Wind turbine generator voltage sag compensation method and system based on deep learning

Abstract

The invention discloses a wind turbine generator voltage sag compensation method and system based on deep learning, and relates to the technical field of low voltage ride through control, and the method comprises the steps: collecting the three-phase voltage output and operation state data of a wind turbine generator in real time; judging whether voltage sag occurs or not, and if so, extracting voltage sag element features; judging and generating trigger source information in combination with the unit operation state data and the voltage sag element characteristics; performing compensation analysis according to the voltage sag compensator and outputting a compensation decision; and performing voltage sag compensation control. The technical problems that the traditional voltage sag compensation technology cannot accurately identify the trigger source and element characteristics of the voltage sag of the wind turbine generator set, the compensation decision pertinence is insufficient, the voltage stability control and compensation effect in a circuit system is poor, and the requirement is difficult to meet are solved. The technical effects of accurately identifying the trigger source and the element characteristics of the voltage sag of the wind turbine generator and effectively improving the control and compensation of the voltage stability are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of low voltage ride-through control, and in particular to a method and system for voltage sag compensation of a wind turbine generator set based on deep learning. Background Art

[0002] Voltage sags often occur during wind turbine operation, significantly impacting the stable operation of circuit systems. Existing technologies often use traditional circuit regulation methods to compensate for voltage sags in wind turbines, which have been effective in relatively stable grid environments or simple operating scenarios. However, as the scale of wind power grid integration expands and the grid environment becomes increasingly complex, these traditional compensation technologies have exposed limitations in their application: Due to the inability to accurately identify the triggering sources and element characteristics of voltage sags, compensation decisions are not targeted enough, resulting in poor compensation results and making it difficult to meet the voltage stability control and compensation requirements of power supply / distribution circuit systems. Summary of the Invention

[0003] This application provides a method and system for wind turbine voltage sag compensation based on deep learning, which is used to solve the technical problem that traditional voltage sag compensation technology cannot accurately identify the trigger source and element characteristics of wind turbine voltage sag, and the compensation decision is not targeted enough, resulting in poor control and compensation effects of voltage stability in the circuit system, which is difficult to meet the needs.

[0004] In a first aspect of the present application, a deep learning-based method for compensating for voltage sags in wind turbines is provided, the method comprising: real-time collection of three-phase voltage output data and unit operating status data of the wind turbine; determining whether a voltage sag occurs based on the three-phase voltage output data, and if so, extracting sag elements to generate voltage sag element features; determining fault triggering and fluctuation triggering based on the unit operating status data and the voltage sag element features to generate trigger source information; performing compensation analysis through a voltage sag compensator based on the trigger source information and the voltage sag element features to output a compensation decision; and performing voltage sag compensation control based on the compensation decision.

[0005] The second aspect of the present application provides a wind turbine voltage sag compensation system based on deep learning, the system comprising: a wind turbine data acquisition module, for collecting three-phase voltage output data and unit operating status data of the wind turbine in real time; a voltage sag behavior judgment module, for judging whether voltage sag behavior occurs based on the three-phase voltage output data, and if so, extracting sag elements to generate voltage sag element features; a trigger source information acquisition module, for judging fault triggering and fluctuation triggering based on the unit operating status data and the voltage sag element features, and generating trigger source information; a compensation decision output module, for performing compensation analysis through a voltage sag compensator based on the trigger source information and the voltage sag element features, and outputting a compensation decision; and a voltage sag compensation control execution module, for performing voltage sag compensation control based on the compensation decision.

[0006] One or more technical solutions provided in this application have at least the following technical effects or advantages: This application collects the three-phase voltage output data and unit operating status data of the wind turbine in real time, obtains voltage sag related information through sag element extraction, trigger source judgment and other processing, combines the trigger source information and sag element characteristics to generate a compensation decision through a voltage sag compensator, and then makes adjustments through global collaborative analysis and local optimization, so as to accurately compensate for the voltage sag of the wind turbine, make the control and compensation effect of voltage stability in the power supply / distribution circuit system better, meet the stable operation requirements of the wind turbine, and achieve the precise identification of the trigger source and element characteristics of the wind turbine voltage sag, effectively improving the technical effect of voltage stability control and compensation. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0008] Figure 1 This is a flow chart of a method for compensating for voltage sag of a wind turbine generator system based on deep learning provided in an embodiment of the present application.

[0009] Figure 2 This is a structural diagram of a wind turbine voltage sag compensation system based on deep learning provided in an embodiment of the present application.

[0010] Description of the accompanying symbols: wind turbine data acquisition module 1, voltage sag behavior judgment module 2, trigger source information acquisition module 3, compensation decision output module 4, voltage sag compensation control execution module 5. DETAILED DESCRIPTION

[0011] This application provides a method and system for wind turbine voltage sag compensation based on deep learning, which is used to solve the technical problem that traditional voltage sag compensation technology cannot accurately identify the trigger source and element characteristics of wind turbine voltage sag, and the compensation decision is not targeted enough, resulting in poor control and compensation effects of voltage stability in the circuit system, which is difficult to meet the needs.

[0012] The following will be combined with the accompanying drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only some of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0013] It should be noted that the terms "first", "second", etc. in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices.

[0014] Example 1, as Figure 1 As shown, a method for compensating for voltage sag of a wind turbine generator system based on deep learning is provided, wherein the method comprises: Step A100: collecting three-phase voltage output data and unit operating status data of the wind turbine in real time.

[0015] In the embodiment of the present application, the three-phase voltage is the voltage of the three-phase AC circuit used by the wind turbine generator set during power generation and transmission.

[0016] Specifically, real-time collection of three-phase voltage output data from wind turbines is primarily achieved by deploying high-precision voltage sensors at the turbine output and grid-connected interface. These sensors must cover the A, B, and C phases, capturing both the instantaneous voltage fluctuation data and the RMS value (root mean square value) of each phase voltage in real time, while also recording the precise timestamp of each acquisition.

