplitude and duration of this transient event.

[0046] Based on the maximum peak extracted from the transient voltage amplitude sequence in the previous step (this sequence is composed of RMS voltage values calculated period by period or window by window), for example volts, and the duration during which the voltage amplitude exceeds the normal operating range calculated through the formula , for example milliseconds, the system combines these two key quantification metrics. Specifically, in implementation, these two numerical values are paired and stored to form a data structure or object containing two elements. For example, it can be represented as a tuple: Or a record with explicit field names (such as max_rms_peak and duration_outside_limits). In this example, the corresponding relationship formed is (240.0 volts, 40 milliseconds). This structure clearly associates the most severe voltage amplitude deviation of the transient event (represented in the form of RMS peak) with the total time it exceeds the normal limit. Without any further complex transformations or statistical processing, simply placing these two directly calculated and extracted parameters side by side generates the core voltage disturbance parameter pair.

[0047] The steps to obtain the islanding event compliance index are as follows: According to the voltage amplitude peak and the cumulative over-limit duration in the core voltage disturbance parameter pair, respectively retrieve the corresponding voltage amplitude threshold range and duration threshold range within the preset islanding event characteristic reference interval. Calculate the difference between the voltage amplitude peak and the duration and the upper and lower thresholds of the reference interval one by one to obtain the voltage amplitude difference and the duration difference; Based on the voltage amplitude difference and the duration difference, taking the preset islanding event characteristic reference interval as the standard, calculate the relative proportion values of the two differences in their respective threshold intervals item by item. Divide the differences into three levels: high compliance, medium compliance, and low compliance according to the proportion values to obtain the voltage amplitude matching level and the duration matching level; According to the voltage amplitude matching level and the duration matching level, call the comprehensive compliance level decision table for item-by-item matching. Use the lower value of the two levels as the comprehensive matching level, define the comprehensive matching level as the compliance degree of the islanding event, and generate the islanding event compliance index.

[0048] Specifically, based on the core voltage disturbance parameter pairs obtained in the previous stage, which include the peak voltage amplitude in this transient event (for example, the calculated peak effective value is 240.0 volts) and the cumulative over-limit duration when the voltage amplitude exceeds the normal operating range (for example, 40 milliseconds), the system first retrieves the "preset islanding event characteristic reference interval" from a pre-configured parameter library. This reference interval is comprehensively defined based on statistical analysis of a large number of known real islanding event data and in combination with relevant industry standards (such as the description of voltage and frequency behavior in the islanding state in the requirements for distributed power grid connection operation in IEEE1547), aiming to delimit the characteristic range of voltage amplitude and duration most likely to be exhibited by typical islanding events. This reference interval specifically includes two parts: one is the "voltage amplitude threshold range", for example, set as [215.0 volts, 245.0 volts] (this range indicates that the voltage in the initial stage of islanding may be in an interval close to but slightly offset from the normal value. This value is set based on the common voltage fluctuation characteristics when a 230-volt nominal voltage system is independently powered by distributed power after losing the support of the main grid. The lower limit is the nominal voltage minus about 6.5%, and the upper limit is the nominal voltage plus about 6.5%), and the other is the "duration threshold range", for example, set as [0.5 seconds, 2.0 seconds] (this range is set based on the fact that the islanding detection algorithm usually needs to respond within a few seconds, while excluding extremely short transient disturbances and considering the possible duration before the islanding becomes unstable). Subsequently, the system compares the peak voltage amplitude (240.0 volts) in the core voltage disturbance parameter pair with the upper and lower limits of the retrieved voltage amplitude threshold range [215.0 volts, 245.0 volts] respectively and calculates the difference. Specifically, if the measured value is within the interval, the difference is defined as 0, and if it is outside the interval, the difference is defined as the absolute difference between the measured value and the nearest interval boundary. For example, if the peak voltage amplitude of 240.0 volts is within the interval [215.0 volts, 245.0 volts], the voltage amplitude difference is 0 volts. Similarly, the cumulative over-limit duration (40 milliseconds, i.e., 0.04 seconds) is compared with the upper and lower limits of the duration threshold range [0.5 seconds, 2.0 seconds]. Since 0.04 seconds is less than the lower limit of 0.5 seconds, the difference is calculated as , thus obtaining the voltage amplitude difference and the duration difference respectively.

