Reactive power control method, reactive power control device and reactive power control equipment for permanent magnet direct-driven wind turbine generator
By using dynamic weight allocation and pre-trained models to generate fundamental reactive power regulation commands, the reactive power control problem of permanent magnet direct-drive wind turbines under nonlinear operating conditions was solved, achieving coordinated optimization of grid voltage stability and unit safety.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-11-19
- Publication Date
- 2026-04-10
AI Technical Summary
Existing reactive power control methods for permanent magnet direct-drive wind turbines are difficult to adapt to nonlinear dynamic conditions such as sudden wind speed changes and grid voltage fluctuations in wind power scenarios, resulting in grid voltage overshoot or oscillations and an inability to coordinate responses to multiple demands such as grid dispatch commands, reactive power deficit compensation, and harmonic suppression.
By employing a dynamic weight allocation algorithm combined with grid voltage deviation parameters, the weight ratio of multi-source reactive power demand signals is dynamically adjusted, and a fundamental reactive power regulation command is generated through a pre-trained reactive power regulation optimization model to achieve reactive power control of permanent magnet direct-drive wind turbine units.
It enables flexible adjustments under scenarios of grid voltage fluctuations and changes in dispatching instructions, ensuring grid voltage stability and power quality, guaranteeing the safe and stable operation of generating units, and balancing the coordinated satisfaction of multi-source demands with the rigid constraints of unit safety.
Smart Images
Figure CN121840808A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power system new energy power generation control, and in particular to a reactive power control method, device and electronic equipment for permanent magnet direct drive wind turbine. Background Technology
[0002] With the acceleration of the global energy transition, wind power, as a core component of clean energy, is experiencing a continuous increase in penetration. Permanent magnet direct-drive wind turbines, with their advantages of simple drive chains, high operating efficiency, and strong adaptability to low wind speeds, are widely used in large-scale wind farms. When these units are connected to the grid, not only is efficient transmission of active power required, but their reactive power control performance is also directly related to grid voltage stability, power quality compliance, and transient operational safety. They must respond to reactive power adjustment commands issued by the grid dispatch center, compensate for grid reactive power deficits to maintain the power factor within the standard range, and suppress harmonic distortion generated during grid connection. They must also cope with complex operating conditions such as grid voltage fluctuations and load surges. This is a key technical support for ensuring the safe and stable operation of the grid after a high proportion of wind power is connected to the grid.
[0003] In existing technologies, reactive power control methods for permanent magnet direct-drive wind turbines mostly rely on PI (proportional-integral) regulators. These regulators calculate reactive power deviations by collecting grid-side voltage and current signals, generating converter control commands to achieve a single objective of constant voltage or constant power factor regulation. However, this approach has significant limitations. PI regulators rely on fixed parameters to achieve linear regulation, making it difficult to adapt to nonlinear dynamic conditions such as sudden wind speed changes and grid voltage fluctuations in wind power scenarios. In rapidly changing scenarios, they are prone to regulation lag, leading to grid voltage overshoot or oscillation. Moreover, the control objective is singular, focusing only on constant voltage or constant power factor, and cannot coordinate responses to multiple demands such as grid dispatch commands, reactive power deficit compensation, and harmonic suppression. This results in insufficient adaptability in complex grid environments.
[0004] Therefore, how to dynamically integrate multi-source reactive power demand and adapt to nonlinear operating conditions is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention
[0005] The purpose of this invention is to provide at least one method, device, and electronic equipment for reactive power control of permanent magnet direct-drive wind turbine generators, which can dynamically integrate multi-source reactive power demand and adapt to nonlinear operating conditions.
[0006] To address the aforementioned technical problems, at least one embodiment of this application provides a reactive power control method for permanent magnet direct-drive wind turbine generators, comprising: Collect multi-source reactive power demand signals, unit constraint parameters, and grid status parameters of permanent magnet direct-drive wind turbine generators; Voltage deviation features are extracted based on the power grid state parameters to obtain the power grid voltage deviation parameters; Based on the grid voltage deviation parameter, a dynamic weight allocation algorithm is used to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal; Based on the calculated weights, the multi-source reactive power demand signals are fused to obtain demand data. The demand data and the unit constraint parameters are input into the pre-trained reactive power regulation optimization model to generate the fundamental reactive power regulation command. The fundamental frequency reactive power regulation command is sent to the grid-side converter, and the grid-side converter executes the command to achieve reactive power control of the permanent magnet direct drive wind turbine.
[0007] In one embodiment, the step of calculating the weight ratio of each type of demand signal in the multi-source reactive power demand signal using a dynamic weight allocation algorithm based on the grid voltage deviation parameter includes: The grid voltage deviation parameter is compared with the preset division threshold of each interval to determine the corresponding interval, which is then used as the target interval. Extract the weighted benchmark values of the power grid dispatch instructions and voltage deviation-derived reactive power corresponding to the target interval; Calculate the rate of change of the grid voltage deviation parameter; Based on the preset rate-adjustment rules, the weight benchmark value is dynamically corrected according to the rate of change to obtain the final weight ratio of each type of demand signal.
[0008] In one embodiment, the step of dynamically correcting the weighted benchmark value based on the rate of change according to a preset rate-adjustment rule includes: When the rate of change is positive, the weight of the reactive power derived from the voltage deviation is increased based on the weighted benchmark value. When the rate of change is negative, the weight of the reactive power derived from the voltage deviation is reduced based on the weighted benchmark value, while the weight of the power grid dispatching command is increased. The adjustment range of the weight is positively correlated with the absolute value of the rate of change.
[0009] In one embodiment, the grid-side converter execution command further includes: Receive the actual reactive power output value fed back by the grid-side converter in real time; Calculate the deviation between the actual reactive power output value and the demand data; A correction amount is generated based on the deviation; The fundamental reactive power adjustment command is updated and reissued based on the correction amount.
[0010] In one embodiment, the reactive power control method for the permanent magnet direct-drive wind turbine further includes: Receive the real-time grid voltage value fed back by the grid-side converter; Based on the real-time grid voltage value, the grid voltage deviation parameter is recalculated, and the demand data is updated based on the recalculated grid voltage deviation parameter. The step of calculating the deviation between the actual reactive power output value and the demand data includes: calculating the deviation between the actual reactive power output value and the updated demand data; The step of generating a correction amount based on the deviation includes: generating a correction amount based on the deviation and the voltage state corresponding to the real-time value of the grid voltage.