[0017] The collection of unit operating status data involves a variety of specialized sensors and data interfaces. Wind speed data is collected in real time by an anemometer mounted on the nacelle roof, recording both instantaneous and average wind speeds every second to aid in determining whether wind speed fluctuations have triggered voltage sags. Blade and generator speeds are acquired via speed sensors deployed on the blade hub and generator shaft, recording minute-by-minute speed changes in real time to reflect the unit's mechanical operating status. Reactive and active power data is collected by a power sensor at the generator output, updated every 0.5 seconds, to assess the unit's power output stability.

[0018] The operating temperature of equipment such as generators, transformers, and inverters is collected through temperature sensors mounted on the surface of the equipment to monitor in real time whether the voltage output of the equipment is affected by overheating; the internal voltage, current, and frequency data of the wind turbine are synchronously collected by the current sensor and frequency meter in the unit control cabinet, and a set of data is generated every 0.1 second to reflect the internal circuit status of the unit; the start / stop status of the unit is recorded by the status monitoring module in the control cabinet, and is updated in real time with the switch signal starting as 1 and stopping as 0; the connected power grid operation data is obtained through the dedicated data acquisition module of the grid-connected interface, including grid voltage, frequency, and fault signals, etc., which are synchronized every 0.1 second to determine whether the voltage sag is triggered by a grid fault.

[0019] All unit operating status data is accompanied by a timestamp consistent with the three-phase voltage output data, ensuring the timing matching of different types of data and providing a complete status background for subsequent trigger source judgment.

[0020] Step A200: determining whether a voltage sag occurs based on the three-phase voltage output data; if so, extracting sag elements to generate voltage sag element features.

[0021] In the embodiment of the present application, voltage sag is a phenomenon in which the output voltage of the wind turbine generator set temporarily decreases due to grid failure, short circuit, load fluctuation, etc.

[0022] Optionally, based on the previously collected three-phase voltage output data, core thresholds for voltage sag detection are first set: 80% of the rated voltage is used as the voltage amplitude threshold, and 10 milliseconds is used as the duration threshold. The data processing module continuously monitors the real-time three-phase voltage RMS values ​​and compares each phase voltage with the amplitude threshold moment by moment.

[0023] Next, when the RMS voltage of any phase falls below the voltage amplitude threshold for the first time, a duration timer is started and the timestamp of that moment is recorded. If the voltage of that phase remains below the voltage amplitude threshold for the next 10 milliseconds and does not recover above the threshold, a voltage sag is initially determined to be a possible event. If the voltage rises above the threshold during this period, it is considered a transient fluctuation and the timer is terminated.

[0024] To avoid misjudgment of a single phase, the system simultaneously checks the voltage status of the other two phases. If only one phase voltage is abnormal and its duration does not reach the threshold, it may be due to sensor failure or local interference. If at least one phase voltage meets the amplitude below the threshold and lasts for more than 10 milliseconds, a voltage sag is confirmed, triggering the subsequent voltage sag feature extraction process.

[0025] Finally, the voltage drop amplitude is calculated based on the three-phase voltage output data, the phase is identified through Fourier transform, and the duration is extracted. These three are then used to construct the voltage sag element feature, which is described in detail in steps A210-A240.

[0026] By setting dual thresholds of voltage amplitude and duration, and combining high-frequency collected three-phase voltage data for continuous monitoring and multi-dimensional verification, it is possible to accurately determine whether a wind turbine has experienced a voltage sag, providing a reliable basis for subsequent voltage sag factor analysis and compensation control.

[0027] Step A300: Based on the unit operating status data and the voltage sag element characteristics, fault triggering and fluctuation triggering judgment are performed to generate trigger source information.

[0028] In one embodiment of the present application, by constructing a voltage sag judgment sample library, the samples, unit operating status data, and voltage sag element characteristics are input into the twin network for similarity identification, and the corresponding historical trigger factors are obtained based on the similarity of meeting the standards, and the trigger source information is generated. The specific steps are described in detail in A310-A330.

[0029] Among them, the fault trigger is the voltage sag caused by internal wind turbine or grid fault, and the fluctuation trigger is the voltage sag caused by wind speed fluctuation, which is explained in detail in A340.

[0030] Step A400: performing compensation analysis through a voltage sag compensator according to the trigger source information and the voltage sag factor characteristics, and outputting a compensation decision.

[0031] In the embodiment of the present application, the voltage sag compensator is a device constructed by building an integrated twin unit model including the wind turbine itself, the connected power grid and meteorological conditions, loading a historical voltage sag recovery record set to learn the relationship between voltage compensation parameters and voltage sag elements and trigger sources.

[0032] Specifically, by constructing a twin unit model including the unit, power grid, and weather, the voltage sag compensator is trained with historical records, the trigger source and sag element characteristics are input and verified by the model, and the compensation decision is output. The specific steps are detailed in A410-A440.

[0033] Step A500: performing voltage sag compensation control according to the compensation decision.

[0034] Specifically, when performing voltage sag compensation control based on the compensation decision, the generated compensation decision parameters are first converted into specific control instructions. These parameters include the reactive compensation amount, converter modulation ratio, and compensation execution time. These parameters are transmitted as digital signals via the wind turbine control system bus to the corresponding execution devices, such as the converter and static VAR generator (SVG). For example, upon receiving a command for a modulation ratio of 0.75, the converter adjusts the switching frequency and conduction angle of its internal power modules, changing the amplitude and phase of the output voltage. The SVG, in response to the 120 kvar command, controls the switching state of its bridge arms to inject the specified reactive power into the grid to support voltage recovery.

[0035] During execution, the system collects real-time output voltage data from the wind turbines, updating it every 10ms to monitor the effectiveness of compensation. If the compensation decision is for a voltage sag caused by a grid fault, the system prioritizes voltage recovery speed, requiring a voltage recovery from the sag within 200ms. If the sag is triggered by wind speed fluctuations, the post-recovery voltage fluctuation amplitude must be controlled to ensure it remains within ±3% of the rated voltage. If the voltage does not recover as expected, feedback is immediately provided to the control system, temporarily increasing the compensation amount, such as by adding additional reactive power, until the voltage reaches the target.