[0049] Based on the voltage amplitude difference (for example, 0 volts) and the duration difference (for example, 0.46 seconds) calculated in the previous step, and referring to the "preset islanding event characteristic reference interval" as the evaluation criterion, the system then calculates the relative proportion values of these two differences respectively accounting for the widths of their corresponding characteristic reference intervals. First, determine the widths of each reference interval: the width of the voltage amplitude threshold range [215.0 volts, 245.0 volts] is , the width of the duration threshold range [0.5 seconds, 2.0 seconds] is , the relative proportion value of the voltage amplitude difference is calculated as the voltage amplitude difference divided by the width of the voltage amplitude reference interval, that is , the relative proportion value of the duration difference is calculated as the duration difference divided by the width of the duration reference interval, that is , then, according to the preset level classification rules, the system classifies each difference (actually the deviation degree) into three levels: high compliance, medium compliance, and low compliance according to these calculated relative proportion values. The level classification rules are as follows (the smaller the proportion value, the more in line with the characteristics): for the relative proportion values of voltage amplitude and duration, if the proportion value is less than or equal to 0.1 (that is, the difference does not exceed 10% of the corresponding interval width), it is rated as "high compliance"; if the proportion value is greater than 0.1 and less than or equal to 0.3 (that is, the difference is between 10% and 30% of the interval width), it is rated as "medium compliance"; if the proportion value is greater than 0.3 (that is, the difference exceeds 30% of the interval width), it is rated as "low compliance". These thresholds (0.1 and 0.3) are set based on the sensitivity analysis of the distribution of islanding event characteristic parameters and combined with expert experience, aiming to effectively distinguish different degrees of matching situations. According to this rule, the relative proportion value of the voltage amplitude is 0, so the voltage amplitude matching level is "high compliance", and the relative proportion value of the duration is approximately 0.3067, so the duration matching level is "low compliance". Finally, the voltage amplitude matching level and the duration matching level are obtained.

[0050] According to the voltage amplitude matching level (for example, "high compliance") and the duration matching level (for example, "low compliance") obtained in the previous step of evaluation, the system then calls an internally stored "comprehensive compliance level decision table" or executes an equivalent logical rule to comprehensively evaluate these two single-item matching levels. The construction principle of this decision table is that only when all the characteristic parameters participating in the evaluation show highly similar characteristics to the typical islanding event, the overall compliance is considered high. In this method, the specific comprehensive rule adopted is to take the lower level of the voltage amplitude matching level and the duration matching level as the final comprehensive matching level. For example, if "high compliance" is assigned 3 points, "medium compliance" is assigned 2 points, and "low compliance" is assigned 1 point, then the score of the voltage amplitude matching level is 3 points, and the score of the duration matching level is 1 point. The lowest value of the two is 1 point, corresponding to "low compliance". Therefore, the comprehensive matching level is "low compliance". This comprehensive matching level is defined as a quantitative indicator of the degree of compliance between the current transient disturbance event and the true islanding event. Finally, based on this comprehensive matching level ("low compliance"), the system generates the islanding event compliance indicator for this analysis.

[0051] The steps to obtain the power grid disturbance source classification code are as follows: Based on the islanding event compliance index, extract the compliance level, combine the voltage change rate and the amplitude jump direction within a fixed interval before and after the trigger moment, form a triple of the compliance level value, the voltage change rate value, and the amplitude change sign, and generate a disturbance feature input combination; According to the disturbance feature input combination, calculate the disturbance source type identification value, and the calculation formula is: ; Wherein, is the disturbance source type identification value, is the number of disturbance feature points, is the th moment's instantaneous voltage amplitude increment, is the th moment's islanding event compliance level, is the compliance level reference constant; Based on the disturbance source type identification value, perform classification matching with the disturbance type identification reference interval to generate a power grid disturbance source classification code.