[0011] In one embodiment, the reactive power control method for the permanent magnet direct-drive wind turbine further includes: Collect power grid harmonic distortion rate; If the distortion rate exceeds a preset benchmark, a compensation component with the opposite phase to the power grid harmonics is generated; the amplitude of the compensation component increases as the harmonic distortion rate increases. The compensation component is superimposed on the modified fundamental reactive power regulation command.
[0012] In one embodiment, the step of sending the fundamental reactive power regulation command to the grid-side converter, and then using the grid-side converter to execute the command to achieve reactive power control of the permanent magnet direct-drive wind turbine, includes: The fundamental frequency reactive power regulation command is converted into a communication protocol format supported by the converter. The encrypted instruction content after format conversion is encrypted using an encryption algorithm to generate a verification code containing the instruction and a timestamp; The verification code is sent to the grid-side converter so that the grid-side converter can receive it and execute the instruction after successful verification.
[0013] In one embodiment, the step of extracting voltage deviation features based on the power grid state parameters to obtain power grid voltage deviation parameters includes: Extract the real-time grid voltage value for the current period from the grid state parameters, and obtain the instantaneous deviation by the difference between the real-time grid voltage value and the rated value. Calculate the ratio of the instantaneous deviation to the nominal value, as an instantaneous characteristic; The real-time voltage values of the power grid for multiple consecutive cycles are extracted from the power grid state parameters, and the deviation change trend is obtained by fitting the data using a linear regression algorithm, which is then used as the trend feature. The instantaneous characteristics and the trend characteristics are combined to form the grid voltage deviation parameter.
[0014] At least one embodiment of this application also provides a reactive power control device for a permanent magnet direct-drive wind turbine generator, comprising: The data acquisition module is used to collect multi-source reactive power demand signals, unit constraint parameters, and grid status parameters of permanent magnet direct-drive wind turbines. The deviation feature extraction module is used to extract voltage deviation features based on the power grid state parameters to obtain power grid voltage deviation parameters. The weight calculation module is used to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal based on the grid voltage deviation parameter and using a dynamic weight allocation algorithm. The data fusion module is used to perform data fusion on the multi-source reactive power demand signals according to the calculated weights to obtain demand data; The instruction generation module is used to input the demand data and the unit constraint parameters into the pre-trained reactive power regulation optimization model to generate the fundamental reactive power regulation instruction. The instruction execution module is used to send the fundamental reactive power regulation instruction to the grid-side converter, and the grid-side converter executes the instruction to realize reactive power control of the permanent magnet direct drive wind turbine.
[0015] At least one embodiment of this application also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the above-described reactive power control method for permanent magnet direct drive wind turbine generators.
[0016] The reactive power control method for permanent magnet direct-drive wind turbines provided in this application simultaneously acquires multi-source reactive power demand signals, unit constraint parameters, and grid state parameters. It transforms the original grid state parameters into quantified voltage deviation parameters, extracts grid voltage deviation state characteristics, and dynamically adjusts the weight ratio of multi-source demand signals based on the voltage deviation parameters. This deeply binds demand priority with the real-time grid state, overcoming the limitation of fixed weights failing to adapt to dynamic grid changes. Then, guided by these weights, it integrates multi-source heterogeneous demand signals, eliminating redundancy and conflicts between different demands, forming standardized demand data. Finally, using a pre-trained reactive power regulation optimization model, it combines demand data with unit constraint parameters to achieve intelligent decision-making.
[0017] This method enables the control strategy to adapt to the grid state in real time. It can flexibly adjust the control logic according to grid voltage fluctuations, changes in dispatching instructions, and other scenarios. While ensuring grid voltage stability and optimizing power quality, it also ensures the safe and stable operation of the generating units. Furthermore, it takes into account the coordinated satisfaction of multiple source demands and the rigid constraints of unit safety, and achieves multi-objective coordinated optimization of grid demand, unit capacity and control effect. Attached Figure Description
[0018] One or more embodiments are illustrated by way of example with reference to the accompanying drawings, and these illustrative descriptions do not constitute a limitation on the embodiments.
[0019] Figure 1 This is a flowchart of a reactive power control method for a permanent magnet direct-drive wind turbine provided in one embodiment of this application; Figure 2 This is a schematic diagram of a reactive power control device for a permanent magnet direct-drive wind turbine provided in one embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in one embodiment of this application. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been provided in the various embodiments of this application to help readers better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments. The division of the various embodiments below is for the convenience of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.
[0021] This invention proposes a reactive power control method for permanent magnet direct-drive wind turbine generators. The implementation details of the reactive power control method for permanent magnet direct-drive wind turbine generators in this embodiment are described below. The following content is only for the convenience of understanding and is not necessary for implementing this solution.
[0022] Example 1: The specific process of the reactive power control method for the permanent magnet direct-drive wind turbine in this embodiment can be described as follows: Figure 1 As shown, it includes: Step 101: Collect multi-source reactive power demand signals, unit constraint parameters, and grid status parameters of the permanent magnet direct-drive wind turbine.
[0023] Acquire the multi-source reactive power demand signals, unit constraint parameters, and grid status parameters required for reactive power control of permanent magnet direct-drive wind turbine units, providing comprehensive input for subsequent control strategy formulation.
[0024] Among them, multi-source reactive power demand signals refer to reactive power regulation demands from different scenarios, including but not limited to: reactive power instructions issued by the power grid dispatch center (to meet the overall dispatch requirements of the power grid), reactive power compensation demands derived from the deviation of the grid voltage from the rated value (to maintain voltage stability), and correction demands generated based on the power factor standard (to optimize power quality), etc. Unit constraint parameters are parameters that reflect the safe operating boundaries of wind turbine units, including but not limited to: the maximum output reactive power capacity of the converter (hardware capability limitations), the allowable limit value of the stator flux linkage (motor magnetic circuit safety constraints), and the threshold range of the DC bus voltage (converter system stability conditions), etc.; power grid status parameters are used to characterize the real-time operating status of the power grid, including but not limited to: the real-time value and rated value of the grid voltage and the current amplitude of each reactive power demand signal.
[0025] Step 102: Extract voltage deviation features based on grid state parameters to obtain grid voltage deviation parameters.