[0036] At the same time, the control process needs to coordinate with other wind turbine protection mechanisms. For example, if the generator speed is detected to be outside the safe range during compensation, such as exceeding 1800 rpm, the compensation intensity will be temporarily reduced, and the pitch system will prioritize adjusting the blade angle to stabilize the speed. Once the speed returns to below 1500 rpm, the original compensation decision will be resumed to avoid equipment overload caused by a single compensation target.

[0037] By converting compensation decisions into specific control instructions and driving the execution equipment, combined with real-time monitoring and coordinated protection mechanisms, it is ensured that voltage sags are effectively alleviated, and the output voltage of the wind turbine is quickly restored to a stable range, ensuring the safe operation of the unit and the power supply quality of the power grid.

[0038] Furthermore, step A200 in the method provided in the embodiment of the present application includes: A210: Calculate the voltage drop amplitude based on the three-phase voltage output data to generate voltage sag amplitude information.

[0039] A220: Perform Fourier transform on the three-phase voltage output data, identify phase characteristics based on the transformation results, and generate voltage sag phase information.

[0040] A230: Extract the duration based on the timestamp in the three-phase voltage output data to generate sag duration information.

[0041] A240: Construct the voltage sag element feature based on the voltage sag amplitude information, the voltage sag phase information, and the voltage sag duration information.

[0042] In the embodiment of the present application, the sag duration information is the duration of the voltage sag, rather than the time from the voltage sag to the voltage recovery.

[0043] Specifically, when extracting sag elements, the voltage drop amplitude is first calculated based on the three-phase voltage output data. For a wind turbine with a rated voltage of 690V, the RMS values ​​of the three phases are obtained in real time: phase A is 510V, phase B is 505V, and phase C is 515V. Using the formula (rated voltage minus current voltage) / rated voltage, the voltage drop amplitudes for phase A are approximately 26.1%, phase B is approximately 26.8%, and phase C is approximately 25.4%. This information is then integrated to generate voltage sag amplitude information, reflecting the extent of the voltage drop for each phase.

[0044] Subsequently, a Fourier transform is performed on the three-phase voltage output data. This first obtains the time-varying voltage signals of each phase, such as the instantaneous voltage value sequence of phase A from time t0 to t1. These time-domain signals are then converted into frequency-domain signals using a Fourier transform algorithm, thereby decomposing the phase components of the fundamental wave (50Hz or 60Hz) and the second and third harmonics in each phase voltage. The fundamental wave phase is then extracted, as it is the primary component of the voltage signal, and its phase variation is crucial for sag analysis. The fundamental wave phase is then compared with the rated fundamental wave phase during normal operation of the wind turbine (e.g., the rated phase of phase A is 0°) to obtain the phase offset value for each phase. For example, the fundamental wave phase of phase A lags by 5°, the fundamental wave phase of phase B leads by 2°, and the phase of phase C has virtually no offset. Finally, these phase offset data are integrated to generate voltage sag phase information. This clearly reflects the phase differences between phases A, B, and C during the voltage sag, providing a key basis for determining the phase dimension of the sag characteristic.

[0045] At the same time, the duration is extracted based on the timestamps in the three-phase voltage output data. The time difference between the time t1, when the voltage sag is determined to have occurred, and the time t2, when the current data is collected, is calculated to obtain the duration of the sag. This information is then generated, recording the duration of the sag from its onset to the current state, not the time it took to recover from the sag.

[0046] Finally, the voltage sag amplitude information, voltage sag phase information, and sag duration information obtained above are integrated, as shown in Table 1, to construct a complete voltage sag element feature, which comprehensively describes the key features of the voltage sag.

[0047] By extracting and integrating the amplitude, phase and duration information of the voltage sag in steps, accurate voltage sag element characteristics are constructed, providing a detailed and critical basis for subsequent trigger source judgment and compensation decisions based on these characteristics.

[0048] Table 1: Voltage sag factor characteristic parameters Furthermore, step A300 in the method provided in the embodiment of the present application includes: A310: Build a voltage sag judgment sample library.

[0049] A320: Traverse each voltage sag judgment sample in the voltage sag judgment sample library, input them together with the unit operation status data and the voltage sag element features into the twin network for similarity identification, and obtain similarity output information.

[0050] A330: If the similarity output information meets the preset threshold, collect historical voltage sag trigger factors corresponding to the voltage sag judgment sample, complete fault trigger and fluctuation trigger judgment, and generate the trigger source information.

[0051] In this embodiment of the present application, the voltage sag judgment sample library is a collection of samples containing historical voltage sag events. Each sample is associated with unit operating status data, voltage sag element characteristics, and the corresponding historical voltage sag trigger factor. This sample is used to perform a similarity comparison with current data to determine the trigger source. A twin network is a network used for similarity identification. By receiving samples from the voltage sag judgment sample library, unit operating status data, and voltage sag element characteristics, it performs similarity calculations and outputs similarity information to assist in fault trigger and fluctuation trigger determination.

[0052] Optionally, when constructing a voltage sag judgment sample library, all types of voltage sag event data that occurred in the historical operation of the wind turbine are fully collected, covering the complete three-phase voltage output data of each event, including the instantaneous value, effective value, and timestamp of each phase voltage; the corresponding unit operating status data, such as wind speed, blade speed, generator power, equipment temperature, etc.; and clear triggering factors manually marked or recorded by the system, such as grid short circuit, unit inverter failure, sudden increase in wind speed, etc.

[0053] The above raw data are preprocessed, including eliminating outliers, such as extreme voltage values ​​falsely reported by sensors; standardization, converting data such as voltage amplitude and speed of different magnitudes into the same interval; time alignment, ensuring that the timestamps of three-phase voltage data and operating status data accurately match, and then classified and stored according to the trigger factor type of fault trigger or fluctuation trigger, forming a voltage sag judgment sample library containing thousands to tens of thousands of valid samples, each sample is associated with complete feature data and trigger factor labels.

[0054] After the sample library is built, each voltage sag detection sample in the library is traversed. The sample's characteristic data, namely historical three-phase voltage characteristics and historical unit operating status data, is input into the pre-trained twin network along with the currently detected unit operating status data and voltage sag element characteristics. The twin network uses two identical sub-networks to extract the historical sample features and current data features, respectively, to generate high-dimensional feature vectors. The cosine similarity between these vectors is then calculated to obtain similarity output information. For example, the similarity between one sample and the current data is 0.85, while that of another sample is 0.62.