[0052] Specifically, based on the islanding event compliance index generated in the previous stage, this index quantifies the similarity between the current voltage disturbance and the characteristics of typical islanding events. For example, when the compliance is obtained as "low compliance", the system first extracts this compliance level and numericalizes it. Specifically, referring to the preset mapping rule: map "high compliance" to the numerical value 3, "medium compliance" to the numerical value 2, and "low compliance" to the numerical value 1. Therefore, the compliance level value in the current example is 1. Subsequently, the system combines the precise trigger moment information recorded in an earlier stage (i.e., when the voltage transient was first triggered), and defines a symmetric fixed time interval around this trigger moment. For example, this fixed interval is set to be from 50 milliseconds before the trigger moment to 50 milliseconds after the trigger moment, totaling 100 milliseconds. The selection of this time interval is based on the fact that the key dynamic characteristics of most power grid disturbances (especially in the initial stage of faults and island formation) are usually most significant within dozens to one hundred milliseconds after the disturbance occurs. The system extracts the original high-frequency sampled instantaneous voltage amplitude data within this 100-millisecond interval from the real-time voltage data cache, and selects several (for example, ones, and select one feature point every 10 milliseconds) disturbance feature points within this interval. For each feature point (from 1 to ), calculate the "instantaneous voltage amplitude increment" of this point relative to its immediately previous sampling point. At the same time, record or determine the "amplitude jump direction" near each feature point (i.e., 's sign, positive for increase, negative for decrease), and combine the overall event compliance level value extracted previously (such as 1) with the instantaneous voltage amplitude increment calculated at each feature point (e.g., sequence ) and the corresponding amplitude change symbol (e.g., sequence ) together form a set of triples containing groups of information (or equivalently, each feature point corresponds to a triple ), and this set is the generated perturbed feature input combination.

[0053] Formula: , the advantage of the formula is that it combines the dynamic characteristics of voltage disturbance (reflected by the instantaneous voltage amplitude increment ) with the static compliance characteristics of events (reflected by the difference between the islanding event compliance level and the reference ) to achieve comprehensive identification of the type of disturbance source. The denominator term plays a regulating role: when the islanding compliance level of the event is closer to the preset reference representing typical islanding characteristics, the denominator is closer to 1, making the dynamic change of voltage contribute more significantly to ; conversely, if the islanding compliance of the event is very low (i.e., is far from ), the denominator will increase, thus weakening the influence of . This design enables the value to more accurately reflect the nature of the disturbance. If a disturbance is dynamically intense (large ) but its characteristics do not match those of islanding well, its value will be effectively suppressed. Conversely, even if the dynamic change is not extreme, but if it highly conforms to the islanding characteristics, its contribution will be retained. By averaging multiple feature points, the robustness of the recognition result is enhanced, effectively distinguishing whether it is an islanding event or other types of power grid faults; The steps to obtain the parameter are as follows. represents the total number of disturbance feature points used to calculate the disturbance source type recognition value . These feature points are selected within a fixed time interval defined around the voltage transient trigger moment. For example, this fixed time interval is 100 milliseconds. If it is set to uniformly select feature points for analysis at a step of 10 milliseconds within this interval, then the number of selected feature points points; in this embodiment, feature points are selected.

[0054] The steps to obtain the parameter are as follows. represents at the th disturbance feature point moment ( From 1 to ), the observed instantaneous voltage amplitude increment, in volts (V), which is calculated by comparing the instantaneous voltage amplitude at the moment of this characteristic point with the instantaneous voltage amplitude at the immediately preceding original sampling moment, i.e., , where is the instantaneous voltage amplitude at the characteristic point, is the high-frequency sampling period of the system (e.g., 0.1 milliseconds), and these values directly reflect the microscopic dynamic changes of the voltage near the characteristic point. For example, for cases, the three consecutive voltage instantaneous amplitude increments extracted from the generated disturbance characteristic input combination may be: , , , which means the voltage drops by 15 volts at the first characteristic point, 25.5 volts at the second point, and rises by 8.2 volts at the third point.

[0055] Parameter is obtained as follows. represents the numerical representation of the islanding event compliance level evaluated at the th disturbance characteristic point. This level value is derived from the extraction and quantification of the "islanding event compliance index" for the entire transient event, i.e., "high compliance" corresponds to 3, "medium compliance" corresponds to 2, "low compliance" corresponds to 1. Since the "islanding event compliance index" is a single evaluation result given for the entire identified transient event, in the summation term of the current formula, for all characteristic points (from 1 to ), values are the same, i.e., . For example, if the obtained event compliance level is "low compliance", its numerical value is 1. Therefore, for all , , this parameter reflects the similarity of the current disturbance to the islanding in macroscopic characteristics.