[0026] Based on the collected power grid state parameters, key information that can quantify the deviation of the power grid voltage from its rated state is extracted through feature analysis, forming the power grid voltage deviation parameter. Voltage deviation is the core basis for determining the weighting of multi-source reactive power demand signals. For example, when the voltage deviation is severe and widening, voltage compensation demand should be prioritized. This embodiment does not limit the specific feature types for voltage deviation characteristics. For instance, it can extract the absolute deviation between the real-time voltage value and the rated value (intuitively reflecting the specific numerical value of the deviation), the voltage deviation rate (the ratio of the absolute deviation to the rated value, standardized to characterize the degree of deviation); the slope of the voltage deviation over N consecutive cycles (reflecting the rate of deviation expansion or contraction), the acceleration of the deviation change (the rate of change of the slope, characterizing the steepness of the deviation trend); the number of cycles in which the voltage deviation persists (reflecting the duration of the deviation), the standard deviation of the deviation within consecutive cycles (characterizing the stability of voltage fluctuations); and the peak and trough values of the deviation (reflecting extreme deviation states), etc. These features quantify the voltage deviation state from multiple dimensions, including instantaneous deviation degree, dynamic change trend, continuous characteristics, and fluctuation stability. They can be flexibly selected or combined according to actual control accuracy requirements, providing more comprehensive decision support for subsequent dynamic weight allocation.
[0027] To enhance understanding, this embodiment further proposes a method for calculating the grid voltage deviation parameter, specifically including the following steps: extracting the real-time grid voltage value for the current period from the grid state parameters; obtaining the instantaneous deviation by calculating the difference between the real-time grid voltage value and the rated value; calculating the ratio of the instantaneous deviation to the rated value as an instantaneous feature; extracting the real-time grid voltage values for multiple consecutive periods from the grid state parameters; fitting the deviation change trend through a linear regression algorithm as a trend feature; and combining the instantaneous feature and the trend feature as the grid voltage deviation parameter.
[0028] First, the real-time grid voltage value for the current period is extracted from the grid state parameters. The difference between this value and the grid's rated voltage is calculated to obtain the instantaneous deviation, which directly reflects the absolute degree of voltage deviation from the rated state. Then, this instantaneous deviation is proportionally calculated with the rated voltage to obtain the instantaneous feature, eliminating the influence of the rated voltage benchmark difference and achieving a normalized representation of the grid deviation degree at different voltage levels. Subsequently, real-time grid voltage values for multiple consecutive periods are extracted, and the data change pattern is fitted using a linear regression algorithm to quantify the deviation change trend (e.g., continuous expansion, gradual reduction, or stabilization), i.e., the trend feature, capturing the dynamic development trend of voltage deviation. Finally, the normalized instantaneous feature and the quantified trend feature are integrated to form a grid voltage deviation parameter that combines the current deviation degree with the future development trend. This deviation calculation method can solve the problem of the one-sidedness of single-dimensional deviation features, generating a more comprehensive, accurate, and three-dimensional voltage deviation quantification index, providing a scientific and reliable decision-making basis for subsequent dynamic weight allocation algorithms. It should be noted that other types of deviation calculations can refer to the description in this embodiment, and will not be elaborated further here.
[0029] This step, by integrating voltage deviation parameters, overcomes the limitations of relying solely on voltage values at a single moment. It can more comprehensively and dynamically capture the deviation state of the grid voltage, providing accurate and reliable input for the subsequent dynamic weight allocation algorithm. This ensures that the weight adjustment matches the actual needs of the grid, thereby improving the adaptability and pertinence of the reactive power control strategy.
[0030] Step 103: Based on the grid voltage deviation parameter, use a dynamic weight allocation algorithm to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal.
[0031] The priority of multi-source reactive power demand signals varies under different power grid operating conditions, and there may be target conflicts (such as the reactive power output required by dispatch instructions being inconsistent with the reactive power required for voltage compensation). If a fixed weight allocation is adopted, it is difficult to adapt to the dynamic changes of the power grid. This may lead to critical demands (such as emergency voltage compensation) being suppressed by low-priority demands, or over-responding to a certain demand while ignoring other reasonable demands, ultimately affecting the effectiveness of reactive power control and the stability of power grid operation.
[0032] To address this, this method proposes a dynamic weighting algorithm based on extracted grid voltage deviation parameters. This algorithm quantifies and allocates the priority of multi-source reactive power demand signals (such as grid dispatch commands, reactive power derived from voltage deviation, and power factor correction demands) according to the degree and trend of grid voltage deviation from its rated state. The algorithm dynamically adjusts the importance of reactive power demands from different sources, ultimately determining the weight of each type of demand signal in the total demand. For example, when the grid voltage deviates significantly and the deviation continues to widen, the weight of reactive power derived from voltage deviation is increased to prioritize voltage stability. When the voltage returns to the normal range, the weight of grid dispatch commands or power factor correction demands is increased to meet dispatch requirements or power quality standards.
[0033] The grid voltage deviation parameter directly reflects the core requirement of grid voltage stability. Adjusting the weights based on this parameter allows the weight allocation to be directly linked to the real-time operating status of the grid. This ensures that voltage compensation demands receive priority response when the grid is unbalanced, solving the problem of the disconnect between traditional fixed weights and the actual needs of the grid. On the other hand, dynamic adjustment of the weights overcomes the limitation of fixed weights in adapting to complex operating conditions. It can respond in real time to dynamic scenarios such as grid voltage fluctuations and changes in dispatching instructions. When the grid state changes, the weights can be updated synchronously to match the new demand priorities. This avoids the inadequacy of weight settings for multiple sources of demand under a single operating condition, and can flexibly balance multiple objectives such as grid dispatching, voltage stability, and power quality optimization under different operating scenarios. This significantly improves the adaptability and robustness of reactive power control strategies, ensuring that the fusion results of multi-source reactive power demand signals always align with the current core needs of the grid.
[0034] Step 104: Based on the calculated weights, perform data fusion on the multi-source reactive power demand signals to obtain demand data.
[0035] Multi-source reactive power demand signals originate from different scenarios and target different control objectives. Directly inputting these dispersed signals into the model to generate commands would lead to chaotic model decisions. In this step, based on the previously calculated weight proportions of various types of reactive power demand signals, a weighted fusion algorithm is used to integrate multi-source reactive power demand signals (such as grid dispatch commands, reactive power derived from voltage deviation, power factor correction demands, etc.). This transforms the dispersed, multi-dimensional demand signals into unified, quantified, single demand data, providing a standardized core input for the subsequent generation of fundamental reactive power regulation commands.