[0055] Furthermore, the training of the twin network needs to first filter data from the voltage sag judgment sample library, construct positive sample pairs composed of different historical events of the same trigger source (such as voltage sag events caused by two power grid failures) and negative sample pairs composed of different trigger source events (such as a voltage sag event caused by an internal fault and a wind speed fluctuation); then perform standardization, time alignment and other preprocessing on the three-phase voltage characteristics and unit operating status data in the sample to unify the data format and dimension; then build two sub-networks with the same structure, which are used to extract the features of historical samples and current data respectively, and through training, let the sub-networks learn the key features that can distinguish different trigger sources; use a suitable loss function in training to make the feature vectors of the same type of samples closer and the feature vectors of different types of samples farther apart. After multiple rounds of iterative optimization of network parameters, the accuracy of the model's similarity judgment on the verification set is stabilized at a high level. At this time, the twin network training is completed.

[0056] Next, a preset threshold (such as 0.8) is set. When the similarity output information for a sample reaches or exceeds this threshold, the current voltage sag event is determined to be highly similar to the historical sample. The corresponding historical voltage sag trigger factors are then collected. If the trigger factor for the sample with the highest similarity is a grid fault, the current event is determined to be a fault trigger; if it is a wind speed fluctuation, it is determined to be a fluctuation trigger. Ultimately, clear trigger source information is generated, such as "Trigger Source: Fault Trigger (Grid Short Circuit)."

[0057] By building a sample library containing rich historical data and combining it with the twin network for similarity matching, the trigger source type of voltage sag can be accurately identified, providing a key basis for subsequent targeted compensation decisions and improving the accuracy and reliability of trigger judgment.

[0058] Furthermore, step A300 in the method provided in the embodiment of the present application includes: A340: Fault triggering refers to a voltage sag caused by an internal fault in the wind turbine or a grid fault, and fluctuation triggering refers to a voltage sag caused by wind speed fluctuations.

[0059] Specifically, during the operation of a wind turbine, when a voltage sag is detected, its trigger type needs to be further distinguished, i.e., fault trigger or fluctuation trigger.

[0060] To determine whether a fault has been triggered, the internal equipment status data of the wind turbine must be combined with the grid operation data: if the generator winding temperature is detected to rise sharply from the normal operating temperature of 70°C to 105°C within 10 seconds, and at the same time the inverter output current is distorted, the total harmonic distortion rate exceeds 5%, and the three-phase voltage imbalance exceeds 2%, then the voltage sag may be caused by an internal fault in the wind turbine (such as an inverter fault); if the voltage at the common connection point connected to the grid is detected to drop sharply from the rated voltage within 50 milliseconds, for example, from 690V to 480V, and other wind turbines in the area experience voltage sags at the same time, and the grid frequency fluctuation range exceeds ±0.5Hz, then it can be preliminarily determined that the voltage sag was caused by a grid fault (such as a line short circuit).

[0061] The judgment of fluctuation triggering mainly relies on the correlation analysis between wind speed and unit operating status: for example, the anemometer collects that the wind speed rises sharply from 12m / s to 20m / s within 3 seconds, exceeding the normal fluctuation range of ±3m / s. At the same time, the blade speed increases from 18rpm to 25rpm, and the generator active power suddenly increases from 800kW to 1200kW. The internal equipment temperature, current and grid voltage of the wind turbine are all within the normal range. At this time, it can be preliminarily determined that the voltage sag is caused by wind speed fluctuation.

[0062] Furthermore, during the actual judgment process, the system traverses the voltage sag judgment sample library, comparing the real-time collected internal fault characteristic data such as temperature and current distortion rate, grid fault characteristic data such as common connection point voltage and frequency, and wind speed fluctuation characteristic data such as wind speed change rate and rotational speed correlation with the corresponding characteristic data of all historical samples in the library, and calculating the similarity between each sample and the current data. A preset threshold is set, such as 0.8. When there are samples with a similarity exceeding this threshold, the sample with the highest similarity is selected. If the historical trigger factor corresponding to the sample is an internal fault of the wind turbine or a grid fault, the current event is determined to be a fault trigger; if it is a wind speed fluctuation, it is determined to be a fluctuation trigger, thereby clarifying the specific trigger source of the voltage sag.

[0063] By clarifying the judgment basis and data characteristics of fault triggering and fluctuation triggering, the different causes of voltage sag are accurately distinguished, providing a clear basis for the subsequent formulation of targeted compensation strategies based on the trigger source, ensuring the effectiveness and accuracy of the compensation measures.

[0064] Furthermore, step A400 in the method provided in the embodiment of the present application includes: A410: Perform digital simulation modeling on the wind turbine to construct a twin turbine model, wherein the modeling includes integrated modeling of the turbine itself, the connected power grid, and meteorological conditions.

[0065] A420: Perform trigger source statistics based on historical voltage sag record information, generate multiple preset trigger sources, and collect historical voltage sag recovery record sets.

[0066] A430: Load the historical voltage sag recovery record set into the twin unit model, learn the relationship between voltage compensation parameters and voltage sag elements and trigger sources, and construct the voltage sag compensator, wherein the voltage sag compensator is connected to the twin unit model.

[0067] A440: Input the trigger source information and the voltage sag element characteristics into the voltage sag compensator, perform compensation parameter analysis, and then call the twin unit model to perform compensation verification. After the verification is passed, output the compensation decision.

[0068] In the embodiment of the present application, the twin unit model is a model constructed by digital simulation modeling of the wind turbine, wherein the modeling covers the integrated modeling of the unit itself, the connected power grid and the meteorological conditions.