[0056] Parameter is obtained as follows. is a preset compliance level reference constant, dimensionless, which is used as a reference point to evaluate the relative position of the actual islanding event compliance level . Its value is set to optimize 's ability to distinguish different disturbance types. The setting strategy is to set to the value representing a "highly compliant" islanding event so that when is also "highly compliant", the term is minimized, thus maximizing for Regarding the influence, if "high compliance" is mapped to the value 3, it can be set .

[0057] Calculation process: Take , as an example. From the above parameter acquisition steps, we have: V, (low compliance).

[0058] Calculate term: Since and is the same for all , so . . .

[0059] Now calculate each term of the sum : For , for , for ; Sum ; Calculate ; This result indicates that after comprehensively considering the instantaneous dynamic changes of the voltage and the compliance between the overall event and the island characteristics, the calculated identification value of the disturbance source type is approximately -4.815. This negative value and its specific magnitude are the key numerical bases for distinguishing the disturbance source type. A larger negative value may indicate a certain type of power grid fault (such as a rapid voltage drop), while a larger positive value may indicate another type of disturbance or a more island-like characteristic (depending on the design intention and subsequent classification criteria).

[0060] Based on the identification value of the disturbance source type calculated in the previous step, for example, its value is -4.815. Next, it will be classified and matched with a pre-defined set of "disturbance type identification reference intervals". This reference interval set is divided and calibrated based on the value statistical distribution characteristics of a large number of known types of power grid disturbance events (including various faults, operation disturbances, and real and simulated island events). Each interval corresponds to a specific power grid disturbance source classification code. For example, the reference intervals and corresponding classification codes may be set as follows (these thresholds are empirically determined based on historical data analysis and simulation results to ensure the distinguishability of the value distributions of various events): If , the classification code is "GF01" (indicating a severe voltage dip type power grid fault). If , the classification code is "GF02" (indicating a general voltage sag or short-term fluctuation type grid fault). If , the classification code is "UD00" (indicating an undefined or minor disturbance, non-atypical islanding and non-atypical fault). If , the classification code is "IS01" (indicating suspected islanding or specific non-fault disturbance). If , the classification code is "IS02" (indicating a highly suspected islanding event). Match the currently calculated with the above intervals. Since , it falls into the interval corresponding to the classification code "GF02". Based on this, the system generates the grid disturbance source classification code for this disturbance as "GF02".

[0061] The steps to obtain the anti-islanding regulation instruction are as follows: According to the grid disturbance source classification code, analyze the coding bit sequence in the grid disturbance source classification code, interpret each coding bit as three identifiers: disturbance type, disturbance intensity level, and duration level. Judge the value of the disturbance type identifier. If the identifier field does not belong to the valid coding range indicating islanding events, it is determined that the current transient disturbance is a non-islanding type grid fault, and a grid fault type determination result is generated; Based on the grid fault type determination result and combined with the current system operating state, extract the inverter operating mode, grid frequency state, and load power fluctuation value. Use these three parameters as judgment conditions for conditional combination reasoning, select the inverter operating mode corresponding to the matching conditions, and determine whether to limit the inverter injection current to generate an inverter response mode; Based on the inverter response mode, expand the control actions included in the response mode, list item by item the values set for active power output suppression, reactive power compensation amplitude adjustment, and grid-connected current phase shift. Combine all control action parameters to generate an anti-islanding regulation instruction.