[0036] By using weighted fusion, demand signals can be rationally selected and integrated based on the priority corresponding to the current state of the power grid. This ensures that the final demand data highlights both core control objectives and secondary demands, avoiding the limitations of single-demand-oriented control. At the same time, it transforms dispersed and heterogeneous multi-source signals into unified and quantified demand data, eliminating redundant information and logical conflicts between signals. This provides standardized and highly reliable input for subsequent reactive power regulation optimization models, reduces the difficulty of model decision-making, and improves the accuracy of fundamental reactive power regulation command generation.
[0037] It should be noted that the fusion results are directly linked to the dynamic weights. When changes in the power grid state trigger weight adjustments, the demand data will be updated synchronously and adaptively to ensure the continuity of the control link from the perception of the power grid state to the quantification of demand, so that subsequent control actions always maintain a high degree of consistency with the actual needs of the power grid.
[0038] Step 105: Input the demand data and unit constraint parameters into the pre-trained reactive power regulation optimization model to generate the fundamental reactive power regulation command.
[0039] The multi-source reactive power demand data, after dynamic weighting and fusion, along with the wind turbine's own operating constraint parameters, are input into a pre-trained reactive power regulation optimization model. This model is an intelligent decision-making model built upon multi-source data training. Its core function is to receive the fused reactive power demand data and the constraint parameters corresponding to the turbine's safe operating boundary. Through built-in algorithm logic and learned mapping rules, it outputs fundamental reactive power regulation commands adapted to the wind turbine's operating state and grid demand. This transforms the multi-objective optimization problem of reactive power control (such as meeting voltage stability, dispatch requirements, and equipment safety) into a quantifiable command generation process, enabling automated decision-making under complex operating conditions without manual intervention.
[0040] The type and structure of the reactive power regulation optimization model are not limited in this embodiment. They can be flexibly selected based on actual control accuracy, response speed, and deployment conditions. For example, traditional optimization algorithm models (such as models based on linear programming, quadratic programming, and mixed integer programming) can be used, suitable for scenarios with clear constraints and simple objective functions. Alternatively, U-shaped EY deep learning models (such as convolutional neural networks, recurrent neural networks, and Transformer models) can be used to mine deep features in time-series power grid data, suitable for highly dynamic and high-dimensional data scenarios. Or, hybrid optimization models (such as models combining optimization algorithms and machine learning) can be used, combining rule interpretability with data-driven accuracy, suitable for scenarios requiring both reliability and flexibility. The above model types are only examples; other types can be referred to in this embodiment and will not be elaborated further.
[0041] The reactive power regulation optimization model is pre-trained offline based on historical data and simulation scenarios to learn the mapping rules between grid status, demand signals and unit constraints. In this embodiment, the training method of the model is not limited and can be implemented with reference to relevant technologies.
[0042] Step 106: The fundamental frequency reactive power regulation command is sent to the grid-side converter, and the reactive power control of the permanent magnet direct drive wind turbine is realized by the grid-side converter executing the command.
[0043] The generated quantized fundamental reactive power regulation commands (such as reactive power setpoints, phase control parameters, etc.) are accurately transmitted to the grid-side converter, the core actuator of the wind turbine. Relying on the power electronic conversion function of the grid-side converter, the commands are converted into actual reactive power regulation actions, ultimately achieving precise control of the reactive power output of the wind turbine.
[0044] Based on the above introduction, the reactive power control method for permanent magnet direct-drive wind turbines provided in this embodiment simultaneously acquires multi-source reactive power demand signals, unit constraint parameters, and grid state parameters. It transforms the original grid state parameters into quantified voltage deviation parameters, extracts grid voltage deviation state characteristics, and dynamically adjusts the weight ratio of multi-source demand signals based on the voltage deviation parameters. This deeply binds demand priority with the real-time grid state, overcoming the limitation that fixed weights cannot adapt to dynamic changes in the grid. Then, guided by the weights, it integrates multi-source heterogeneous demand signals, eliminating redundancy and conflicts between different demands, forming standardized demand data. With the help of a pre-trained reactive power regulation optimization model, it achieves intelligent decision-making by combining demand data and unit constraint parameters.
[0045] This method enables the control strategy to adapt to the grid state in real time. It can flexibly adjust the control logic according to grid voltage fluctuations, changes in dispatching instructions, and other scenarios. While ensuring grid voltage stability and optimizing power quality, it also ensures the safe and stable operation of the generating units. Furthermore, it takes into account the coordinated satisfaction of multiple source demands and the rigid constraints of unit safety, and achieves multi-objective coordinated optimization of grid demand, unit capacity and control effect.
[0046] Example 2: The above embodiments do not limit the specific implementation method of dynamic weight allocation. This embodiment proposes a dual design scheme of interval benchmark + rate correction, which takes into account both the stability and dynamic adaptability of weight allocation, and ensures that the weight adjustment does not deviate from the core requirements of the grid voltage state, and can respond to the deviation change trend in real time.
[0047] Specifically, step 103, based on the grid voltage deviation parameter, uses a dynamic weight allocation algorithm to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal, which can be performed according to the following steps: Step 31: Compare the grid voltage deviation parameter with the preset division threshold of each interval to determine the interval to which it belongs, and use it as the target interval.
[0048] Based on the power grid safety operation standards and voltage stability control requirements, the range of power grid voltage deviation parameters is divided into several discrete intervals (such as normal interval, slight deviation interval, severe deviation interval, etc.), and a corresponding threshold is set for each interval, that is, the critical value that distinguishes different intervals. In this embodiment, the number of intervals is not limited, and can be set according to the configuration requirements of the actual application scenario.
[0049] The extracted and quantifiable grid voltage deviation parameters, which represent the current grid voltage deviation state, are compared with the preset interval division thresholds one by one to determine which preset interval the actual deviation parameter falls into. Finally, the interval is determined as the target interval, providing a clear grade basis for the subsequent extraction of the weight benchmark value of the corresponding interval.
[0050] Step 32: Extract the weighted benchmark values of the power grid dispatch instructions and voltage deviation-derived reactive power corresponding to the target interval.