[0069] Specifically, first, digital simulation modeling of the wind turbine is performed to construct a twin turbine model, which covers the integrated data integration and model building of the turbine itself, the connected power grid, and meteorological conditions. The modeling of the wind turbine itself requires technical personnel to enter core parameters, including hardware parameters such as rated power, blade length, generator rated voltage, converter capacity, as well as operating data such as the correspondence between blade pitch angle and wind speed, and the characteristic curve between generator speed and output power. The modeling of the connected power grid requires the inclusion of parameters such as the line impedance from the public connection point to the substation, the transformer capacity and ratio, and the short-circuit capacity of the power grid to simulate the voltage transfer characteristics during power grid faults. The meteorological condition modeling requires the mapping relationship between environmental factors such as wind speed, air density, and temperature and the turbine output, ultimately forming a digital twin that can reproduce the voltage output characteristics of the turbine under different operating conditions, namely the twin turbine model.

[0070] Next, when compiling trigger source statistics based on historical voltage sag records, it is necessary to sort out all voltage sag events from the past five years and categorize them by triggering factors. For example, grid faults such as line short circuits and transformer failures account for 60%, internal wind turbine faults such as inverter failures and bearing overheating account for 25%, and wind speed fluctuations with sudden changes of ±5m / s or more account for 15%. Based on these, multiple preset trigger sources are generated, such as grid short circuits, inverter failures, and wind speed surges. At the same time, corresponding historical voltage sag recovery records are collected. Each record includes the trigger source type, voltage sag elements such as amplitude drop, phase lag, and duration; compensation parameters used such as reactive power compensation and converter modulation ratio; and recovery effects such as voltage recovery time and post-recovery fluctuations, resulting in thousands of valid records.

[0071] Then, when the historical voltage sag recovery record set is loaded into the twin unit model to build a voltage sag compensator, the unstructured data in the record set needs to be converted into a feature vector that can be recognized by the model through data preprocessing: the trigger source type is converted to a one-hot encoding. For example, there are 8 trigger sources, and the length of the corresponding one-hot encoding vector is 8. The grid short circuit corresponds to: [1,0,0,0,0,0,0,0], the transformer fault corresponds to: [0,1,0,0,0,0,0,0], the inverter fault corresponds to: [0,0,1,0,0,0,0,0] ,0], bearing overheating corresponds to: [0,0,0,1,0,0,0,0], wind speed surge corresponds to: [0,0,0,0,1,0,0,0], wind speed drop corresponds to: [0,0,0,0,0,1,0,0], generator failure corresponds to: [0,0,0,0,0,0,1,0], and line overload corresponds to: [0,0,0,0,0,0,0,1]. Voltage sag factors are normalized to values ​​in the range [-1,1], for example, a 30% drop in amplitude corresponds to -0.3. The compensation parameters are used as labels and retain their original values. The compensator model is then trained using the random forest algorithm, with the trigger source and voltage sag factors as input features and the compensation parameters as output targets. Model parameters are optimized through 5-fold cross-validation, such as using 200 decision trees and a maximum depth of 10, to keep the model's prediction error within 5% on the validation set. At the same time, a real-time communication interface between the compensator and the twin unit model is established to ensure that the parameters output by the compensator can directly drive the twin model for simulation verification.

[0072] Finally, after inputting the current trigger source information and voltage sag factor characteristics into the voltage sag compensator, the compensator quickly outputs initial compensation parameters, such as reactive power compensation and converter modulation ratio, using a built-in training model. The twin unit model is then invoked and these parameters are substituted into the simulation environment to simulate the voltage recovery process. Assuming the model output indicates that the voltage recovers to 92% of the rated value within 15ms after compensation, with a fluctuation range of ±3%, meeting the preset verification criteria of recovery time ≤20ms and fluctuation ≤±5%, the compensator officially outputs the compensation decision.

[0073] By constructing a multi-dimensional twin unit model, training the compensator based on historical data and combining simulation verification, we have achieved accurate mapping from trigger sources and sag factors to compensation decisions, ensuring that the output compensation decisions can effectively respond to different types of voltage sags, and improving the reliability and pertinence of wind turbine voltage stability control.

[0074] Furthermore, step A430 in the method provided in the embodiment of the present application further includes: A431: If the trigger source is a fault-type trigger source, the historical voltage sag recovery record set is loaded into the twin unit model, the relationship between the fault isolation characteristics and the fault source location is learned, and the voltage sag compensator is optimized.

[0075] Specifically, when the trigger source is determined to be a fault-related trigger source, such as a wind turbine internal inverter fault or a grid line short circuit, the system first screens the historical voltage sag recovery records for records related to the fault type. These records contain detailed information about various past fault events, including the specific location of the fault source (e.g., generator side, transformer end, or a specific section of the grid line); fault isolation characteristics at the time of the fault, such as the amplitude and frequency of the current surge, the voltage waveform distortion pattern, and the degree of voltage imbalance between phases; and the compensation parameters and operating procedures for successful voltage restoration in the corresponding event.

[0076] Subsequently, the selected historical fault records are loaded into the twin unit model. The twin unit model simulates the evolution of fault isolation characteristics at different fault source locations by reproducing historical fault scenarios. For example, when the fault source is the inverter, the model will show a surge in current harmonics in a specific frequency band, and the voltage distortion of phase A is particularly obvious. When the fault source is a short circuit in a section of the power grid, the model will show a simultaneous sudden drop in the three-phase voltage, and the magnitude of the temporary drop decreases with increasing distance from the short circuit point. Through simulation learning of a large number of historical scenarios, the model gradually establishes a mapping relationship between fault isolation characteristics and fault source locations, clarifying which feature combinations correspond to inverter faults and which correspond to power grid line faults, as well as the subtle differences in features under faults at different locations.

[0077] Next, the voltage sag compensator is optimized based on the mapping relationship learned by the twin unit model. During the optimization process, the compensator updates its internal decision-making logic: when a new fault-related voltage sag is detected, the compensator can quickly locate the fault source based on the current fault isolation characteristics, such as real-time current harmonic data and voltage distortion, combined with the learned mapping relationship. For faults in different locations, the compensator adjusts the emphasis of the compensation parameters. For example, if it is determined to be an inverter fault, the response speed and accuracy of the reactive compensation will be optimized first to suppress the impact of harmonics. If it is determined to be a grid line fault, the duration of the voltage support will be strengthened, and the grid protection mechanism will be used to quickly restore the voltage.