[0062] Specifically, based on the grid disturbance source classification code generated in the previous stage, for example, the obtained classification code is "GF02". The system first performs a structured analysis on this code. For example, the "grid disturbance source classification code" adopts a four-character structure, where the first two letters represent the major category of the disturbance, and the last two digits represent the disturbance intensity and duration characteristics or sub-classifications respectively. Therefore, for "GF02", the system analyzes it as follows: the first two characters "GF" are used as the disturbance type identifier, representing "Grid Fault"; the third digit "0" is used as the disturbance intensity level identifier. For example, 0 represents low intensity, 1 represents medium intensity, and 2 represents high intensity; the fourth digit "2" is used as the disturbance duration or specific sub-type level identifier. For example, 0 represents instantaneous, 1 represents short-term, and 2 represents persistent or specific fault sub-class. Next, the system makes a value judgment on the parsed disturbance type identifier "GF" and compares it with the preset "valid coding range representing islanding events". This valid coding range is preset according to system definitions and relevant standards to uniquely identify islanding or suspected islanding events. For example, all codes starting with "IS" (IslandSuspected / Detected) (such as "IS01", "IS02") are defined as valid codes representing islanding events. Since the currently parsed disturbance type identifier "GF" is not within the coding range starting with "IS", the system determines that the current transient disturbance is a non-islanding type of grid fault and generates a clear grid fault type determination result of "non-islanding type of grid fault" accordingly.

[0063] Based on the grid fault type determination result generated in the previous step, confirming that the current disturbance is a "non-islanding type of grid fault", the system further combines the current real-time overall operating state information of the photovoltaic system to determine the specific response strategy of the inverter. For this purpose, the system extracts the following three key current system operating state parameters from the supervisory control and data acquisition system: The first item is the "inverter operating mode". For example, by reading the internal status register of the inverter, it is known that the current mode is "maximum power point tracking (MPPT) mode"; the second item is the "grid frequency status". The grid-connected point frequency is measured in real time by a high-precision frequency meter. For example, the currently measured frequency is 49.6 Hz, and the status is determined based on a preset frequency threshold. This threshold is based on grid operation standards (such as the allowable frequency deviation of the State Grid is usually 50 Hz Hz, and can be relaxed to If the frequency is set in hertz, for example, less than 49.5 Hz is defined as "too low frequency", 49.5 to 50.5 Hz is "normal frequency", and higher than 50.5 Hz is "too high frequency", then the current 49.6 Hz belongs to the state of "slightly lower than normal frequency"; the third item is "load power fluctuation value", which is calculated by monitoring the change in the net exchange power at the point of common coupling over a short period of time (for example, the last 1 second) as a percentage of the inverter's rated power. For example, if the calculated fluctuation value is +15% (indicating a sudden reduction in load or a relative increase in PV output), and it is compared with the preset fluctuation thresholds (for example, low fluctuation: less than 5%, medium fluctuation: 5% to 20%, high fluctuation: greater than 20%, these thresholds are set based on local load characteristics and inverter regulation ability experience). The current fluctuation belongs to "medium fluctuation". Subsequently, the system takes the determination result of the "non-island type grid fault" that has been confirmed, as well as the three parameters of the "inverter operating mode (MPPT)", "grid frequency state (normal and slightly lower)", and "load power fluctuation value (medium fluctuation)" extracted and determined as input conditions, and applies the built-in conditional combination reasoning logic (usually a set of IF-THEN-ELSE rule sets, which are pre-programmed by grid safety experts and inverter control engineers according to relevant grid connection guidelines and anti-islanding requirements). For example, a rule might be: "If it is a non-island grid fault, and the frequency is normal, and the load power has a medium positive fluctuation (the trend of net power flowing into the grid increases), then the inverter should slightly suppress the active power output and check the reactive power demand". By matching these conditions, the system selects an inverter operating mode that best suits the current situation and determines whether and to what extent the grid-connected injection current of the inverter needs to be restricted, and finally generates an inverter response mode, such as "active power slightly suppression mode combined with voltage support".