[0051] For each voltage state interval (such as normal, slight deviation, and severe deviation interval), based on the priority requirements of power grid operation (such as voltage stability taking precedence over dispatching instructions, and dispatching compliance taking precedence over non-emergency compensation), initial weight values (i.e. weight benchmark values) corresponding to power grid dispatching instructions and voltage deviation-derived reactive power are set respectively. The weight benchmark values corresponding to different intervals are configured differently (e.g., the weight benchmark value of voltage deviation-derived reactive power in the severe deviation interval is higher than that in the normal interval).
[0052] Once the target interval is determined, the two sets of initial weight benchmark values corresponding to the target interval are retrieved directly from the preset interval-weight benchmark value mapping relationship, providing a basis for subsequent dynamic correction based on the deviation change rate.
[0053] Step 33: Calculate the rate of change of the grid voltage deviation parameter.
[0054] Continuously collect grid voltage deviation parameters (such as voltage deviation rate and deviation trend combination value at different periods) under continuous time sequence. By calculating the difference of grid voltage deviation parameters between two or more consecutive moments, and then dividing by the corresponding time interval, the change of grid voltage deviation parameters per unit time is obtained, that is, the rate of change of grid voltage deviation parameters.
[0055] This rate value can intuitively reflect whether the current voltage deviation is expanding, shrinking, or changing steadily, providing a quantitative basis for subsequent dynamic adjustment of the weight benchmark value based on the changing trend.
[0056] Step 34: Based on the preset rate-adjustment rules, dynamically adjust the weight benchmark value according to the rate of change to obtain the final weight ratio of each type of demand signal.
[0057] The system backend pre-sets rate-adjustment rules corresponding to the rate of change of grid voltage deviation. These rules clarify the correspondence between the sign (reflecting the expansion or contraction of deviation), magnitude (reflecting the speed of deviation change), and weight adjustment direction (increasing or decreasing the weight of a certain type of demand) and adjustment magnitude of the rate of change (e.g., if the rate is positive and the larger the value, the greater the upward adjustment of the reactive power weight derived from the voltage deviation).
[0058] After obtaining the rate of change of the grid voltage deviation parameter calculated above, the system backend matches this rate with the preset rate-adjustment rules. Based on the adjustment logic specified in the rules, the weight benchmark values of the grid dispatch instructions extracted from the target interval and the reactive power derived from the voltage deviation are specifically corrected. Finally, the final weight ratio of each type of reactive power demand signal that can accurately match the current grid voltage deviation degree and trend is obtained, providing a clear priority basis for subsequent multi-source demand data fusion.
[0059] In this step, the specific rules for the rate-adjustment rule are not limited. To deepen understanding, an adjustment rule is proposed here: when the rate of change is positive, the weight of reactive power derived from voltage deviation is increased based on the weight benchmark value; when the rate of change is negative, the weight of reactive power derived from voltage deviation is decreased based on the weight benchmark value, while the weight of grid dispatch instructions is increased; wherein, the adjustment magnitude of the weight is positively correlated with the absolute value of the rate of change.
[0060] When the rate of change is positive, it indicates that the current grid voltage is deviating from its rated state and is expanding (e.g., the voltage is consistently too low or too high). In this case, based on the weighted benchmark value corresponding to the target interval, the weight ratio of reactive power derived from the voltage deviation is increased to prioritize strengthening voltage compensation and curb further deterioration of the deviation. When the rate of change is negative, it indicates that the voltage deviation is gradually decreasing and the grid state is stabilizing. In this case, the weight of reactive power derived from the voltage deviation is reduced, while the weight of grid dispatching instructions is increased accordingly to return to the overall grid dispatching planning requirements. Furthermore, the adjustment magnitude of the weight is positively correlated with the absolute value of the rate of change; that is, the faster the deviation expands or shrinks, the greater the adjustment magnitude of the weight, ensuring a rapid response to changes in the grid state. It should be noted that this embodiment only uses the above rule settings as an example; reactive power control under other rules can refer to the description in this embodiment, and will not be elaborated further here.
[0061] Based on the above introduction, this embodiment determines the target interval by comparing the grid voltage deviation parameter with the preset interval threshold and extracting the corresponding weight benchmark value. This ensures the precise binding of weight allocation with the grid voltage deviation level, providing a clear and quantifiable initial basis for weight setting and avoiding random and blind adjustments. By calculating the rate of change of the voltage deviation parameter and correcting the weight benchmark value according to preset rules, the weight can respond in real time to the development trend of voltage deviation (such as strengthening the weight of voltage compensation demand when the deviation increases and optimizing the weight ratio of dispatching instructions when the deviation decreases), breaking through the limitation of relying solely on static intervals. The entire process is clear and logically closed-loop, ensuring the consistency and reliability of weight allocation under different voltage states and enabling rapid adaptation of weight to dynamic changes in the grid. The final weight ratio can accurately balance the priority of multi-source reactive power demand, providing scientific and efficient decision support for subsequent multi-source demand data fusion, thereby improving the pertinence and robustness of the overall reactive power control strategy.
[0062] Example 3: In actual operation scenarios, the execution of the initial fundamental reactive power regulation command is susceptible to various interference factors: on the one hand, the hardware characteristics of the grid-side converter (such as power switching device losses and device aging) and the dynamic fluctuations of the grid operating status (such as voltage surges and load changes) may cause deviations between the actual reactive power output value and the command expectation; on the other hand, real-time changes in the wind turbine's own operating conditions (such as wind speed changes and component operating status) may also cause a mismatch between the initial command and the actual execution capability, thus resulting in deviations. To correct this deviation, this embodiment further proposes a feedback correction mechanism. Without a feedback correction mechanism, such deviations will persist or even amplify, preventing the achievement of core control objectives such as grid voltage stability and power quality optimization. This closed-loop correction process can capture deviations in real time and dynamically calibrate commands, effectively offsetting the impact of various interference factors. This ensures that reactive power control always aligns with the actual needs of the grid and generating units, avoiding control failures caused by a disconnect between commands and execution, and significantly improving the reliability, robustness, and accuracy of the overall reactive power control strategy.
[0063] In step 106, the grid-side converter execution command can be further executed in the following steps: Step 107: Receive the actual reactive power output value fed back by the grid-side converter in real time.
[0064] During the execution of commands by the grid-side converter, the actual reactive power output data (i.e., actual reactive power output value) of the wind turbine generators is continuously received from the converter. This value is a specific quantitative result of the reactive power actually output to or absorbed from the grid after the grid-side converter executes the fundamental reactive power regulation command, directly reflecting the actual effect of command execution.