[0078] By using historical fault records to train the twin unit model, it can learn the relationship between fault isolation characteristics and fault source location, and then optimize the voltage sag compensator. This enables the compensator to more accurately match the fault type and compensation strategy when facing a fault-type trigger source, thereby improving the pertinence and effectiveness of voltage sag compensation in fault conditions.

[0079] Furthermore, the embodiment of the present application further includes step A600, which includes: A610: Collect information about several other wind turbines connected to the wind turbine.

[0080] A620: Obtain multiple voltage compensation decisions of the multiple other wind turbines through the data sharing channel.

[0081] A630: Performing global synergistic impact analysis and local compensation optimization on the compensation decision and the plurality of voltage compensation decisions to generate a local optimized compensation decision.

[0082] A640: Send the local optimization compensation decision to the corresponding wind turbine for optimization control.

[0083] In one embodiment, after outputting a compensation decision, the first step is to collect data from several other wind turbines connected to the current wind turbine. Based on the wind farm's grid topology, it is necessary to identify associated wind turbines that share the same collector line or common connection point as the current wind turbine. The voltage outputs of these wind turbines are correlated, and the compensation behavior of one wind turbine may affect the voltage status of other wind turbines. Including these wind turbines in the analysis can help avoid overall grid fluctuations caused by compensation from a single wind turbine.

[0084] Next, voltage compensation decisions for these other wind turbines are obtained through a data sharing channel. This data sharing channel must have real-time transmission capabilities to ensure timely collection of compensation plans for each associated unit, including compensation methods (such as reactive power regulation and converter control), planned compensation measures, and implementation timing, providing a comprehensive decision-making basis for subsequent global analysis.

[0085] Afterwards, the compensation decision is globally integrated with the compensation decisions of other units to generate global information. After judging the over-compensation, the unit with the largest compensation amount is adjusted to generate a local optimized compensation decision. The specific steps are described in detail in A631-A632.

[0086] Finally, the locally optimized compensation decisions are distributed to the corresponding wind turbines, ensuring that each unit executes the optimized plan in a coordinated manner. Time synchronization is maintained during execution to avoid voltage instability caused by time differences. This ensures that the compensation actions of all units work together to effectively stabilize the grid voltage.

[0087] By collecting decisions of related units, globally analyzing synergistic effects, optimizing compensation plans and jointly executing them, coordinated cooperation of voltage sag compensation for multiple units in the wind farm is achieved, avoiding the limitations of a single decision and improving the stability and reliability of the overall compensation effect.

[0088] Furthermore, step A630 in the method provided in the embodiment of the present application includes: A631: Globally fuse the compensation effects of the compensation decision and the multiple voltage compensation decisions to generate global fusion compensation effect information.

[0089] A632: Determine whether there is over-compensation based on the global fusion compensation effect information. If so, extract the local unit with the largest local compensation amount, adjust the compensation decision according to the over-compensation parameter, and generate the local optimized compensation decision.

[0090] Optionally, when globally integrating the compensation effect of the compensation decision with the voltage compensation decisions of several other wind turbines, the core contents of all decisions are first summarized, including the compensation methods, compensation intensity and execution rhythm planned by each unit, such as reactive power injection and converter output adjustment.

[0091] Subsequently, a voltage coordination analysis model at the wind farm level is used to simulate the combined impact of these decisions on the overall grid voltage. For example, the voltage recovery curve and fluctuation range of the common connection point after the superposition of the compensation behaviors of each unit are analyzed. The results of these combined impacts are integrated into global fusion compensation effect information to fully reflect the overall effect of multi-unit coordinated compensation.

[0092] Among them, the construction of the voltage coordination analysis model needs to be based on the wind farm as a whole. First, the core parameters of each wind turbine, such as rated voltage and compensation capacity, are integrated; the topological structure of the connected power grid, such as line connection relationship and impedance parameters; and the influence of meteorological conditions are connected to form a basic data layer; then a digital simulation framework is built based on these data to simulate the impact of single-unit compensation behavior on local voltage and the coupling transmission process of multi-unit compensation behavior through power grid lines; then historical multi-unit coordinated compensation case data is introduced to learn the correlation pattern between different compensation combinations and voltage changes at the common connection point (such as recovery curves and fluctuation ranges), and optimize model parameters; finally, through actual operation data verification, it is ensured that the model can accurately reproduce the global voltage effect during multi-unit compensation.

[0093] Overcompensation is then determined based on the globally integrated compensation effect information, based on grid voltage stability criteria, such as the permissible fluctuation range of the rated voltage. If analysis reveals that the combined compensation effect could cause the voltage to exceed this range, overcompensation is considered. At this point, the local units with the largest planned compensation amounts are identified from all participating generators, as these units' compensation behavior typically has the most significant impact on global voltage fluctuations and are key targets for regulation.

[0094] Finally, when adjusting compensation decisions based on overcompensation parameters, the degree of overcompensation, such as the predicted magnitude of voltage exceeding rated values, is considered. The compensation decision for that local unit is then adjusted, perhaps by reducing the compensation intensity or slowing down the compensation process. After these adjustments, global fusion analysis is performed to verify that the overall compensation effect is within a reasonable range. Ultimately, a locally optimized compensation decision is generated that balances global stability with the characteristics of the local unit.

[0095] By globally integrating the compensation effects of multiple units to identify over-compensation risks and targetedly adjusting the units with the greatest impact, the voltage overshoot problem in collaborative compensation is effectively avoided, ensuring the stability and accuracy of the overall voltage recovery of the wind farm.

[0096] Furthermore, step A632 in the method provided in the embodiment of the present application includes: A632-1: Construct a local-global fusion relationship by analyzing the relationship between historical local compensation data and global fusion compensation data.

[0097] A632-2: Based on the local-global fusion relationship, adjust the compensation decision according to the overcompensation parameter.

[0098] In one embodiment, when making compensation decisions based on overcompensation parameters, historical local compensation data and global integrated compensation data are first collected. The historical local compensation data includes compensation measures implemented by each wind turbine in the past, such as reactive power regulation and converter parameter adjustments, as well as the corresponding voltage changes within the turbine itself. The global integrated compensation data includes the overall voltage status of the entire wind farm or associated power grid after the combined effect of these local compensation measures, such as voltage fluctuations and recovery speed at the point of common connection.