[0064] Based on the inverter response mode generated in the previous step, for example, determined as "active power slightly suppression mode combined with voltage support", the system further specifies this abstract response mode into a set of directly executable control parameters and actions. First, the system parameterizes and expands the various standard control actions included in this response mode, that is, it searches for the preset parameter sets associated with the "active power slightly suppression mode combined with voltage support" mode. These parameter sets are predefined and stored in the inverter control logic according to the inverter design specifications, grid connection protocols, and a large number of simulation test results. For example, "slightly suppressing the active power" may be specifically quantified as reducing the current active power output by 10% of its rated value, or reducing it to an absolute safety upper limit value. For example, if the current output is 50 kW and the rated value is 100 kW, the target output after suppression is , or directly set to not exceed 40 kW. "Voltage support" may be expanded to adjust the reactive power output to help stabilize the grid voltage. For example, according to the current grid connection point voltage measurement, if the voltage is low, the inverter is commanded to generate inductive reactive power, and its amplitude may be set to 5% of the rated apparent power, that is, output 5 kVA of reactive power. At the same time, the phase shift of the grid-connected current is also accurately calculated and set according to the required active and reactive power combinations. For example, when the target active power is 45 kW and the reactive power is 5 kVA, the corresponding current amplitude and phase angle reference values are calculated. The system lists these specific setting values item by item: the target value of the active power output (such as 45 kW), the target kVA value of the reactive power compensation (such as 5 kVA), and the adjusted grid-connected current phase angle (for example, lagging behind the voltage by degrees), and then combines and packages all these detailed control action parameters (active power setpoint, reactive power setpoint, current phase setting, etc.) to form a structured data instruction. This instruction is the finally generated anti-counterflow regulation instruction and is sent to the underlying control execution unit of the inverter.

[0065] The above are only the preferred embodiments of the present invention and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change, and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution content of the present invention still fall within the protection scope of the technical solution of the present invention.

Claims

1. A photovoltaic system backflow prevention method, characterized in that: The following steps are involved: Periodically sample the voltage at the grid connection point where the photovoltaic system is connected to the grid, continuously obtain instantaneous voltage measurements, and establish and update real-time voltage data cache; Based on the instant voltage data cache, extract the voltage data in a recent fixed time window, calculate the maximum rate of change of the voltage amplitude in the time window, and determine and generate a voltage transient trigger signal; In response to the generated voltage transient trigger signal, locking the data of the voltage transient occurrence section corresponding to the triggering moment from the instant voltage data cache to form a locked transient voltage segment; Based on the locked transient voltage segment, the maximum peak value of the voltage amplitude in the segment and the duration of the voltage exceeding the normal operating range are calculated to obtain a core voltage disturbance parameter pair; According to the voltage peak amplitude and duration in the core voltage disturbance parameter pair, the preset island event characteristic reference interval is matched and compared, the consistency between the disturbance event and the real island event is judged, and the island event consistency index is established; based on the island event consistency index, the source of the current transient disturbance is evaluated, and the grid disturbance source classification code is determined; According to the obtained grid disturbance source classification code, it is judged whether the current transient disturbance belongs to a non-islanding type grid fault, the response mode of the inverter is determined, and an anti-backflow regulation instruction is generated.

2. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps of obtaining the instant voltage data cache are: By setting the voltage sampling trigger cycle, calling the voltage sensor at the grid connection point, continuously recording the instantaneous measurement value of the grid connection point voltage obtained at each sampling trigger moment, and inputting all the measurement values ​​into the data temporary storage area in sequence, obtaining the instantaneous measurement value sequence of the grid connection point voltage; According to the instantaneous measurement value sequence of the grid connection point voltage, the voltage information is entered one by one in chronological order, a voltage record table structure with a timestamp is established, and after each sampling is completed, data is overwritten or added to update, so as to generate a voltage measurement record set with a time index; Based on the voltage measurement record set with time index, the newly collected data is judged for duplication, the voltage record set arranged in time order is dynamically maintained, and the real-time voltage data cache is generated.

3. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps of obtaining the voltage transient trigger signal are: Based on the real-time voltage data cache, trace back to a preset fixed time period from the current time point, extract all voltage amplitude data and corresponding timestamp information arranged in sequence within the fixed time period, and form a voltage amplitude sequence and time series within the fixed time window; According to the voltage amplitude sequence and time series within a fixed time window, the fluctuation intensity of the voltage amplitude change rate is calculated; The intensity of the voltage amplitude change rate fluctuation is compared with the rate fluctuation critical value set under stable operating conditions. If the intensity of the voltage amplitude change rate fluctuation is greater than the critical value, it is determined that a transient disturbance has occurred and a voltage transient trigger signal is generated.

4. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps of acquiring the locked transient voltage segment are: Based on the triggering time information carried in the voltage transient trigger signal, the voltage record position corresponding to the triggering time in the instant voltage data cache is retrieved, the voltage record index position corresponding to the time when the voltage transient trigger signal is generated is determined, and the voltage transient start index is generated; According to the voltage transient start index, forward and backward extension searches are respectively performed from the real-time voltage data cache to track the process of the voltage transient amplitude first deviating from the stable state and recovering to the stable state, and the start and end index intervals of the voltage transient occurrence are determined to form a voltage transient occurrence section; Based on the voltage transient occurrence section, all voltage amplitude data and corresponding time data in the real-time voltage data cache within the voltage transient occurrence section are extracted and stored separately, and an independent voltage amplitude and time information sequence is constructed to obtain a locked transient voltage segment.

5. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps of obtaining the core voltage disturbance parameter pair are: Based on the locked transient voltage segment, all voltage amplitude data and corresponding timestamp information in the segment are extracted, and a voltage amplitude sequence and a timestamp sequence with structural matching are constructed to obtain a transient voltage amplitude sequence and a timestamp sequence; According to the transient voltage amplitude sequence and the timestamp sequence, the maximum peak value is calculated and the duration of the voltage amplitude exceeding the normal operating range is counted; Based on the maximum peak value extracted from the transient voltage amplitude sequence and the duration of the voltage amplitude exceeding the normal operating range, a corresponding relationship between the voltage amplitude peak value and the duration is formed to generate a core voltage disturbance parameter pair.

6. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps for obtaining the island event compliance index are as follows: According to the voltage amplitude peak value and the cumulative over-limit duration in the core voltage disturbance parameter pair, the voltage amplitude threshold range and the duration threshold range corresponding to the preset island event characteristic reference interval are retrieved respectively, and the voltage amplitude peak value and the duration are calculated one by one with the upper and lower thresholds of the reference interval to obtain the voltage amplitude difference and the duration difference; Based on the voltage amplitude difference and the duration difference, taking the preset island event characteristic reference interval as the standard, the relative proportion of the two differences in their respective threshold intervals is calculated item by item, and the differences are divided into three levels of high compliance, medium compliance and low compliance according to the proportion values, and the voltage amplitude matching level and duration matching level are obtained; According to the voltage amplitude matching level and the duration matching level, the comprehensive compliance level decision table is called to perform item-by-item matching, the lowest value of the two levels is used as the comprehensive matching level, the comprehensive matching level is defined as the compliance degree of the island event, and the island event compliance index is generated.

7. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps for obtaining the power grid disturbance source classification code are: Based on the island event compliance index, the compliance level is extracted, and the voltage change rate and amplitude jump direction in a fixed interval before and after the triggering moment are combined to form a triple of the compliance level value, the voltage change rate value and the amplitude change sign to generate a disturbance feature input combination; Calculating a disturbance source type identification value according to the disturbance feature input combination; Based on the disturbance source type identification value, classification matching is performed with the disturbance type identification reference interval to generate a power grid disturbance source classification code.

8. The photovoltaic system backflow prevention method according to claim 1, characterized in that: The steps for obtaining the anti-backflow adjustment instruction are as follows: According to the power grid disturbance source classification code, the coding bit sequence in the power grid disturbance source classification code is parsed, and the coding bits are interpreted bit by bit as three identifications of disturbance type, disturbance intensity level and duration level, and the disturbance type identification is judged. If the identification field does not belong to the valid coding range representing the islanding event, the current transient disturbance is judged to be a non-islanding type power grid fault, and a power grid fault type judgment result is generated; According to the grid fault type determination result, combined with the current system operation status, the inverter working mode, grid frequency status and load power fluctuation value are extracted, and the three parameters are used as judgment conditions for conditional combination reasoning, the inverter working mode corresponding to the matching conditions is selected, and it is determined whether the inverter injection current needs to be limited, and the inverter response mode is generated; Based on the inverter response mode, the control actions included in the response mode are parameterized, the values ​​of active power output suppression, reactive power compensation amplitude adjustment and grid-connected current phase shift setting are listed item by item, all control action parameters are combined to generate an anti-backflow regulation instruction.

Images