[0065] Step 108: Calculate the deviation between the actual reactive power output value and the demand data.
[0066] The actual reactive power output value is compared with the target demand data (i.e., the ideal target value that reactive power control needs to achieve) determined after multi-source demand fusion. Through preset calculation logic (such as actual value minus target value or target value minus actual value), a numerical result that can quantitatively characterize the magnitude and direction of the difference between the two is obtained. This result is the deviation. This deviation directly reflects the specific degree to which the current reactive power output has not reached or exceeded the target. It is the core basis for subsequent judgment on whether to adjust the command and how to generate the correction amount, providing accurate quantitative support for realizing dynamic calibration of the command.
[0067] Step 109: Generate correction amount based on deviation.
[0068] A pre-defined mapping logic between deviation and correction amount is established to determine the direction of the deviation (e.g., a positive deviation if the actual output is lower than the required data, and a negative deviation if it is higher than the required data), its magnitude, and its trend of change (e.g., the deviation continues to expand or shrink), and then calculates the corresponding correction amount based on the pre-defined logic.
[0069] For example, a positive correction is generated when there is a positive deviation (to increase reactive power output), and a negative correction is generated when there is a negative deviation (to reduce reactive power output). The magnitude of the correction is matched with the severity of the deviation (the larger the deviation, the larger the correction is usually to ensure rapid correction).
[0070] Step 110: Update the fundamental reactive power adjustment command according to the correction amount and reissue it.
[0071] The generated correction value is numerically integrated with the original fundamental reactive power regulation command to form a new fundamental reactive power regulation command adapted to the current deviation correction requirements. Subsequently, the system performs communication protocol adaptation and format conversion on the updated command to ensure that it can be accurately recognized and parsed by the grid-side converter. Finally, the updated command is reissued to the grid-side converter to guide the converter to adjust the operating status of its internal power switching devices and correct the deviation between the original execution action and the target requirement.
[0072] Based on the above introduction, this embodiment can quickly capture command execution deviations by collecting actual reactive power output values in real time and comparing them with demand data. Then, through correction amount generation and command update, the converter's execution action is dynamically calibrated, effectively narrowing the gap between actual output and target demand. This solves the problem of control deviation caused by static commands being susceptible to interference from grid fluctuations, device losses, etc. Facing complex interference factors such as sudden changes in grid load, drift of converter hardware characteristics, and fluctuations in wind turbine operating conditions, the closed-loop mechanism can respond to and adjust commands in real time, avoiding the accumulation and expansion of deviations, and ensuring that the reactive power control effect always stably matches the target, breaking through the limitation of weak anti-interference capability of traditional open-loop control.
[0073] Example 4: Building upon Example 3, to further improve the response speed and adaptation accuracy of reactive power control to dynamic changes in the power grid, and to strengthen the dynamic linkage between the real-time status of the power grid voltage and the control link, this example proposes a collaborative logic for adding real-time sensing of the power grid voltage and dynamic updating of demand data. Specifically, in addition to the steps described above, the following steps can be further performed: Step 111: Receive the real-time grid voltage value fed back from the grid-side converter.
[0074] Step 112: Based on the real-time grid voltage value, recalculate the grid voltage deviation parameter, and update the demand data based on the recalculated grid voltage deviation parameter.
[0075] During the execution of commands and correction of existing feedback by the grid-side converter, the system backend needs to synchronously receive the real-time grid voltage value fed back by the grid-side converter. Using this as the new state input, the grid voltage deviation parameter is recalculated. That is, based on the difference between the current real grid voltage and the rated value, the quantified deviation index is updated. Based on the updated voltage deviation parameter, the weight allocation and data fusion process of multi-source reactive power demand signals are readjusted to generate updated demand data that adapts to the current grid state.
[0076] Step 108 calculates the deviation between the actual reactive power output value and the demand data, including: calculating the deviation between the actual reactive power output value and the updated demand data; subsequent deviation calculation steps no longer use the initial demand data, but instead use the updated demand data and the actual reactive power output value to perform difference calculation.
[0077] Step 109 generates a correction amount based on the deviation, including: generating a correction amount based on the deviation and the voltage state corresponding to the real-time grid voltage value. The generation of the correction amount no longer depends solely on a single deviation, but rather combines the magnitude of the deviation with the voltage state corresponding to the real-time grid voltage value (such as normal, slight deviation, severe deviation, etc.) to comprehensively determine the correction direction and adjustment range.
[0078] This method dynamically updates demand data by feeding back the real-time value of the grid voltage, ensuring that the control target is always precisely aligned with the current state of the grid and avoiding control mismatch caused by the lag in demand data. On the other hand, the generation of correction amount is combined with the deviation and the current voltage state, which not only ensures the effective correction of deviation at the execution end, but also adjusts the correction intensity according to the urgency of the grid (such as increasing the correction amount when there is a severe deviation to quickly respond to the demand for voltage stability), avoiding the limitations of a single correction logic under complex operating conditions.
[0079] Example 5: Building upon Example 3, to further expand the power quality optimization dimension of reactive power control methods and achieve the dual objectives of precise fundamental reactive power regulation and proactive harmonic pollution control, this example further proposes a power grid harmonic control method. Specifically, in addition to the steps described above, the following steps can be further performed: Step 113: Collect the power grid harmonic distortion rate.
[0080] Throughout the entire process of reactive power control, the harmonic distortion rate of the power grid is collected synchronously. This is a quantitative indicator of the degree to which the voltage or current waveform of the power grid deviates from the sine wave, reflecting the severity of harmonic pollution in the power grid.
[0081] Step 114: If the distortion rate exceeds the preset benchmark, a compensation component with the opposite phase to the power grid harmonics is generated; the amplitude of the compensation component increases as the harmonic distortion rate increases.
[0082] The collected harmonic distortion rate is compared with the preset qualified benchmark value. If the distortion rate is determined to be outside the benchmark range, it indicates that the harmonic pollution of the power grid has affected the power quality. Then, a harmonic compensation component that is completely opposite to the phase of the current power grid harmonics is automatically generated. The amplitude of the compensation component follows the dynamic adaptation rule that the higher the harmonic distortion rate, the larger the amplitude.
[0083] It should be noted that if the distortion rate does not exceed the preset benchmark, this embodiment does not limit the processing method in this case, and can jump to step 113 to continuously collect and monitor data.