[0099] By analyzing the above historical data, we can find the correlation pattern between local compensation behavior and global voltage changes. For example, how the increase in the compensation intensity of a certain unit will affect the increase in global voltage, and then establish a corresponding relationship between local compensation and global effects, that is, the local-global fusion relationship.

[0100] Based on the established local-global fusion relationship, compensation decisions are adjusted in conjunction with overcompensation parameters. The overcompensation parameter reflects the extent to which the current global voltage may exceed the stable range. Based on the local-global fusion relationship, it is determined which local units' compensation behavior has the greatest impact on global overcompensation, and the amount of local compensation adjustment required to bring the global voltage back to the stable range. For example, if the relationship shows a significant positive correlation between a unit's compensation and the global voltage increase, and the current overcompensation parameter indicates significant voltage overshoot, the compensation intensity of that unit is specifically reduced until global analysis verifies that the overall voltage is within the stable range, completing the compensation decision adjustment.

[0101] By analyzing historical data to establish local and global correlations, and accurately adjusting local compensation decisions based on overcompensation parameters, effective control is achieved under overcompensation conditions, ensuring the stability and safety of the overall voltage of the wind farm.

[0102] By globally integrating the compensation effects of each unit to identify the risk of overcompensation and targetedly adjusting the unit with the largest compensation amount, the voltage overshoot problem during collaborative compensation of multiple units is effectively avoided, ensuring the stability and accuracy of the overall voltage recovery of the wind farm.

[0103] In summary, the wind turbine voltage sag compensation method based on deep learning provided by the embodiments of the present application has the following technical effects: This application collects the three-phase voltage output data and unit operating status data of the wind turbine in real time, obtains the trigger source information and voltage sag element characteristics through sag element extraction, twin network similarity identification and other processing, combines this information to generate a compensation decision through a voltage sag compensator, and then adjusts it through global collaborative impact analysis and local compensation optimization, so as to accurately compensate for the voltage sag of the wind turbine, make the voltage stability control and compensation effect in the power supply / distribution circuit system better, meet the stable operation requirements of the wind turbine, and achieve accurate identification of the trigger source and element characteristics of the wind turbine voltage sag, effectively improving the technical effect of voltage stability control and compensation.

[0104] Example 2, as Figure 2 As shown, based on the same inventive concept as the aforementioned embodiment 1, the embodiment of the present application provides a wind turbine voltage sag compensation system based on deep learning, the system comprising: The wind turbine data acquisition module 1 is used to collect the three-phase voltage output data and the wind turbine operating status data of the wind turbine in real time.

[0105] The voltage sag behavior judgment module 2 judges whether a voltage sag behavior occurs based on the three-phase voltage output data, and if so, extracts sag elements to generate voltage sag element features.

[0106] The trigger source information acquisition module 3 performs fault triggering and fluctuation triggering judgment based on the unit operation status data and the voltage sag element characteristics to generate trigger source information.

[0107] The compensation decision output module 4 is used to perform compensation analysis through a voltage sag compensator according to the trigger source information and the voltage sag element characteristics, and output a compensation decision.

[0108] The voltage sag compensation control execution module 5 is configured to perform voltage sag compensation control according to the compensation decision.

[0109] Furthermore, the voltage sag behavior judgment module 2 is configured to perform the following steps: The method further comprises the steps of: calculating a voltage drop amplitude based on the three-phase voltage output data to generate voltage sag amplitude information; performing Fourier transform on the three-phase voltage output data, identifying phase characteristics based on the transform result, and generating voltage sag phase information; extracting an elapsed time based on a timestamp in the three-phase voltage output data to generate sag duration information; and constructing a voltage sag element feature based on the voltage sag amplitude information, the voltage sag phase information, and the sag duration information.

[0110] Furthermore, the trigger source information acquisition module 3 is configured to perform the following steps: Construct a voltage sag judgment sample library; traverse each voltage sag judgment sample in the voltage sag judgment sample library, and input it together with the unit operation status data and the voltage sag element characteristics into the twin network for similarity identification to obtain similarity output information; if the similarity output information meets the preset threshold, collect the historical voltage sag trigger factors corresponding to the voltage sag judgment sample, complete the fault trigger and fluctuation trigger judgment, and generate the trigger source information.

[0111] Fault triggering refers to a voltage sag caused by an internal fault in the wind turbine or a grid fault, and fluctuation triggering refers to a voltage sag caused by wind speed fluctuations.

[0112] Furthermore, the compensation decision output module 4 is configured to perform the following steps: Digital simulation modeling is performed on the wind turbine to construct a twin turbine model, wherein the modeling includes integrated modeling of the turbine itself, the connected power grid, and meteorological conditions; trigger source statistics are performed based on historical voltage sag record information to generate multiple preset trigger sources, and a historical voltage sag recovery record set is collected; the historical voltage sag recovery record set is loaded into the twin turbine model, and the relationship between voltage compensation parameters and voltage sag elements and trigger sources is learned to construct the voltage sag compensator, wherein the voltage sag compensator is connected to the twin turbine model; the trigger source information and the voltage sag element characteristics are input into the voltage sag compensator, and after performing compensation parameter analysis, the twin turbine model is called to perform compensation verification, and the compensation decision is output after the verification is passed.

[0113] If the trigger source is a fault-type trigger source, the historical voltage sag recovery record set is loaded into the twin unit model, the relationship between the fault isolation characteristics and the fault source location is learned, and the voltage sag compensator is optimized.

[0114] Furthermore, the system further includes a local optimization compensation decision control module, which is configured to perform the following steps: Collect data from several other wind turbines connected to the wind turbine; obtain several voltage compensation decisions of the several other wind turbines through a data sharing channel; perform global synergistic impact analysis and local compensation optimization on the compensation decisions and the several voltage compensation decisions to generate a local optimized compensation decision; and send the local optimized compensation decision to the corresponding wind turbine for optimized control.