[0084] In addition, this embodiment does not limit the specific value setting of the preset benchmark, and it can be set according to the actual application scenario.
[0085] Step 115: Add the compensation component to the corrected fundamental reactive power regulation command.
[0086] The harmonic compensation component is superimposed onto the fundamental reactive power regulation command after closed-loop feedback correction to form a composite control command that combines fundamental reactive power regulation and harmonic mitigation functions, driving the grid-side converter to execute the two control tasks simultaneously.
[0087] Based on the above introduction, this embodiment can perceive the state of harmonic pollution in the power grid in real time by collecting and monitoring the harmonic distortion rate, and achieve accurate identification of harmonic problems. In addition, by generating an anti-phase compensation component and superimposing it on the control command, the power electronic conversion capability of the grid-side converter is used to output a compensation current / voltage to the power grid that is opposite in phase and matches the amplitude of the harmonics, thereby offsetting the influence of the original harmonics in the power grid, reducing the harmonic distortion rate, and improving power quality.
[0088] Example 6: To ensure the security, accuracy, and immutability of the fundamental frequency reactive power regulation command from issuance to execution, and to avoid equipment malfunctions or grid security risks caused by abnormal command transmission, this embodiment proposes a secure command transmission mechanism. Specifically, step 106 sends the fundamental frequency reactive power regulation command to the grid-side converter. The grid-side converter executes the command to achieve reactive power control of the permanent magnet direct-drive wind turbine. This can be performed according to the following steps: Step 61: Convert the fundamental frequency reactive power regulation command into a communication protocol format supported by the converter.
[0089] The fundamental reactive power regulation command generated earlier needs to be converted into a standard communication protocol format supported by the grid-side converter hardware (such as Modbus, CANopen, EtherCAT, and other industrial communication protocols) to ensure that the command can be accurately recognized and parsed at the hardware level. The specific communication protocol format is set according to the actual application scenario and is not limited here.
[0090] Step 62: Encrypt the converted instruction content using an encryption algorithm to generate a verification code containing the instruction and a timestamp.
[0091] The converted instruction content is encrypted using a preset encryption algorithm (such as symmetric encryption AES, asymmetric encryption RSA, etc.), and timestamp information is embedded to generate an integrated verification code containing encrypted instruction data and timestamp verification information, ensuring the confidentiality and uniqueness of the instruction content.
[0092] Step 63: Send a verification code to the grid-side converter so that the grid-side converter can receive it and execute the command after the verification is successful.
[0093] The verification code is sent to the grid-side converter through a reliable communication channel. After receiving it, the converter must first verify the validity of the timestamp in the verification code and the integrity of the encrypted instruction, such as whether the timestamp is within the valid window and whether the encrypted signature matches. Only after the verification is passed can the instruction be decrypted and the corresponding reactive power regulation action be executed.
[0094] Based on the above introduction, this embodiment firstly effectively resists malicious attacks such as tampering and forgery through encryption processing and verification code generation, ensuring the authenticity and confidentiality of the instruction content and preventing illegal instructions from intervening in the control link. The timestamp embedding and verification mechanism eliminates replay attacks and the execution of outdated instructions, ensuring that the converter executes valid instructions that are precisely matched to the current control scenario. In addition, the protocol format conversion solves the compatibility problem of different hardware devices, ensuring the feasibility of instructions at the physical transmission level and avoiding execution blockages caused by format differences, providing key guarantees for the safe and stable operation of the power grid and generating units.
[0095] Example 7: This embodiment relates to a reactive power control device for a permanent magnet direct-drive wind turbine generator set. A schematic diagram of the reactive power control device for this embodiment is shown below. Figure 2 As shown, it includes: a data acquisition module 201, a deviation feature extraction module 202, a weight calculation module 203, a data fusion module 204, an instruction generation module 205, and an instruction execution module 206.
[0096] Among them, the data acquisition module 201 is used to acquire multi-source reactive power demand signals, unit constraint parameters and grid status parameters of permanent magnet direct drive wind turbine units; The deviation feature extraction module 202 is used to extract voltage deviation features based on power grid state parameters to obtain power grid voltage deviation parameters. The weight calculation module 203 is used to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal based on the grid voltage deviation parameter and using a dynamic weight allocation algorithm. The data fusion module 204 is used to fuse multi-source reactive power demand signals according to the calculated weights to obtain demand data; The instruction generation module 205 is used to input demand data and unit constraint parameters into a pre-trained reactive power regulation optimization model to generate fundamental reactive power regulation instructions. The instruction execution module 206 is used to send the fundamental wave reactive power regulation instruction to the grid-side converter, and realize the reactive power control of the permanent magnet direct drive wind turbine through the grid-side converter executing the instruction.
[0097] It should be noted that the contents of the reactive power control device for permanent magnet direct drive wind turbine provided in this embodiment can be referred to in conjunction with the reactive power control method for permanent magnet direct drive wind turbine provided in the above embodiments, and the repeated parts will not be described again in this embodiment.
[0098] In the reactive power control device for permanent magnet direct-drive wind turbines provided in this embodiment, the data acquisition module comprehensively covers multi-source reactive power demand signals, unit constraint parameters, and grid status parameters, providing complete data support for subsequent control decisions based on external demand, internal constraints, and real-time status, avoiding the bias of decisions caused by a single data source; the deviation feature extraction module generates grid voltage deviation parameters that combine the current voltage deviation degree and trend through multi-dimensional feature extraction logic, providing accurate status basis for weight allocation; the weight calculation module adopts a dynamic weight allocation algorithm to adapt the priority of multi-source demands to the real-time status of the grid, breaking through the limitations of fixed weights; the data fusion module integrates heterogeneous demands guided by weights, eliminates redundant conflicts, and forms standardized, highly reliable demand data; the instruction generation module relies on a pre-trained reactive power regulation optimization model, taking into account both demand satisfaction and unit constraints, and quickly outputs physically achievable and accurate instructions; the instruction execution module is directly connected to the grid-side converter to ensure efficient instruction implementation. This device, through a closed-loop chain of state perception, dynamic weight adjustment, precise command generation, and execution, can achieve rapid response to dynamic scenarios such as grid voltage fluctuations and demand changes, ensuring dynamic matching between control effects and real-time grid demands.
[0099] Furthermore, it should be noted that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. In addition, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units do not exist in this embodiment.