[0115] The compensation effect of the compensation decision and the several voltage compensation decisions are globally integrated to generate global integrated compensation effect information; based on the global integrated compensation effect information, it is determined whether over-compensation occurs; if so, the local unit with the largest local compensation amount is extracted, and the compensation decision is adjusted according to the over-compensation parameter to generate the local optimized compensation decision.

[0116] By analyzing the relationship between historical local compensation data and global fusion compensation data, a local-global fusion relationship is constructed; and based on the local-global fusion relationship, compensation decision adjustment is performed according to the over-compensation parameter.

[0117] The deep learning-based wind turbine voltage sag compensation system provided in an embodiment of the present invention can execute the deep learning-based wind turbine voltage sag compensation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0118] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules may be used and run on the user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other and are not used to limit the scope of protection of the present invention.

[0119] The above specific embodiments do not constitute a limitation to the scope of protection of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements made within the spirit and principles of this application should be included in the scope of protection of this application. In some cases, the actions or steps recorded in this application can be performed in an order different from that in the embodiments and can still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A wind turbine voltage sag compensation method based on deep learning, characterized in that: include: Real-time collection of three-phase voltage output data and unit operating status data of wind turbines; Determine whether a voltage sag occurs based on the three-phase voltage output data, and if so, extract sag elements to generate voltage sag element features; Perform fault triggering and fluctuation triggering judgment based on the unit operating status data and the voltage sag factor characteristics, and generate trigger source information; Perform compensation analysis by a voltage sag compensator according to the trigger source information and the voltage sag factor characteristics, and output a compensation decision; Voltage sag compensation control is performed according to the compensation decision.

2. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 1, wherein: Extract voltage sag elements and generate voltage sag element features, including: Calculating the voltage drop amplitude based on the three-phase voltage output data to generate voltage sag amplitude information; Performing Fourier transform on the three-phase voltage output data, identifying phase characteristics based on the transform result, and generating voltage sag phase information; Extracting the duration based on the timestamp in the three-phase voltage output data to generate sag duration information; The voltage sag element feature is constructed based on the voltage sag amplitude information, the voltage sag phase information, and the sag duration information.

3. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 1, wherein: Based on the unit operating status data and the voltage sag factor characteristics, fault triggering and fluctuation triggering judgment are performed to generate trigger source information, including: Construct a voltage sag judgment sample library; Traversing each voltage sag judgment sample in the voltage sag judgment sample library, inputting the sample together with the unit operation status data and the voltage sag element features into the twin network for similarity identification, and obtaining similarity output information; If the similarity output information meets a preset threshold, historical voltage sag trigger factors corresponding to the voltage sag judgment sample are collected, fault trigger and fluctuation trigger judgment are completed, and the trigger source information is generated.

4. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 3, wherein: Fault triggering refers to a voltage sag caused by an internal fault in the wind turbine or a grid fault, and fluctuation triggering refers to a voltage sag caused by wind speed fluctuations.

5. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 1, wherein: Performing compensation analysis by a voltage sag compensator according to the trigger source information and the voltage sag factor characteristics and outputting a compensation decision includes: Performing digital simulation modeling on the wind turbine to construct a twin turbine model, wherein the modeling includes integrated modeling of the turbine itself, the connected power grid, and meteorological conditions; Perform trigger source statistics based on historical voltage sag record information, generate multiple preset trigger sources, and collect historical voltage sag recovery record sets; Loading the historical voltage sag recovery record set into the twin unit model, learning the relationship between voltage compensation parameters and voltage sag factors and trigger sources, and constructing the voltage sag compensator, wherein the voltage sag compensator is connected to the twin unit model; The trigger source information and the voltage sag element characteristics are input into the voltage sag compensator, and after compensation parameter analysis, the twin unit model is called to perform compensation verification. After the verification is passed, the compensation decision is output.

6. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 5, wherein: If the trigger source is a fault-type trigger source, the historical voltage sag recovery record set is loaded into the twin unit model, the relationship between the fault isolation characteristics and the fault source location is learned, and the voltage sag compensator is optimized.

7. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 1, wherein: After outputting the compensation decision, it includes: Collecting information of several other wind turbines connected to the wind turbine; Obtaining a plurality of voltage compensation decisions of the plurality of other wind turbine generator sets through a data sharing channel; Performing global synergistic impact analysis and local compensation optimization on the compensation decision and the plurality of voltage compensation decisions to generate a local optimized compensation decision; The local optimization compensation decision is sent to the corresponding wind turbine for optimization control.

8. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 7, wherein: Performing a global synergistic impact analysis and local compensation optimization on the compensation decision and the plurality of voltage compensation decisions to generate a local optimized compensation decision, including: performing global fusion of compensation effects on the compensation decision and the plurality of voltage compensation decisions to generate global fusion compensation effect information; Based on the global fusion compensation effect information, it is determined whether there is over-compensation. If so, the local unit with the largest local compensation amount is extracted, and the compensation decision is adjusted according to the over-compensation parameter to generate the local optimized compensation decision.

9. The method for compensating for voltage sag of a wind turbine generator system based on deep learning according to claim 8, wherein: Compensation decision adjustment is performed according to the overcompensation parameters, including: By analyzing the relationship between historical local compensation data and global fusion compensation data, a local-global fusion relationship is constructed; Compensation decision adjustment is performed according to the overcompensation parameter based on the local-global fusion relationship.

10. A wind turbine voltage sag compensation system based on deep learning, characterized in that: A system for implementing the deep learning-based wind turbine voltage sag compensation method according to any one of claims 1 to 9, comprising: Wind turbine data acquisition module, used to collect three-phase voltage output data and unit operating status data of wind turbines in real time; a voltage sag behavior judgment module, which judges whether a voltage sag behavior occurs based on the three-phase voltage output data, and if so, extracts sag elements to generate voltage sag element features; a trigger source information acquisition module, which determines fault triggering and fluctuation triggering based on the unit operating status data and the voltage sag element characteristics, and generates trigger source information; a compensation decision output module, configured to perform compensation analysis through a voltage sag compensator according to the trigger source information and the voltage sag factor characteristics, and output a compensation decision; The voltage sag compensation control execution module is configured to perform voltage sag compensation control according to the compensation decision.

Images