[0100] Example 8: Another embodiment of this application relates to an electronic device, such as... Figure 3 As shown, it includes: at least one processor 301; and a memory 302 communicatively connected to at least one processor 301; wherein the memory 302 stores instructions that can be executed by at least one processor 301, and the instructions are executed by at least one processor 301 to enable at least one processor 301 to perform the steps of the reactive power control method of the permanent magnet direct drive wind turbine in the above embodiments.
[0101] The memory and processor are connected via a bus, which can include any number of interconnecting buses and bridges, connecting various circuits of one or more processors and memories. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and will not be described further herein. The bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by the processor is transmitted over the wireless medium via an antenna, which further receives data and transmits it to the processor.
[0102] The processor manages the bus and general processing, and also provides various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory is used to store data used by the processor during operation.
[0103] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.
Claims
1. A reactive power control method for a permanent magnet direct-drive wind turbine generator, characterized in that, include: Collect multi-source reactive power demand signals, unit constraint parameters, and grid status parameters of permanent magnet direct-drive wind turbine generators; Voltage deviation features are extracted based on the power grid state parameters to obtain the power grid voltage deviation parameters; Based on the grid voltage deviation parameter, a dynamic weight allocation algorithm is used to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal; Based on the calculated weights, the multi-source reactive power demand signals are fused to obtain demand data. The demand data and the unit constraint parameters are input into the pre-trained reactive power regulation optimization model to generate the fundamental reactive power regulation command. The fundamental frequency reactive power regulation command is sent to the grid-side converter, and the grid-side converter executes the command to achieve reactive power control of the permanent magnet direct drive wind turbine.
2. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 1, characterized in that, The step of calculating the weight ratio of each type of demand signal in the multi-source reactive power demand signal using a dynamic weight allocation algorithm based on the grid voltage deviation parameter includes: The grid voltage deviation parameter is compared with the preset division threshold of each interval to determine the corresponding interval, which is then used as the target interval. Extract the weighted benchmark values of the power grid dispatch instructions and voltage deviation-derived reactive power corresponding to the target interval; Calculate the rate of change of the grid voltage deviation parameter; Based on the preset rate-adjustment rules, the weight benchmark value is dynamically corrected according to the rate of change to obtain the final weight ratio of each type of demand signal.
3. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 2, characterized in that, The method of dynamically adjusting the weighted benchmark value based on the preset rate-adjustment rule according to the rate of change includes: When the rate of change is positive, the weight of the reactive power derived from the voltage deviation is increased based on the weighted benchmark value. When the rate of change is negative, the weight of the reactive power derived from the voltage deviation is reduced based on the weighted benchmark value, while the weight of the power grid dispatching command is increased. The adjustment range of the weight is positively correlated with the absolute value of the rate of change.
4. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 1, characterized in that, The grid-side converter execution command also includes: Receive the actual reactive power output value fed back by the grid-side converter in real time; Calculate the deviation between the actual reactive power output value and the demand data; A correction amount is generated based on the deviation; The fundamental reactive power adjustment command is updated and reissued based on the correction amount.
5. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 4, characterized in that, Also includes: Receive the real-time grid voltage value fed back by the grid-side converter; Based on the real-time grid voltage value, the grid voltage deviation parameter is recalculated, and the demand data is updated based on the recalculated grid voltage deviation parameter. The step of calculating the deviation between the actual reactive power output value and the demand data includes: calculating the deviation between the actual reactive power output value and the updated demand data; The step of generating a correction amount based on the deviation includes: generating a correction amount based on the deviation and the voltage state corresponding to the real-time value of the grid voltage.
6. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 4, characterized in that, Also includes: Collect power grid harmonic distortion rate; If the distortion rate exceeds a preset benchmark, a compensation component with the opposite phase to the power grid harmonics is generated. The amplitude of the compensation component increases as the harmonic distortion rate increases; The compensation component is superimposed on the modified fundamental reactive power regulation command.
7. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 1, characterized in that, The step of sending the fundamental reactive power regulation command to the grid-side converter, and then using the grid-side converter to execute the command to achieve reactive power control of the permanent magnet direct-drive wind turbine, includes: The fundamental frequency reactive power regulation command is converted into a communication protocol format supported by the converter. The encrypted instruction content after format conversion is encrypted using an encryption algorithm to generate a verification code containing the instruction and a timestamp; The verification code is sent to the grid-side converter so that the grid-side converter can receive it and execute the instruction after successful verification.
8. The reactive power control method for permanent magnet direct-drive wind turbine generators according to claim 1, characterized in that, The step of extracting voltage deviation features based on the power grid state parameters to obtain power grid voltage deviation parameters includes: Extract the real-time grid voltage value for the current period from the grid state parameters, and obtain the instantaneous deviation by the difference between the real-time grid voltage value and the rated value. Calculate the ratio of the instantaneous deviation to the nominal value, as an instantaneous characteristic; The real-time voltage values of the power grid for multiple consecutive cycles are extracted from the power grid state parameters, and the deviation change trend is obtained by fitting the data using a linear regression algorithm, which is then used as the trend feature. The instantaneous characteristics and the trend characteristics are combined to form the grid voltage deviation parameter.
9. A reactive power control device for a permanent magnet direct-drive wind turbine generator set, characterized in that, include: The data acquisition module is used to collect multi-source reactive power demand signals, unit constraint parameters, and grid status parameters of permanent magnet direct-drive wind turbines. The deviation feature extraction module is used to extract voltage deviation features based on the power grid state parameters to obtain power grid voltage deviation parameters. The weight calculation module is used to calculate the weight ratio of each type of demand signal in the multi-source reactive power demand signal based on the grid voltage deviation parameter and using a dynamic weight allocation algorithm. The data fusion module is used to perform data fusion on the multi-source reactive power demand signals according to the calculated weights to obtain demand data; The instruction generation module is used to input the demand data and the unit constraint parameters into the pre-trained reactive power regulation optimization model to generate the fundamental reactive power regulation instruction. The instruction execution module is used to send the fundamental reactive power regulation instruction to the grid-side converter, and the grid-side converter executes the instruction to realize reactive power control of the permanent magnet direct drive wind turbine.
10. An electronic device, characterized in that, include: At least one processor; And, a memory communicatively connected to the at least one processor; The memory stores instructions that can be executed by the at least one processor, which are executed by the at least one processor to enable the at least one processor to perform the reactive power control method for permanent magnet direct drive wind turbines as described in any one of claims 1 to 8.